Emerging Trends in Mobile Robotics

The following sections are included:

• PREFACE

• CONFERENCE ORGANISERS

• CONFERENCE COMMITTEES

• CONFERENCE SPONSORS AND CO-SPONSORS

• CONTENTS

Since time immemorial, human like artifact has been one of human dreams. From the genesis of robotics research, many persons have researched for the way of realizing the humanoid robot. In the end of the last century, Honda Co. gave humanoid robotics the big impact with humanoid robot P2. We know now humanoid robot is not a dream but is there for the taking. The relatively recent availability of high computational power, energy-dense batteries, and new theories in robotics is boosting up humanoid capability. Now humanoid robots can perform not only bipedal walking but perceiving environments, manipulating objects and interacting humans. In this talk I will be addressed by presenting relevant ongoing research in Intelligent Systems Research Institute, National Institute of Advanced Industrial Science and Technology (AIST) along with other relevant results in this field. The most advanced humanoid robot "Cybernetic human HRP-4C" which has a realistic head and the average figure of a young Japanese female and our attempts to its possible applications in contents industries and show business will be also addressed.

Note from Publisher: This article contains the abstract only.

The biorobotics program at Case Western Reserve University (CWRU) has been active for more than 20 years. This presentation highlights many of the projects undertaken during that time and describes how neuromechanical principles have benefited a number of robots. As this list of principles grows, so does the functionality and performance of the biorobots…

During the last decades, mobile robots have been a classic subject for robotics researchers, covering a large number of topics ranging from image processing, SLAM (Simultaneous Localisation and Mapping), all the way to control technologies and swarm techniques. However, these techniques did not see many applications in road transport until the Prometheus Project in Europe (1986-1994) and the AHS (Automated Highway Systems) in Japan and in the USA and for a long time it was considered by many to be impossible to implement safety critical function in large production road vehicles…

The theoretical method of slip estimation is proposed. The theoretical considerations concerning the slip reduction pattern are followed by simulation example. Presented research is realised in the frame of PROTEUS project aiming the development of autonomous robots for inspection and exploration. The robots will move on natural terrain. Proposed method of slip reduction was tested experimentally and will be implemented in robot control systems.

The understanding of the environment is mandatory for any type of autonomous robot. The ability to put semantics on self-generated sensor data is one of the most challenging tasks in robotics. While navigation tasks can be performed by pure geometric knowledge, high-level planning and intelligent reasoning can only be done if the gap between semantic and geometric representation is narrowed. In this paper, we introduce our approach for recovering 3D scene information from unorganized point clouds, generated by a tilting laser range scanner in a typical indoor environment. This unorganized information has to be analyzed for geometric and recognizable structures so that a robot is able to understand its perception. We discuss in this paper how this spatial information, which is based solely on segmented shapes and their extractable features, can be used for semantic interpretation of the scenery. This will give an idea of how the gap between semantic and spatial representation can be solved by spatial reasoning and thereby increasing robot autonomy.

A system of two-wheeled vehicle is considered in this work. The vehicle is powered with two direct current (DC) motors for driving the vehicle and a linear actuator to activate the payload to achieve a specific height. The driving motors are driven independently to activate the vehicle wheels. A control strategy is developed and implemented on the system in order to steer the vehicle to perform a certain wheels trajectory. Different wheel trajectory profiles are used to check the capability of the developed control scheme to tackle the control problem while maintaining the balance condition of the whole mechanism.

The purpose of this paper is to design and implement a Middle size soccer robot to conform RoboCup MSL league. First, according to the rules of RoboCup, we design the middle size soccer robot, The proposed autonomous soccer robot consists of the mechanical platform, motion control module, omni-directional vision module, front vision module, image processing and recognition module, investigated target object positioning and real coordinate reconstruction, robot path planning, competition strategies, and obstacle avoidance. And this soccer robot equips the laptop computer system and interface circuits to make decisions. In fact, the omni-directional vision sensor of the vision system deals with the image processing and positioning for obstacle avoidance and target tracking. The boundary-following algorithm (BFA) is applied to find the important features of the field. We utilize the sensor data fusion method in the control system parameters, self localization and world modeling. A vision-based self-localization and the conventional odometry systems are fused for robust self-localization. The localization algorithm includes filtering, sharing and integration of the data for different types of objects recognized in the environment. In the control strategies, we present three state modes, which include the Attack Strategy, Defense Strategy and Intercept Strategy. The methods have been tested in the many Robocup competition field middle size robots.

Robocup is an international competition for multi agent research and related subject like: Artificial intelligence, Image processing, machine learning, robot path planning, control, and obstacle avoidance. In a soccer robot game, the environment is highly competitive and dynamic. In order to work in the dynamically changing environment, the decision-making system of a soccer robot system should have the features of flexibility and realtime adaptation. In this paper we will focus on the Middle Size Soccer Robot league (MSL) and new hierarchical hybrid fuzzy methods for decision making and action selection of a robot in Middle Size Soccer Robot league (MSL) are presented. First, the behaviors of an agent are introduced, implemented and classified in two layers, the Low_Level_Behaviors and the High_Level_Behaviors. In the second layer, a two phase mechanism for decision making is introduced. In phase one, some useful methods are implemented which check the robot's situation for performing required behaviors. In the next phase, the team strategy, team formation, robot's role and the robot's positioning system are introduced. A fuzzy logical approach is employed to recognize the team strategy and further more to tell the player the best position to move. We believe that a Dynamic role engine is necessary for a successful team. Dynamic role engine and formation control during offensive or defensive play, help us to prevent collision avoidance among own players when attacking the ball and obstacle avoidance of the opponents. At last, we comprised our implemented algorithm in the Robocup 2007 and 2008 and results showed the efficiency of the introduced methodology. The results are satisfactory which has already been successfully implemented in ADRO RoboCup team. This project is still in progress and some new interesting methods are described in the current report.

Autonomous mobile robots have been popularly employed in several applications especially in soccer player robots considered in Robocup competitions. This paper wants to give a general description to development of an autonomous robot and the cooperative team behavior in dynamic and uncertain environments. We have tried to focus on areas such as omni directional mechanism, cooperative behavior, world modeling, fuzzy decisions and behavior learning. The project is described in two major sections: Hardware and Software. The software is developed in two main parts, one is the server application which consists Network, World Modeling, Global AI, Condition Monitoring and the second is a player application which consists Image processing, Network, local AI, Trajectory Planning and a MIMO Motion Controller. We utilize the sensor data fusion method both in the self localization and world modeling. The localization algorithm includes filtering, sharing and integration of the data for different types of objects recognized in the environment. The hardware consists of the robot itself and the driver circuit board. The methods have been tested in the many Robocup competition field middle size robots. Some new interesting methods are described in the current report.

This article describe the force-based walking with timing control to provide better force control for hydraulic driven hexapod robot named COMET-IV. Impedance controller is combined in order to provide compliant mechanism with environment stiffness during walking phase. The proposed method is verified with some experiments by walking the real-time robot through the random uneven terrain. Results of comparison between methods are presented and discussed.

Large vertical concrete structures are still a great challenge for autonomous climbing robots, which should be able to perform different service tasks like inspection or coating of the rough surface. With the behavior-based obstacle avoidance system of our robot CROMSCI this paper presents a step towards such an autonomous climbing system. Main problems arise from a limited payload for environmental sensor systems. Therefore a special sensor setup and internal representation of the environment has to be found. In this paper we will show the overall structure of our climbing robot CROMSCI using negative pressure adhesion and omnidirectional wheels for locomotion and focus on its sensor systems and the behavior-based components for obstacle avoidance.

The article presents a method to improve stability performance for motion control of modular walking robots moving on uneven terrain. Real time robot's control method through virtual projection used by the author has allowed the development, design, testing and experimenting in real-time control of reaction forces through mathematical modeling. Also, it is presented the analytical calculation method of the elastic element of force measurement sensor, together with the compliant control system with tracking functions for whom have been integrated 6 force sensors achieved through this method.

This work treats the dynamic control of multi-mobile robot formation taking robot dynamic interconnections. The dynamic of each agent is modeled by a nonlinear second order differential equation, and its behavioral control will depends on attractive or repulsive interconnection function. The interconnection dynamic function is built around certain estimated parameters, and taking the dynamic of agents in neighbor. Once the target/objectif is fixed, the formation convergence in presence of known obstacles is obtained through a stabilizing nonlinear sliding mode controller, and under the bound of the interconnection parameters. Some bio-inspired examples can be concerned by our modeling and control approaches, one thinks to the autonomy of a herd of sheep in displacement, a flock of birds or a school of fish, and in generally the problem of swarms.

This work will present the U-Go Robot, a rugged outdoor vehicle under development at DIEES Robotic Laboratories. The main target for this new machine is to be a multifunction system able to comply with different field of applications. The DIEES Robotic Group currently has different active projects in the field of autonomous navigation, new GNSS technologies, agriculture and precision farming applications, artificial vision, data fusion and so on.

In This paper, a mixed robust sliding mode with fuzzy logic controller is applied to a nonlinear quadrotor Unmanned Aerial Vehicle (UAV). Based on the technique of variable structure control (VSC) with sliding mode, we can construct the controller that possesses the merits of both fuzzy Logic Controller (FLC) and Sliding Mode Controller (SMC), It is called a Fuzzy Sliding Mode Controller (FSMC). It provides a simple way to solve the main drawback of VSC by reducing the amount of chattering effect using a fuzzy control part in the proposed controller. Also, the VSC part makes the system stable which treats the main disadvantage of fuzzy controller when being used alone. Numerical simulation of the proposed controller is presented and discussed. Furthermore, a comparison with the SMC is also made.

This paper deals with the clearance capabilities of mobile robots in rough terrain. A way of using kinematic reconfigurability is proposed to allow the crossing of obstacle that would normally be impossible by choosing a configuration that will guaranty static equilibrium. The control uses force control on the legs and try to decrease the internal forces needed to insure stability.

There are many legged robots which perform well statically, but none that can execute robust dynamic maneuvers. One limitation of current systems lies in their leg designs, which are not well suited to dynamic maneuvering. A new leg is designed based on consideration of four areas: joint type, foot design, actuation type, and mechanical coupling. This leg has an articulated hip, knee, and ankle, a substantial (non-point) foot, pneumatic actuation at the knee, and mechanical coupling between the knee and ankle joints. This leg is attached to a torso to form a monopod , and produces stable hopping in simulation, which is confirmed in initial experimentation.

Derived from the results of biological research on dry adhesion mechanisms a technical adhesion module based on similar principles is introduced. The function principle of this module relies on a combination of dry adhesion and a passive suction cup. The area of application covers smooth silicon-based materials and substrates, e.g. glasses. The adhesion module enables all important phases of interaction between adhesive system and substrate including 1. attachment to the substrate, 2. maintenance of contact without the consumption of energy and 3. detachment from the substrate. Therefore the adhesion module is suited as a gripping tool for handling tasks as well as a module for a small-sized glass-surface-climbing robot.

The number of pipe accidents has been increasing recently. Because most of them were caused by corrosion or deterioration, in-pipe inspections have been performed with fibre scopes to prevent such problems; however, these scopes cannot inspect pipes that are greater than 15 m in length or complex pipes. Therefore, in-pipe inspection robots are needed. In this paper, we propose the development of a robot that mimics the peristaltic crawling of earthworms as a locomotion mechanism. In addition, we performed several experiments in a 27-mm-diameter acrylic pipe to examine the relationships between locomotion speed and motion pattern and between propulsive force and motion pattern.

Currently many hazardous maintenance and inspection tasks, such as paint inspection and corrosion condition monitoring of steel structures, are being performed manually by workers, which causes serious health and safety problems. This paper presents a concept climbing robot, with the aim of exploring highly complex ferrous structures such as steel bridges, for the purpose of inspection duties. To demonstrate this concept, a quadruped prototype is developed. A modular architecture that simplifies the development process and improves reusability has been implemented. Permanent magnet compliant pads on each foot provide a simple method of adhesion on the highly complex and unsmooth surface of a bridge. A simple detachment mechanism has been employed. Experiments have been conducted to prove the concept and test the design of the prototype.

Decentralized control is a key concept to understand the mechanism underlying versatile and adaptive locomotion of animals under various environments. However, a systematic design of an autonomous decentralized control system is yet to be realized. To address this gap, we have so far focused on true slime mold, and have extracted a design scheme for brain-body interaction based on discrepancy function. In this paper, we investigate the validity of this design scheme by applying it to the control of a quadruped locomotion. The simulation results show that the quadruped robot exhibits remarkably adaptive behavior against changes of environments and body properties.

Decentralized control is a key concept to understand the mechanism of versatile and adaptive locomotion of animals under various environments. However, a systematic design of an autonomous decentralized control system is yet to be realized. To address this gap, we have so far developed a design scheme based on discrepancy function. However, the spatial distribution of muscle tonus has not been considered in our previous study, although it is crucial for animal locomotions. In this paper, we propose a design scheme where both the phasic and tonic control are considered, by focusing on serpentine locomotion. Simulation results show that adaptive locomotion is successfully emerged through the coordination of the phasic and tonic control.

In this paper we present some first aspects of the design of an arthropod-like walking machine that will serve as a carrier for bioinspired sensors as well as a research platform for walking. The focus in the design of the robot was to create an optimal bodily framework for bioinspired control approaches for walking. Since body features and control structures co-evolved in biological systems, it is fundamental to idenify the most important body aspects in order to benefit from the related bioinspired control aspects. According to this, the robot mirrors the relative dimensions of the stick insect Carausius morosus. This involves the scaled-up dimensions of the body and the legs including the leg onsets, the joint axes orientations and additional 2 DoF joints that were introduced in between the individual body segments. In addition to these structural features, an important functional feature regarding biomimetics is the implementation of passive serial elastic elements in the leg joints.

Insects have incredible walking and climbing capabilities. Therefore, it is not surprising that researchers and engineers try to transfer these capabilities to biologically-inspired robots. This paper will present a biologically-inspired free gait based on one simple coordination mechanism. The resulting statically stable gait is then compared to the gait of a walking stick insect. Finally, the results will show that the presented behaviour-based free gait is able to create a gait very similar to a biological one.

We studied of crawl-motion of lizard type robot and humanoid robot. The stride of lizard type robot can be enlarged by twisting the waist. Using Humanoid robot SANDY-3, the experiment of crawl on the hands and knees was done. SANDY-3 moved forward with making the body wriggled and using the arm and the leg of diagonal position.

A planar pneumatic endoscopic manipulator powered by compressed air is presented. The tip joint is manipulated by the operator through a Bowden cable and all other joints are driven and controlled by mechanically implemented feedback control law that produces propulsion like a snake. Unlike conventional active endoscopes, the proposed mechanism produces propulsion at every part of the body and thus can reach deeper part of a complicated cavity.

Biological systems such as insects have often been used as a source of inspiration when developing walking robots. Insects' ability to nimbly navigate uneven terrain, and their observed behavioral complexity have been a beacon for engineers who have used behavioral data and hypothesized control systems to develop some remarkably agile robots. However, it is now possible to implement models of relatively recent discoveries of the stick insect's local control system (its thoracic ganglia) for hexapod robot controllers. Walking control based on a model of the stick insect's thoracic ganglia, and not just observed insect behavior, has now been implemented in a complete hexapod able to walk, perform goal-seeking behavior, and obstacle surmounting behavior, such as searching and elevator reflexes. Descending modulation of leg controllers is also incorporated via a head module that modifies leg controller parameters to accomplish turning in a role similar to the insect's brain and subesophageal ganglion. While many of these features have been previously demonstrated in robotic subsystems, such as single-and two-legged test platforms, this is the first time that the neurobiological methods of control have been implemented in a complete, autonomous walking hexapod. This paper discusses the implementation of the biologically-grounded insect control methods and descending modulation of those methods, and demonstrates the performance of the robot for navigating obstacles and performing phototaxis.

Reconnaissance robot is a useful robot type for not only the possibility of preparing FCS (future combat system) but of overcoming various obstacles such as military operation, sewer pipe exploration, building collapse spot and so on. For this reason, researches for snake type robot, which has merits, its flexibility and expansion is epidemic in many places in the world. In this research, the robot mechanism, which is capable of not only moving with hiding its position on rough terrain in operation area, but also guaranteeing its efficiency of reconnaissance is concerned with behavioral characteristics of mobility of snake when climbing trees. In this reason, firstly, for satisfying proposition, three modules based on the wheel mechanism were proposed and the spiral climbing method was considered for overcoming the situation when moving on the trees.

This paper discusses the state of the art on biomimetic inspiration in climbing robots. It also studies the anatomical specialization of climbing animals and their application on current climbing robots.

In this paper, we propose a compact kick-and-slide robot that moves by kicking and sliding on the ground repeatedly. The proposed robot jumps low in the air for high moving efficiency. This robot has the large repeated jumping capability by selecting robotic parameters optimally. We show that the developed palmtop size robot achieves the maximum velocity of 1.3[m/s] with the weight of only 69[g].

In this paper, an artificial passive spine and an actuated rear foot with two DOFs for a small and lightweight ape-like robot are presented.

The aim of the LittleApe project is to build a robot that is capable of walking adequately on two and four legs, change from a four-legged posture to a two-legged posture, has the ability to manipulate small objects, and is able to climb. Specific motion sequences require not only movements of the limb, but also of the body. For LittleApe an abstraction was made to find out which DOFs are necessary to realise some of these motion sequences. The genus chimpanzee was chosen as an antetype for the robot. LittleApe is modelled with the same characteristics regarding limb proportions, spinal column, center of mass, walking pattern, and range of motion as the biological antetype. One approach of this work is to equip the torso with a flexible artificial spine which adds central flexibility and additional damping capabilities to the robot's body. The design of the light weight rear foot features force sensors to detect touchdown, measure the applied force, and evaluate the torque applied to the feet in real time.

In this paper, we develop a compact jumping robot based on closed elastic structure utilizing bending and twisting deformation of an elastic strip. By using snap-through buckling generated from not only bending but also twisting of thin rectangle elastic strip, impulsive forces can be generated repeatedly by the frequency of about 4[Hz] without changing added torque directions. Since it can generate impulsive motion repeatedly, the proposed robot can swim against flow of water. In addition, this simple mechanism is effective at a hybrid environment. By using paddle attached to the elastic strip as a leg, it can jump by kicking pieces of ice or floating ground on water.

In this paper, we develop an impulsive turning mechanism for swimming robot based on closed elastic structure utilizing bending and twisting deformation of an elastic strip. By using snap-through buckling generated from not only bending but also twisting of thin rectangle elastic strip, impulsive forces can be generated repeatedly by the frequency of 4[Hz] without changing added torque directions. The proposed impulsive mechanism can turn the swimming robot by 90 degrees only 0.4[s]. In addition, since it can generate impulsive motions repeatedly, the swimming robot which use the proposed turning mechanism can be turned by 270 degrees only on 1.2[s].

This paper presents an investigation into leg passive parameters using passive pendulum test. The passive parameters are viscosity and passive knee stiffness in spastic paraplegic patients. The study is carried out with a representative spastic paraplegic whose knee angle is measured using a twin axis SG series Biometrics ltd goniometer. The study includes a few minutes of physiotherapy between two pendulum test measurements involving heat packs, Infra Red heat, and passive range of motion exercises. The measured data is analysed with the view to extract features of muscle fatigue and development of a more comprehensive knee joint model of paraplegia. Further analysis is made of the biomechanical stimulation to assess the potential of combining physiotherapy with muscle fatigue testing, and viscosity and passive resistance changes in rehabilitating spastic paraplegic patients.

Chemical search is perhaps the most fundamental skill in animals, but it is typically slow. Biological studies have already demonstrated the use of various search methods (e.g., chemotaxis and biased random walk) by various animals and robotics research also tested a number of search methods (Lytridis, et a., 2006). The aim of the present study is to provide a biologically motivated but artificially enhanced search strategy for a single agent in odour source localisation in a confined space. The search performance is boosted by a preliminary fast sampling process to approximate the gradient field of the diffused chemical, which helps the robot move closer to the target at a faster rate than biological strategies would normally do. Using noiseless and noisy fields with two targets, the search strategy is chemotaxis but the pre-search sampling is a random walk search. Based on the cost function of search time with success and average path lengths, the results clearly show the benefit of pre-search sampling. These findings are promising but we need additional testing of other biological search strategies (biased random walk, combined strategies, etc.) under various field conditions, before robotic search performance could be assessed in natural conditions.

This paper is about the design and implementation of artificial neural computation systems for a virtual anthropomorphic hand which can learn to reach and grasp different objects by an unsupervised learning. The grasping fingers are the thumb (finger number 1) and middle finger (number 3). The information about the hand shape can be compressed by a numerical parameters called synergies. The synergies have been studied by Santello et al. [1] and [2]. In this paper, the information of hand shape is compressed by these synergies. This way of compress information simplifies the neural computation [3].

To enhance dexterity of multi-fingered robotic hands, this paper analyzes grasp stability of two objects in three dimensions. Each finger is replaced with 3D spring model. The stability is evaluated from the viewpoint of potential energy stored in the grasp. The contributions of this paper are as follows. Grasp stiffness matrices of two spatial objects with any friction property are derived. Grasp stability and its motion direction are evaluated by eigenvalues and eigenvectors of the matrices. Effectiveness of this analysis is demonstrated through numerical examples.

This paper presents a seemingly novel approach to accomplish cooperative tasks with two dual-arm mobile robots in a vision-based intelligent space. Specifically, the cooperative task considered is to grasp and move a box object to a target position with a ceiling-mounted IP camera. To grasp the box, the desired poses of the dual arms are determined based on the poses of the mobile robots and the box. The movement of the box to its desired position is then performed by two decentralized visual servo controllers for the two mobile robots. In particular, the master robot is controlled to reach its desired pose while the slave robot is driven to maintain the grasping employing compliance control and to keep relative pose to the box using real-time visual servoing requiring no communication with the master robot. The proposed approach has been successfully validated by experimenting with two Dr Robot i90 robots equipped with 6-DOF dual arms. The system can be further applied to cooperative tasks in a large workspace equipped with a set of fix-mounted vision sensors.

A control scheme and sensor system is proposed that allows groups of distributed mobile robots connected by physical link sensors to work co-operatively. Each robot in the group senses rotations and motions in order to follow changes in direction and position of neighbouring robots using information provided by link sensors mounted on the robots. A model of the proposed link sensor system is presented in this paper. The distance between each of the robots can be changed by requiring the robots to maintain the desired link length as necessary for the task in hand. Differences in robot dynamics which would occur between the robots in the real world have been incorporated into our simulations. The control scheme is able to keep the desired robot configuration while executing complex motion trajectories despite the differences in robot dynamics and positioning errors. Simulation results of the control scheme applied to five group robot configurations (horizontal chain of robots, vertical chain of robots, cross configuration, array configuration and a changeable position configuration) demonstrate that the scheme enables the robots in our system to accurately follow the changes in direction and position of the leader robot while maintaining the link length to the required amount.

This paper presents development of fuzzy logic control scheme for positioning an overhead crane system and minimizing the sway of the payload when the system is in motion. Proportional-Derivative (PD)-type fuzzy logic controllers are developed to move the payload to the target location to achieve accurate positioning and minimum sway. The system is realized within the Visual Nastran (vn4d) software environment, and integrated with Matlab-Simulink for analysis and control purposes. Simulation results are then presented in time and frequency domains to assess the effectiveness of the control approach. It is demonstrated that accurate positioning and significant reduction in the system sway is achieved.

This paper presents an alternative for controlling and positioning a two degree of freedom flexible link -whisker-. The aim is the end-point precise position control with effective vibration suppression. The proposed control scheme makes use of an inner loop to control the whisker actuator position and an outer loop that cancels the vibrations due to the whisker flexural dynamics. A new algebraic regulator is designed to control the actuators. This regulator has to cancel the non-modeled components of the friction without a previous estimation. In order to control the vibration, an input state feedback linearization is used and the closed loop poles of the system are placed taking into account the control signal limit that is allowable before the actuators saturation. The simulations show a perfect tracking of the trajectory at the whisker tip.

Advances in control theory has provided a wide variety of artificial intelligence design techniques in addition to the more traditional PID approaches which have been applied to control of flexible manoeuvring systems. Robust optimal control of flexible structures with active feedback techniques requires good models of the base structure, and knowledge of the uncertainties of these models. Such information may not be easy to acquire for certain systems. It has been known that fuzzy-logic-based modelling and control could serve as a powerful methodology for dealing with imprecision and nonlinearity efficiently. On the other hand, classical control laws are designed for linear systems and they provide a preferable cost/benefit ratio. However, the presence of nonlinear effects limits their performances. Hybridization of fuzzy logic techniques and classical PID control schemes will exploit the beneficial features of both categories. This paper addresses the problem of input tracking control of a nonlinear flexible system namely a twin rotor multi-input multi-output (MIMO) system (TRMS) in hovering mode using hybrid control approaches in simulation and real time situations. An ANFIS model of the TRMS is used to represent its behaviour in the simulation environment. The obtained results in real-time prove the efficiency of the proposed scheme.

The following sections are included:

• Introduction

• Prototype of Fish-like Robot with Elastic Fin

• Experiment of Fish-like Robot

• Simulator of Robot with Elastic Fin

• Conclusions

• References

The following sections are included:

• Introduction

• Development of O-VOW

• Control Scheme of O-VOW

• Experiments

• Conclusions

• References

This paper represents rolling of a deformable polyhedral robot with a tensegrity structure. In this paper, we make a prototype of a tensegrity locomotion robot, which corresponds with the right-handed snub cube. We show that a moving strategy of tensegrity locomotion robots using the body deformation depends on the structure itself through analyses of the gravitational potential energy. In the snub cube, the gravitational potential energy at four-point contact is smaller than one at three-point contact, which results in moving directionality.

Based on results of biomechanical analyses, a climbing robot is designed. The main aim of the design process is the reduction of degrees of freedom of this mechatronic system. Therefore, the motion is mainly generated in the trunk. The functions "gripping" und "generation of motion" are separated for the design process. The principle of a flexible trunk-like construction is introduced. The degree of freedom of the trunk is four. A consequent light-weight design leads to an over all mass of only 1 kg. Climbing capabilities are shown and quantified on cylindrical substrates.

A Magnetic Switchable Device (MSD) is a ferromagnetic circuit using permanent magnets where the flux can circulate between different paths when its configuration is changed. This routes or cancels the flux trough specific surfaces, and thus turns on or off adhesion forces. We present classic and innovative magnetic configuration to realize powerful MSD. We designed and prototyped some miniature systems and give their characteristics. Finally various robotics applications for gripper, anchor and climbing robot are unveiled where the MSD solution has proved to be advantageous.

Cy-mag^3D is a miniature climbing robot with advanced mobility and magnetic adhesion. It is very compact: a cylindrical shape with 28 mm of diameter and 62 mm of width. Its design is very simple: two wheels, hence two degrees of freedom, and an advanced magnetic circuit. Despite its simplicity, Cy-mag^3D has an amazing mobility on ferromagnetic sheets. From an horizontal sheet, it can make transition to almost any intersecting sheet from 10° to 360° – we baptise the last one surface flip. It passes inner and outer straight corners in any almost inclination of the gravity. Cy-mag^3D opens new possibilities to use mobile robots for industrial inspection with stringent size limitations, as found in generators. A patent is pending on this system.

In this paper the design and development of a new type of robust fault-tolerant hexapod robot named OSCAR-X is described. The introduced novel design of its legs allows the hexapod walking robot to perform on demand reconfiguration and amputation of its legs. For future adaptation of our robot we have introduced several other design ideas for reconfiguration mechanisms that will improve the reconfiguration ability of the robot. The described ideas can be also universally applied to design of various other types of future legged, wheeled or hybrid reconfigurable robots.

This paper describes the concept, design and prototype implementation of a wheeled pole-climbing-robot with high payload capability, which is planned for climbing palm trees, masts of telephone cables or lamp posts. It uses a new clamping mechanism which is mainly inspired by the rope-clamps that are used for human climbing (common name: "jumars"). By adapting this principle to the field of wheeled pole-climbing robots, a very high payload capability can be achieved – without any additional constraints concerning the stiffness of springs or the position of the robot's center of mass. After an introduction chapter that points to the importance of the above-mentioned advantages in specific applications and the drawbacks of previous designs; the basic mechanical principle of the clamping mechanism is explained with a 2D calculation model. This analysis is followed by the description of the first prototype robots: A preliminary prototype at small size for proving the functionality of the concept; and the conceptual design of a pole-climbing robot for real industrial applications – planned for pole-diameters from Ø0.2 to Ø0.5m and a payload capability of up to 100kg. The paper concludes with successful experiments that have been performed with the prototype robot at small size, and a discussion of the most critical components for the planned design at big size.

This paper describes the design, implementation and application of a lightweight magnetic foot with variable force, which was developed for docking/landing on vertical walls with a micro-helicopter in order to perform inspection tasks there; and afterwards undock to fly away without problems. Even if not being a complete climbing robot, the magnetic foot on the helicopter uses technologies which were originally developed for climbing robots, operates in an environment which is difficult to access (furnaces in coal-fired boilers) and could be useful for future climbing robots with inchworm-locomotion. By using a lightweight actuation concept based on miniature pulley-wire- transmissions, a simple and robust control strategy and a mechanical design that is well shielded against ferromagnetic dust or other types of aggressive dirt; the here presented magnetic foot with variable adhesion force achieves several advantages over previous designs – such as an excellent compromise concerning lightweight design at strong adhesion force (Fadh,max/(m*g)>100), robustness against most environmental hazards, low residual force (Fadh,max/Fred<30) and low manufacturing cost. After explaining the environment constraints, the basic concept of docking with an unmanned micro-helicopter and the requirements for the magnetic foot; a brief overview on most principles for force variation in magnetic feet is provided. This comparison is followed by the presentation of the basic design concept and its implementation in a prototype. The paper concludes with the measured values of this prototype and provides an outlook on future work and how this technology can be used in other applications.

This article describes our progress toward simplifying and streamlining the low level systems integration of experimental robots, combining a System on Chip (SoC) approach with conventional modular approaches. The combined approach has increased flexibility, improved the embedded integration, and decreased the complexity of programming, compared to conventional modular approaches. We show the impact of the SoC approach in a simple demonstration and teaching model of a walking robot.

This paper addresses how to design and construct a glossy surface climbing biped robot and it includes few factors like mechanics, electronics and physics behind mechanism. Glossy surfaces include glossy metal walls, plastic walls, glass walls, ceramic walls, which are used for covering multistory buildings. Generally human cleaners are appointed for periodic cleaning, but to work at sky touching heights for a whole day with single safety belt; it is tedious and dangerous job. While doing that lot of people were badly injured and some of them lost their life. To avoid such accidents, we need to design an autonomous machine that can safely survive on the vertical surfaces of different shapes and sizes and it can perform maintenance works like cleaning, surface leakage checking, surface polishing, etc. More over in the urban areas, surveillance is required to keep watch on activities to avoid any unexpected natural or human disasters. For surveillance high resolution cameras are used and they are placed at safe places, generally urban heights are used. So the simple glassy surface climbing robot can be modified to surveillance robot. This paper explains mechanical construction details, working methods and climbing patterns of two legged robot that can climb on vertical glossy surfaces. It has optimized technology, small structure, light weight, simplicity, easy controlling, can work efficiently at heights, sustainability. It can perform locomotion by using vacuum suckers connected at the bottom of each leg.

This proposed work explains designing procedures for a reconfigurable modular robot and its reconfiguration planning method for locomotion system and three dimensions by hardware. A group of the modules can thus generate various three-dimensional robotic structures and motions. In this paper, we develop a combination of different separately operated mobile robotic modules having reconfigurable structure and when they combine together they will form an assembly that will be adaptive, reconfigurable, and modular in structure. Although the module itself is a simple mechanism, reconfiguration planning for locomotion presents a computationally difficult problem due to the many combinatorial possibilities of modular configurations. It will be suitable for extra terrestrial explorations and having combined locomotion. This locomotion mode is based on multi-modules. Assembly consists of number of parts. Each individual part showing its own existence by locomotion and it will do own work separately. These all parts combine together when they have to lift something and carry and when they have to do some combined tasks. It will become multitasking machine which can survive on different terrains and difficult environments also. MK8 Reconfigurable Modular Robot consists of multi joint robotic arms. These arms have low level of locomotion and are able to reach to each others. There are six robotic arms and there is one rectangular mobile platform. After assembling all parts the structure will become Hexapod which is having dual locomotion modes, first is rolling mode and second is walking mode.

This paper aims to develop a multi-body mobile robot which has the capabilities to climb walls and make wall-to-wall transitions. The developed robot consists of three connected bodies, two links, and ten tracked wheels actuated by nine motors. Six vacuum suction pads are installed on each tracked wheel and one additional suction pad is attached to the 2nd body for the steering motion of the entire robot. While each tracked wheel rotates on the vertical plane, the suction pads are automatically activated by the sequential opening of the mechanical valves in pneumatic cylinders, thus enabling the continuous locomotive motion. The kinematics of the proposed mechanism is analytically studied and the capabilities of the robot are experimentally verified in the case of vertical wall climbing and wall-to-wall transition between 90 degree walls. The overall size of the robot is 1000mm × 1600mm × 300 mm with a mass of about 70kg. The maximum climbing speed and carrying payload are 3m/min and 10 kg respectively.

The explosive bomb deactivator robot (LOCO-EBD Robot) can replace man to recognize, remove and deal with explosive bombs or other dangerous articles in a dangerous environment. The design is composed of a platform mobile vehicle, an articulated mechanical arm, a tele-operated control system, vision system and a wireless communications system. The LOCO-ERD Robot is developed to replace a human deactivator with high manoeuvrability and strong capacity to defeat obstacles, stairs, etc. It can be used in urban areas and in wild environments of sand, grass and soft soil, etc. The robot design enables a very low cost robot to be constructed and uses several commonly available industrial recycled parts to further reduce costs.

A novel hybrid climbing pruning robot that fills the requirement of sustainable forest management is presented. The climbing principal is an imitation of the climbing approach of timberjacks in Japan. The robot's main features include having the center of its mass outside of the tree and an innovative climbing strategy fusing straight and spiral climbs. This novel design brings both lightweight and high climbing speed features to the pruning robot. We report our progress in developing the robot and describe the experimental results of hybrid climbing.

This paper deals with the comparison between human and humanoid walking. The objective is to see if current humanoids are sufficient to faithfully reproduce human walking kinematics. To proceed, motion captured data of human walking were injected in a humanoid model based on human anthropometry and current humanoid robot kinematics. Then, we compared simulation results and human walking measures. This shows that current humanoid robot degree of freedom (DoF) induce significant quantitative errors in reproducing human walking kinematics. In particular, during foot internal-external rotation, knees tend to step out. We believe this could impede energy consumption by increasing speed and torque requirements in robot joints. Thus, we propose different solutions to solve this problem: either changing one flexion-extension joint orientation of the leg, or adding a new DoF.

The problem of dynamic postural equilibration taking into account the role of compliant feet is solved. The equilibrium conditions are split between the feet attachment points and the points within the feet-end area. The presented method is useful for motion synthesis taking into account robot parameters.

This paper deals with an approach that combines passive walking of biped-robots with active driving. First, it proposes a hybrid transmission based on a planetary gear system that can receive two input torques; one is an active torque generated by the rotary actuator and the other is those caused by an elastic element such as springs. This mechanism allows passive walking in almost all walking period but impulsively propels it to produce the walking on a flat surface with minimum energy consumption. The passive joint inevitably requires static stability to sustain the body weight, for which the four-bar linkage mechanism is employed as a rotary joint. The swing motion of the shank requires some amount of active torque about the knee joint during a swing phase. Therefore this paper also proposes a self-energized system, by which pneumatically stored energy by the passive rotation of the ankle joint during a stance phase is utilized to swing the knee joint during the subsequent swing phase. This mechanism provides a human like motion of the knee and ankle joints with no external energy feeding.

The goal of this paper is to clarify the influences of toe trajectories and head motions on the walking and trotting gaits of a four-legged robot, called "Ryuma." We chose four and five trajectories relative to the body for swing phase and stance phase, respectively. Through simulation and experiments with the real robot, we found suitable trajectories for walking and trotting gaits. Then, we found that head motion affects the forelegs' real toe trajectory

In this paper we introduce the robotic quadrupedal platform ALoF that is designed to aid research on perception in legged locomotion. A well-balanced size and complexity of the robot results in a robust platform that is easy to handle, yet able to perform complex maneuvers as well as to carry sophisticated 3D sensors. A very large range of motion allows the robot to actively explore its surroundings through haptic interaction, and to choose between a wide range of planning options. This robot was employed and tested in the lunar robotics challenge organized by the European Space Agency, for which we also developed a novel crawling gait, in which the weight of the robot is alternately supported by scaled plates under the main body and the four shank segments. This allowed for stable locomotion in steep terrain with very loose soil.

In this study, dynamic model, simulation and control of a six-legged hexapod robot have been performed for a bounding gait. To obtain the numerical simulation of the system with bounding gait, the different dynamic structures have been solved in a sequential closed loop (stance, flight phase). Spring Loaded Inverted Pendulum (SLIP) has been used as a dynamic model to simplify the simulation of the system and ease to understand the bounding gait. In addition, the hexapod robot with bounding gait has been controlled by Fuzzy Logic Control algorithm. The number of the performed control action during the phases of bounding gait in this study is smaller than those of the studies in the literature. Consequently, the proposed improvements in this study make both the system control easy and the system performance increase by decreasing the run time for each loop.

While more mature method of gait planning for humanoid robots are approached, there are not many applied to soccer robots. A kid-size humanoid soccer robot is developed and this paper introduces it's mechanical design and put forward a kind of off-line gait planning. From the results of simulation on MOS2009, the application of methods above was proved acceptable.

This paper deals with a control method for an achieving walking in the case of an angular sensorless of partial leg links. In this study, the partial leg links are thigh links which support the weight of the body. Concretely, Kalman Filter is composed by using the mathematical model, and the posture control is applied.

To enhance the well known, self-stabilizing effects of Spring Loaded Inverted Pendulum (SLIP) models, researchers have proposed a variety of dead-beat controllers that adjust model parameters (angel of attack, spring stiffness), such that a disturbance is rejected within a single step. While such laws can be nicely encoded for disturbances in hopping height (by using the time of flight as a measure of vertical position), they suffer from substantial drift due to the missing information about the forward velocity and inaccuracies in the actual system model. Without requiring additional complex sensors, we propose a method to estimate the forward velocity of a SLIP model based solely on measuring the time of stance. This method is additionally able to perform realtime parameter estimation, which paves the road to implement a full state dead-beat controller that can reject arbitrary disturbances even in the presence of model and sensor errors.

This paper presents a passive dynamic walking model with segmented feet. The model extends the Simplest Walking Model with the addition of flat feet and torsional springs based compliance on ankle joints and toe joints, to achieve stable walking on a slope driven by gravity. The push-off phase includes foot rotations around toe joint and around toe tip, which shows a great resemblance to human normal walking. Experimental results show the effects of adding segmented feet to passive dynamic walkers on energetic efficiency.

Passive running robot can run by interaction between machine dynamics and environment only. 1–legged, 2–legged, and 4–legged passive running exist. Each leg has a translational spring. This study focuses on a passive running robot using the bouncing rod dynamics and not the robot using spring. In our earlier work, we have experimentally demonstrated a passive running based on bouncing rod dynamics. However, the passive running was too unstable. In this paper, we realize a stabilization of the passive running by lizard–like foot. Moreover, we demonstrate a possibility of existence of stable fixed point.

This paper presents a walking control method to achieve stable biped walking on unknown and uneven terrain. It is important to suppress an impact force and to place a foot on the actual ground for landing on such terrain. Because the conventional foot generally has a rigid body, it is difficult to suppress the impulsive reaction force. We first propose to attach a spring on each foot to reduce the impulsive reaction force. Then, to achieve these two objectives simultaneously by adjusting the foot position, this paper presents a PI force controller with a desired foot position to land on unknown terrain. This controller provides the robust stability of control system with respect to terrain variance and exact positioning of the foot to actual ground. This controller is applied to a simple biped robot that has spring embedded feet. The effectiveness of the proposed walking control method on uneven terrain is demonstrated by using a dynamic simulator.

Quadruped robot has the advantage of static and dynamic gait on rough terrain. But if the robot uses hydraulic power system, the fluid consumption must be considered. Especially in a restrictive hydraulic power system for stand-alone, a large fluid consumption can leads to the pressure drop and pressure drop caused lack of torque. Therefore an efficient utilization of hydraulic power system is very important. This paper proposes the fluid consumption minimized motion planning algorithm for the hydraulic actuated quadruped robot.

In this paper we introduce two new mechanisms that developed for a passive walking robot. First mechanism is upper body mechanism which develops to reduce waist assist force and body's stabilizer's energy consumption. This mechanism is employed by one servo and four urethane sheets to operate stabilizer function. This mechanism's experiment results show its low energy consumption and the next step of humanoid robot's upper body mechanism. Second mechanism is ankle spring mechanism which designed to reduce impact force upon walking and reduce waist assist force. This mechanism presents some interesting results and makes a different in waist assist energy consumption. Finally we demonstrate experiment results of two mechanisms that show some interesting points.

Contact interaction with the environment is crucial in the design of locomotion controllers for legged robots, to prevent slipping for example. Therefore, it is of great importance to be able to control the effects of the robots movements on the contact reaction forces. In this contribution, we extend a recent inverse dynamics algorithm for floating base robots to optimize the distribution of contact forces while achieving precise trajectory tracking. The resulting controller is algorithmically simple as compared to other approaches. Numerical simulations show that this result significantly increases the range of possible movements of a humanoid robot as compared to the previous inverse dynamics algorithm. We also present a simplification of the result where no inversion of the inertia matrix is needed which is particularly relevant for practical use on a real robot. Such an algorithm becomes interesting for agile locomotion of robots on difficult terrains where the contacts with the environment are critical, such as walking over rough or slippery terrain.

The paper presents a modified linear actuator for the biped robot "ROTTO". The elastic element of a special design is discussed as well. On the basis of the designed electromechanical system the force control structure is researched and the realization of an artificial spring with the prescribed rigidity and positioning with the given impedance is also investigated.

Terrain classification of the legged robot is one of the most important objects which can determine robot's performance, because ground surface of the fields is often extremely diverse. In the flat surface case, robot can move fast and smoothly. However, it cannot move fast in the rough terrain. Unless robot knows which terrain, robot will be falling down and slippery. Therefore, robot must know their terrain when they are walking. In this paper, we composed a 1-legged robot and terrain environment (flat, sand, gravel, and grass) for terrain classification experiment. A load cell mounted on the 1-legged robot measures the ground reaction force and torque sensors are located each of the 3-joints for measuring torque. Then we present a method for terrain classification using statistical method (Variance, Skewness, and Kurtosis). After that we compare these relationships with statistical values.

This work presents an integral approach to tackle energy issues in bipedal robots. It introduces three combined ideas to increase these robots' autonomy: First, the exploitation of the inherent equilibrium that should exist in the rest position of a well-designed mechanism; then, the efficient usage of the energy to walk based on the natural limit cycle of the system; and finally, the harvest of energy based on the new idea of regenerative walking. Simulations and experimental tests show promising results of this approach, built under a delicate equilibrium between appropriate control scheme, suitable mechanical design and proper actuators choice.

This paper addresses the motion planning problem (MPP) on the kinematic model of six-legged robots. It is assumed that nonholonomic constraints and drift are present which make the motion planning nontrivial while the stratified nature requires additional attention to path planning. An algorithm to drift neutralization derived from stratified framework to six-legged robot is proposed. The algorithm yields generally an approximation. The main motivation is to move the robot along a reference path that assists resolving obstacle avoidance problem. Configuration switching is supported by the use of supervisory control theory. The results are illustrated on a simple kinematic model of six-legged robot.

In this paper, we discuss the effect of upper body on the stability of passive dynamic running (PDR). To this end, we have modeled a passive dynamic running biped with upper body in a numerical simulator. Through our numerical simulations, we have found the following results: (i) Periodic stable solutions exist; (ii) The higher position of COM destabilizes PDR; and (iii) The inertia of upper body stabilizes PDR. Our results shed new light on the principles underlying bipedal locomotion; in particular to what extent the stabilization mechanism can be explained by passive dynamics underlying whole body dynamics properties.

The success of an interaction control scheme is tightly related to the capacity of the actuators to follow the force references as much accurately as possible. The use of conventional actuators, usually modest at force control, lead to a inadequate performance of the interaction control system. This paper presents a leg prototype designed and developed for achieving agile locomotion. In the design, the selection of the proper actuation system has been a key issue for the development of the interaction control scheme. The HADE leg is actuated by means of three Series Elastic Actuators, and experimental results show the good performance of the impedance control performed with these actuators.

This paper deals with the stability analysis of a novel biped walking and climbing robot, which is aimed to perform a dynamic walking on flat or irregular surfaces and a static climbing on vertical or sloped walls, because of the action of suction-cups.

Each serial-parallel leg mechanism shows 6 d.o.f.s since composed by a serial connection of two 3-RPS (Revolute-Prismatic-Spherical) parallel modules. The transient motion of the biped robot from a horizontal terrain to a vertical wall is possible because of a suitable rotation of 90° that can be performed by each foot. In particular, a stability analysis is proposed through a suitable investigation of the worst configurations of this biped walking and climbing robot in order to design each rectangular foot properly.

The actuator is an essential part of robotic bipedal walking. The general approach is to transfer joints for bipedal legs from robot arms and therefore they are not well adapted to walking. This paper presents a model for an biologically motivated actuator. The actuator is designed for a human-like biped to perform highly dynamic motions like running and jumping. There are two commonly used concepts in the field of robotics. The first favors a stiff actuator with high gear-ratio and position controller. The second utilize a compliant actuator that usually features low power consumption and is torque controlled. The actuator presented in this paper combines the benefits of both concepts. A theoretical model is derived to improve the dynamic behavior and examine the energy consumption of the actuator and amplifier. Several test with a prototype leg and in a simulation environment are presented.

This paper proposes a novel locomotion control scheme of centipede-like multi-legged robot, which is called Follow-the-Contact-Point (FCP) gait control. A centipede-like multi-legged robot is composed of segmented trunks which have a pair of legs and are connected with fore and/or rear ones by joints. This control scheme realizes locomotion control of multi-legged robot on uneven terrain with perfectly decentralized manner. The main concept of the control scheme is to relay the contact points from the fore leg to the rear leg. By creating contact points of the first legs on the environment adequately, the robot can climb over obstacles and be navigated successfully. Finally, the result of physical simulation of 20-legged robot shows the availability of the proposed method.

The main objective of this paper is to evaluate experimentally the influence of different structural parameters on passive dynamic gaits. For this purpose a passive dynamic walker with knees and arc-shaped feet has been designed and built. Incremental encoders have been added on the hip and knee joints and force sensors have been installed not only in the sole of the feet but also the thighs of the prototype. Furthermore, each leg segment is also equipped with an inertial measurement unit. The acquired information is then used to determine how the changes in the ramp angle affect the step period, the step length and the walking speed. Conclusions could play an important role on the generation of gaits for actuated biped robots.

This paper introduces a novel continuous path tracking method based on curvatures called Weingarten Maps and implementation of this method in a quadruped robot model. The path to be followed is assumed to be defined in Cartesian coordinates. The path tracking method is achieved both using open and closed loop algorithms. Quadruped gaits are generated due to instantaneous center of rotations of the sample path which are calculated using Weingarten Maps. Quadruped robot model is simulated in ADAMS cooperated with MATLAB/Simulink.

Legged locomotion of autonomous humanoid robots is advantageous but also challenging since it inherently suffers from high posture instability. External disturbances such as collisions with other objects or robots in the environment can cause a robot to fall. Many of the existing approaches for instability detection and falling prevention include a large number of sensors resulting in complex multi-sensor data fusion and are not decoupled from the walking motion planning. Such methods can not simply be integrated into an existing low-level controller for real-time motion generation and stabilization of a humanoid robot. A procedure that is both easily implementable using a minimal number of affordable sensors and capable of reliable detection of posture instabilities is missing to date. We propose a simple, yet reliable balance control technique consisting of a filtering module for the used data from two-axes-gyroscopes and -accelerometers located at the trunk, an instability classification algorithm, and a lunge step module. The modules are implemented on our humanoid robots which participate at the yearly RoboCup competitions in the humanoid kid-size league of soccer playing robots. Experimental results show that the approach is suited for real-time operation during walking.

An often used stability criterion in legged locomotion is the zero moment point (ZMP). The ZMP is a virtual point calculated based on the center of gravity (COG) position and acceleration and must be kept within the support polygon at all times for the robot to be stable. Therefore, maintaining stability relies critically on good tracking of the planned COG trajectory and traditionally high gain PID control is the most commonly used to ensure good tracking performance. However, PID control has a severe disadvantage in that it relies critically on exact knowledge of the terrain and exhibits suboptimal stiff behavior in case of external perturbations. Here we will show that the inverse dynamics controller allows us to lower the PID gains while at the same time maintaining sufficient tracking performance. The main contributions of this paper are (a) the formulation of the ZMP planning as constrained optimization problem, and the proof of its performance on difficult terrain (b) inverse dynamics based compliant control to ensure tracking of the ZMP and (c) showing that the compliant controller outperforms high-gain controllers in case of severely degraded terrain perception and thus increases the robustness in locomotion over difficult terrain. We present results on the quadruped robot LittleDog.

This paper proposes an estimation method for a quadruped robot with dynamic gaits to estimate the body configuration and speed, and simultaneously the average angle between the inclined plane and the horizontal. A nonlinear observer including the simplified dynamic model and leg kinematics, and the sensor noise is designed. The sensor dynamics of low frequency governs the estimator characteristics. We filter the sensor uncertainties as well as the kinematic uncertainties using a nonlinear sliding-mode observer with fast convergence. We implement this estimation procedure online. The efficiency and performance of the developed method are verified through computer simulations and experiments using a quadruped robot.

Wheeled mobile system is generally the mainstream as a transporter. However, its activity area is limited at flat ground. Legged robots are expected to walk on irregular terrain with their high ground adaptability. In this paper, we develop a biped transporter robot based on passive walking robot. Passive walking robot can walk down gentle slope by gravity only. Our concept of biped transporter robot is to carry a baggage by human–assist only. We demonstrate that our biped transporter robot can carry the baggage on irregular terrain and soil surfaces and so on.

This paper presents the climbing of high steps by a humanoid robot (AIT-T6; height: 140 cm; weight: 19 kg) with parallelogram linkage legs. It is difficult for robots with such legs to climb high steps because of the limited movable range of the legs. The effectiveness and performance of the proposed motion control are verified using the humanoid robot. The robot was able to climb steps with a riser height that is 40% of the length of the robot's leg.

In the paper, we consider stability of three-dimensional passive dynamic walking for various length of leg, CoM position and foot inclination, etc, through experiment. As a result, we realized stable walking by adjusting balance position so as to adjust position of center of total mass which is varying with respect to leg length. We also found that there is a tendency for a biped to walk well if CoM position is kept constant regardless of leg length for the same foot shape. Unstable tendency for walking with long stride had been already known for flat feet model. The result showed the same tendency for spherical feet model.

The authors have proposed a novel method for high-speed gait generation of limit-cycle walkers based on the forward-tilting impact posture. Based on this approach, the robot can overcome the potential barrier at mid-stance easily and can generate a high-speed level gait only by extending the stance leg during stance phases. The problem was that there is not enough time-margin for stance-leg actuation due to the excessive high-speed motion. In this paper, we then attach forefeet to the legs of a telescopic-legged rimless wheel for the purposes of braking and tilting the impact posture more. The length of forefeet is finite and the stance leg rotates around the tiptoe just prior to heel strike. The simulation results show that the geometric effect of forefeet on the impact posture strongly improves the gait efficiency in terms of walking speed and specific resistance.

In this paper, we derive the stability conditions of passive dynamic walking using Poincaré map and Jury's stability criterion. We studied how the normalized physical parameters of a biped robot influence its stable one periodic walking, via stability conditions and mathematical analysis as well as numerical simulations. As a result, we found that the influence of the variation of slope angle on the stability condition was more reduced as the leg mass came closer to the hip joint.

This paper deals with alternative humanoid robot dynamics modelling, using the screw theory and Lie groups called the special Euclidean group (SE(3)). The forward dynamic model is deduced analitically, which is solved by propagation method from an end-effector to the center of gravity (COG) always on the SE(3). Many tests for reference dynamic walking patterns have been carried out, which are represented in simulation and experimental results. The results will be discussed in order to validate the proposed algorithms.

In our earlier work, we realized a assist level walking of passive biped walker that added upper body by means of spring mechanism. Moreover, we demonstrated that the upper body increases efficiency of walking. However, its mechanism is not understood well. In this paper, we model a passive walking that added upper body by means of spring mechanism. We demonstrate that the passive walking with upper body can generate a stable limit cycle.

This paper presents an analysis on generating turning motion of a humanoid robot by slipping the feet on the ground. Humans unconsciously exploit the fact that our feet slip on the ground and realize smooth turn without stepping many times. In order to generate the slip motion, we need to predict the amount of slip. We propose the hypothesis that the turning motion is caused by the effect of minimizing the power generated by floor friction. A model of rotation is described on the basis of our hypothesis. The case that a robot applies the same force on both feet is discussed. The feet trajectory to maximize the rotational angle is also investigated based on the proposed hypothesis. The arc trajectories of both feet are adopted to generate turning motion in order to realize intended rotational angle. The hypothesis is verified through experiments with a humanoid robot HRP-4C.

Walking machines designed in Volgograd State Technical University moving along relatively flat surface, is that might be used with cycled stepping mechanisms united into movers gears so that one of such mechanisms will be always set into the phase of interaction with soil. Such solving of the problem makes the machine statistically stable as well as it provides the possibility of ignoring the keeping the gait and affords concentration upon the design of controlling system which is analogous to controlling system in machines with traditional types of movers. We consider the problem of control of motion of a many-legged statically stable locomotion machine ensuring the choice of a minimal-power motor. We show that the machine frame can be used as a kinetic energy recuperator, thus providing a significant decrease in the reactive power. It has given an examples the applications elaborated walking machine an different branches of industry and agriculture.

Stair climbing is one of the key issues for walking robots involved in urban search and rescue missions. This article presents a parametrized stair climbing strategy for a six legged walking robot. By the parameters we mean the height and depth of the steps provided on-line. These data connected with the information about the robot dimensions allows automatic stair negotiation. The strategy is a closed-loop control based on the position of the robot on stairs. Additionally, to allow the correction of the horizontal position and orientation of the robot on the stairs the information about the distance from the side-walls and about orientation of the robot w.r.t.the stair-steps should be provided. The strategy was preliminary tested on the robot simulator. The validation on the real robot is in progress.

This paper presents a method for terrain classification based on haptic feedback and its application to natural terrain samples. The samples were composed of eight different kinds of loose ground and three different solid terrains. The method is based on features extracted from contact force, leg joint position, and joint motor current measurements, which are used to train and evaluate a classifier. For classification, a multiclass AdaBoost machine learning algorithm has been employed. In experiments, an average of about 87% of eleven different ground types could be assigned correctly. The robustness of this classification was evaluated with respect to static and dynamic measurement errors.

This paper presents several results regarding the lateral and longitudinal control systems that have been applied for the automation of an articulated bus, using a rolling wheeled box system with special design that moves inside a guide rail. Nowadays, transport systems are achieving major advances by the incorporation of automation based technologies. Recent developments of electronic instrumentation and actuation systems and the increasing speed of processors allows for the implementation of real-time systems. The automation of an articulated bus provides combined advantages of both conventional bus and train, because it can ascend slopes of 15% and turn on curves of low radius. This transport modality is an interesting, low cost and friendly option. In this paper an experimental setup for the development of lateral and longitudinal control of the articulated bus is presented. Comprised by an experimental mobile platform (articulated bus) fully instrumented and a ground test area of asphalt roads inside CSIC installations, this experimental facility allows full testing of automatic driving systems.

Magnetic wheels are a powerful solution to design inspection climbing robots with excellent mobility. Magnetic wheels optimization based on simulations and the results that were obtained on prototypes are presented. The measured adhesion was doubled between the classic configuration and a novel multilayer one sharing exactly the same four magnets and the same total volume of iron. This know-how is then applied to optimize magnetic wheels for the existing robot called MagneBike. The adhesion force has been multiplied by 2 to 3 times depending on the conditions. Those amazing improvements open new possibilities for miniaturization of climbing robots or payload's increase.

Our purpose is to develop a simulator to emulate movements of a real small sized bike robot under feedback control.¹ We configured the bike robot model by using 4 simple rigid links and derived motion equations for the bike robot from simple constraints among those links. Simple viscus friction is considered in the simulator. We compared the results of the simulations with the movements of the real bike robot under the same feedback control algorithm and the same control coefficient with the bike robot. We confirm that the results of the simulations are almost similar to the motions of the bike robot on the whole. In this paper, we report that our bike-robot model, the simulator and the simulation results compared with the motion of the real robot. We consider it is sufficient to simulate a motion of the bike robot.

Up to the presents, omni-directional mobile robot uses a special wheel such as omni-wheel and mechanam wheel. However, it has problems with kinematic performance. In order to improve these problems, we developed a novel omnidirectional moving wheel as the Differential Drive Steering System (DDSS). Experimental results show the proposed system has high performance.

The rough terrain mobile robot "RT-Mover", which is a leg-wheel-type robot built of very simple mechanism, can move on continuous rough terrain.¹ However, in a real environment there is also discontinuous rough terrain, where it can not get about. The step-up gait for an upward step has been studied to walk on discontinuous rough terrain. In this paper, a flow of the step-up gait is introduced. After that, kinematics to climb up a step is discussed in detail, and is evaluated through simulation and experiment.

The inspection of marine vessels is currently performed manually. Inspectors use sensors (e.g. cameras, devices for non-destructive testing) to detect damaged areas, cracks, and corrosion in large cargo holds, tanks, and other parts of a ship. Due to the size and complex geometry of most ships, ship inspection is time-consuming and expensive. The EU funded project MINOAS develops concepts for a Marine Inspection Robotic Assistant System to improve and automate ship inspection. A central part of MINOAS is to evaluate the use of a cooperative fleet of robots, including areal drones, magnetic climbing robots, and underwater crawlers, for ship inspection. In this paper we describe a first concept for one component of the MINOAS robot fleet, a magnetic crawler for the inspection of large cargo holds and large tanks. We show how a light-weight system using magnetic wheels (including hybrid leg-wheels) and a passive magnetic tail can successfully climb tall metallic walls and overcome small obstacles.

In recent years, some tracked mobile robots have been developed that run over irregular terrain with remote control. They have a number of DOF to improve their mobility and are capable of running over stairs and steps. However, their remote operations are complicated, because they are controlled by operators using a control pad such as used in computer games or toys. In those cases, it is hard to capture an operational sense and easy to make operational mistakes. In order to easily control robots with many DOF, we propose an intuitive remote control system, using the conventional master-slave operation. In this paper, we introduce KUROGANE 0&1 crawler robots, controlled by the intuitive remote operation. KUROGANE 0&1 consist of a typical crawler, plus a human-like upper body. We operate these by means of a wearable remote controller. Thus, the KUROGANE0 has achieved the ability to crawl swiftly over irregular terrain with high stability. We investigated the prototype robot KUROGANE0's performance over outdoor rugged terrain, as well as steps and stairs. Based on the result, we newly designed a 'KUROGANE1' and have done some experiments. MPEG footage of these experiments can be seen at: http://fukuoka.ise.ibaraki.ac.jp.

In this paper authors present description of the original type of mobile hopping robot, which have four linear drives. The construction of this robot supports different jumping regimes and has a simple electromechanical structure. The experimental and mathematical model of the robot and results of simulation are suggested.

This paper describes a synchronized control of hip and knee joints of a wheel-driven legged robot to reduce vertical swing and speed irregularity. Robot legs are jointed at the rim of the wheel and are able to rotate by the power transferred coaxially on hip joint axes. Therefore, the robot body swings vertically while it walks by the rotation power of hip and knee joints, normally. To generate comfortableness of the walk by regulating the swing and irregularity, we propose two types of synchronized control among the two joints under the static condition that neglects the mass of the robot. The first type limits the knee joint rotation in a certain range, i.e., swing motion, and the second type allows continuous rotation of the knee joint. The walking based on the two types are simulated and compared for clarifying the motion difference. In the last, the experiment using the PEOPLER-III proves the validity of the two types.

This paper describes a load allocation control in the motion transfer for a multi-locomotion robot which has two legs and two arms. The load allocation control focus on allocating the torques of supporting joints by changing the torques of one or some important motors. Instead of setting the important motors' voltage by trying in our motion transfer, we explain how to set the parameters in theory in this paper. As a result, the maximum torques of joints is reduced by setting the suitable voltages to the hip motors. And all the motors can work in a safe condition even if there are some position errors existing. The robustness of the designed motions with load allocation control is experimentally verified.

In this paper the modeling and simulation of an autonomous vehicle is presented. The goal is to create a real-time capable model of the complete mechatronic system: this model is being used to analyze the functional safety of the vehicle's onboard embedded systems by means of co-simulation and hardware-in-the-loop tests. The paper reflects some modeling thoughts that had to be met in order to guarantee a good simulation performance using the Modelica language and Dymola environment.

This paper proposes a method for enhancing the robustness of the Central Pattern Generagor (CPG) based a 3D neuro-musculo-skeletal walking controller. The CPG has been successfully applied to walking controller and controllers for walking robots. However, the robustness of walking motion with the CPG based controller is not sufficient especially when it is subjected to the external force or environmental variation. In order to achieve a realistic and stable walking motion of the simulator or controller, we propose the attracting controller in parallel with the CPG based controller. The robustness of the proposed simulator is confirmed through the simulation results.

In this paper, an evaluation of a simulation-based morphology design is presented. The designed system is a six-legged walking and climbing robot, developed within the context of the SpaceClimber project. The robot should be able to traverse rough terrain and to climb in lunar craters. In the design phase of the morphology, an extensive use of real-time simulation and evolutionary computation was made. The resulting morphology, obtained by evolutionary optimization, was used to design a real-world robot. By comparing the physical behaviour of the real and the simulated robot, an evaluation of the simulation-based design approach is performed. The question is answered if the performance of the real robot fulfils the expectations generated by the simulated robot.

Mechanical model of miniature robot with sliding seal is considered. Robot move along the incline plane surface. Conditions of secure moving without sliding of the wheels are derived. These conditions are assigned design control system of the robot. Suggesting method of motion control is based on the combination control by the vacuum pump module providing pressing force fore robot and driving wheels module providing motion along the straight line and maneuver rotation. The motion conditions without slipping and oscillations were established depends on the structure of robot's coupling mechanism contacted with surfaces.

This paper presents a method to design a robust controller for interval parallel kinematic machines. The method involves using a systematic approach to design a robust stabilizing controller for a parallel kinematic machine that can be described by nonlinear second order differential equation with interval parameters. To formulate the interval parameter problem in joint space the dynamic equation which is derived in Cartesian space for parallel kinematic machines is transformed to joint space using the Jacobian matrix and its time derivative. The nonlinear joint space model is linearized yielding an interval linear model and a controller with interval parameter gains is then found. A numerical study on a dynamic model of Stewart platform based parallel kinematics machine demonstrates the design procedure and shows its effectiveness. In the study, the model is simulated assuming lower and upper limits for some of the parameters.

In this paper a novel IDSO (Information Driven Self Organizing) controller for a simple parallel kinematic machine is described and discussed. This control strategy maximizes the 'predictive information' of the controller. It is analyzed in depth the case in which the uncertainty on the joints follows (concentrated) Gaussian distributions under Markovian conditions. Under these hypotheses it is possible to derive a compact expression for the predictive information and consequently simple reinforcement learning rules maximizing it. This is possible thanks to the properties of Lie Groups (and related tangent space Lie algebra) of kinematic structures, studied from G. Chirjjkian et al., which allow to carry out the study of stochastic kinematic chains in a simplified way. It is thought that emergent controllers of this kind may enable in the future the development of more robust and adaptive CLAWAR systems.

The main goal of this paper is to report a successful attempt of monitoring the stiffness behavior of a biped walking vehicle during its dynamic operation. In particular, a suitable procedure has been developed and implemented for estimating the stiffness performance of WL-16RV as based on previous experiences at LARM with Milli-CaTraSys design. The proposed experimental set-up is composed of six linear encoder wire sensors and two force/torque sensors embedded in each foot. Experimental tests are conducted on a biped walking vehicle WL-16RV under static and dynamic conditions. The experimental tests provide useful information for both design and control purposes.

Human upright balance control can be explored using movable platforms driven by servo-controlled torque motors (dynamic posturography). The purpose of the paper is the study of the design of a three-degree-of-freedom movable platform to provide postural perturbations for balance assessment and training. The platform consists of a non-overconstrainted spherical parallel mechanism with an adjustable center of rotation. The structural design of this parallel mechanism is discussed with respect to the particular application of postural perturbation. Then the kinematic performance analysis is performed by the aid of screw theory. Dynamic simulations are used to show how the platform may be used for producing postural perturbations and the influence of the variation of the rotation center position on balance-correcting responses.

When a parallel manipulator is intended to became a parallel kinematic machine, practical tools for the end user should be provided. In this paper some guidelines for the definition of the structural analysis procedure are proposed. These rules provide the designer with the information to define some useful and simple maps. These maps include the information about kinematical, static and dynamic characteristics of the parallel robot. Considering three approaches the structural workspace is defined. The first approach consists in a finite element based model. With the information from this model, the strategy for the experimental analysis campaign is obtained. Mode shapes and the range of natural frequencies obtained in the finite element analysis are taken as a starting point for the experimental analysis. Furthermore, analytical dynamical models have been developed. These models are a practical analysis tool with relative low complexity. All these three approaches lead to a practical and simple tool to analyse the structural behaviour of the manipulator in its workspace: the structural workspace. The structural workspace allows to the machine's end user to identify the most suitable zones of the workspace where the manipulator satisfies the structural requirements of each particular task.

In this paper a novel 5 Degrees of Freedoms(DoFs) parallel kinematic polishing machine is presented, which could perform automatic polishing for the freeform parts. The T3R2(3 Translations and 2 Rotations) 4URHU-1URHR parallel mechanism is first proposed, it is composed of four 6-DoF kinematic chain and one 5-DoF constraint chain, the polishing tool is installed on the platform. The inverse kinematics is developed and the dexterity index is derived, so that an optimal design for the parallel mechanism is carried out, based on the optimal results the workspace with specified dexterity index of the parallel mechanism could meet the requirements of polishing space. The singularity-free polishing path generation is then carried out with the adjustments of orientation of the NC rotary table. Simulation results show that in the whole polishing process no singular positions are included within the workspace. The work provides theoretical analysis for the development of prototype.

We develop a new sensing technique to estimate the friction produced between the foot of a biped walking robot and the floor. This technique is based on the criteria of effective non-contact and non-attachment, which are considered when studying a robot's walking motion. Our technique involves measuring the electrostatic induction current generated from triboelectricity that results from a change in the electric potential of a walking robot. The waveform of the induced current contains cadence components of both feet. The walking motion of a commercially available biped robot is detected by measuring the current generated in the robot walking under non-contact and non-attached conditions. An occurrence model is proposed for the electrostatic induction current generated as a result of the change in the electric potential of the robot. The waveforms of the electrostatic induction current generated by the biped robot's walking motion on a wood, paper, an acrylic, and PVC floor are observed. It is evident that the differences in the floor materials are reflected in the waveform and in the amplitude of the observed peaks. It is considered that friction conditions affect the waveform of the electrostatic induction current since the obtained waveform strongly depends on the floor material.

We have developed a new sensing technique for the detection of subtle human motion under rubble; this is an effective noncontact, nonattached technique that does not require data processing. Our technique involves the generation of an electrostatic current associated with a change in the electric capacitance between the human body and a given measurement electrode. The waveforms of the current generated by various types of human motion under simulated rubble were observed. The observed waveform of the induced current contains peaks that can be attributed to various types of human motion. An occurrence model is proposed for the electrostatic induction current generated as a result of a change in the electric capacitance. The validity of the proposed method is confirmed through experimentation performed using simulated rubble. This method is useful for detecting survivors trapped under rubble in real time under noncontact and in-situ conditions, and it does not require data processing.

Autonomous operation of a walking robot on a rough terrain requires to build a 3D map of this terrain. Due to the limited payload and energy in a walking robot its exteroceptive sensors have to be compact and light-weight. The miniature Hokuyo 2D laser scanner is a practical choice. An application of this sensor in local terrain mapping supporting footholds selection is presented. Different geometric configurations of the perception system are analysed, and the one that is best for terrain profile acquisition is chosen. A method for real-time building of a local grid-based elevation map from the noisy 2D range measurements is proposed, including novel algorithms that remove map artifacts resulting from qualitative errors in range sensing. Experimental results are presented.

Modular robotics offers a wide spectrum of applications due to exchangeable functional modules. Compliant drives provide the advantage that stored energy can be used to support the robot's locomotion. Therefore interrelationships have to be analyzed e.g. between elasticity, energy storage and the underlying application. An experimental setup of compliant drives has been developed.

Research in legged locomotion is directed to new service applications such as lower-limb exoskeletons and agile-locomotion robots. The main difficulty that is limiting their further development is the lack of appropriate actuation systems. The excessive material overhead of high power conventional technologies and the need of means to obtain gait energy recovery are some among the features which need to be improved. The aptitude of an actuation system to perform a certain task can be judged from a little set of parameters describing its rated and maximum capabilities. This paper presents experimental data of a variety of promising actuation technologies, and a final discussion on their suitability for empowering legged robots.

This paper describes methods that employ monocular vision in order to increase discriminative properties of the laser-based line segments typically used in SLAM. Visually salient objects are represented by clusters of photometric features and aggregated together with the basic line segments extracted from the 2D laser scanner data. An improved matching strategy for the visually salient features is introduced. Three recent photometric feature detectors are evaluated in the context of the proposed mapping approach, taking into account visual saliency, repeatability, and computing efficiency.

This paper discusses the problem of multisensor perception for autonomous stair climbing. The perception system is mounted on the Messor six-legged walking robot. The robot, due to its static stability while walking, is able to traverse obstacles in urban space, especially stairs. Messor while climbing stairs uses an adaptive algorithm, which exploits on-line perception of the stair geometry and robot pose with regard to the stair. The ascent procedure consist of three main parts. The first – preparation – measurements are performed in order to obtain information about the geometry of the stairs. The second – climbing – ascending each stair with correction of the robot orientation and horizontal position on the stairs. The third – landing – detection of the last stair and the end of the stair climbing procedure. The paper is focused on the multisensor system and the perception algorithms.

This paper describes the perception system for an automatic articulated bus where an accurate tracking trajectory is desired. Among the most promising transport infrastructures of the autonomous or semi-autonomous transportation systems, the articulated bus is an interesting low cost and friendly option. This platform involves a mobile vehicle and a private circuit inside CSIC premises. The perception system, presented in this work, based on 2D laser scanner as a prime sensor generates local terrain maps, where the major concern lies in detecting and tracking a tunnel guide rail built-in the circuit, by using a hybrid-efficient line extraction algorithm and the obstacle detection to guide the vehicle down the lane.

In this work some considerations, design and tests of different kinds of electro-adhesive pads are discussed. Electro-adhesion exploits Columbian Forces due to an electrostatic field, generated by means of a high voltage pulsed DC Power Supply Unit, applied to conductive electrodes laid on different materials. Inter-digited (with different shapes) and parallel electrode couples have been created and tested over different surfaces. Tests have been performed both with polypropylene coated and non-coated electrodes. Analytical model and FEM analysis have been made in order to validate results of experimental tests. These pads can be used as holding devices for climbing machines as they can adhere to different vertical surfaces. In spite of their high holding tangential forces, these pads can be easily detached by using 'peeling' methodology, although powered up.

This paper presents the design of a novel powered ankle-foot prosthesis with compliant joints and segmented foot. The powered compliant ankle is proposed to replace the able-bodied ankle which can provide sufficient power to propel the body upward and forward during bipedal walking. In order to make the walking gaits of the amputees more stable and natural, we introduce segmented foot with toe joint to the prosthesis. Both the ankle and toe joints are driven by two series-elastic actuators (SEA), which not only provide enough torque, but also tolerance shocks. Preliminary experimental results show that the powered prosthesis with compliant ankle and segmented foot can reproduce the human walking gait and be easily used in walking rehabilitation.

In the development of a wearable robot for locomotion, fall prevention is an important problem. In this study, we propose a fall prevention control of a wearable robot using ground reaction force sensors which measure 3-axis ground reaction force and position of center of pressure. We conducted a measurement experiment to evaluate user's compensation for backward instability during walking with a wearable robot. The results suggest that the proposed control is effective to reduce voluntary effort to avoid backward balance loss.

In our current research, we are developing a robotic walker system with standing, walking and seating assistance function. Our developing system is based on a walker which is popular assistance device for aged person in normal daily life and realizes the standing and seating motion using the support pad which is actuated by the novel assistance manipulator mechanism with four parallel linkages. In this paper, we analyze the natural seating motion with fewer loads to the patient which is recommended by nursing specialists and discuss the required condition which realize it. For discussing its required condition, we investigate the seating motion of aged people who requires to power support and typical seating motion by healthy young people. Comparing with two motions, we set the reference of seating motion with our system and we discuss the required assistance condition during seating motion for our robotic walker.

This paper describes the design of a non-linear control law for a reconfigurable stair-climbing device which has the objective of achieving an excellent system stability; which will lead to an acceptable level of comfort and assure security to its user. In order to accommodate the stair-climbing system to the different geometries of the architectural barriers, the prototype needs to change its configuration. For this reason we propose a control law based on a PI controller plus a compensation law for the nonlinear terms, with the aim to control the system inclination. In addition, there is a mechanism that selects the appropriate system configuration. This control scheme has been compared to a classic PI controller scheme based on a linearized model. The result of this simulation demonstrates a better performance in regard to high precision trajectory tracking and therefore a more robust system.

A standing style transfer system, ABLE, is designed to assist a person with disabled lower limbs to travel in a standing position, to stand up from and sit down in a chair, and to go up and down steps. The ABLE system comprises three modules: a pair of telescopic crutches, a powered lower extremity orthosis, and a pair of mobile platforms. In this system, the telescopic crutches are useful not only to maintain the body stability in a standing position, but also to supply power when standing up from a chair, etc. Each hip and knee joint of the powered lower extremity orthosis has an actuator. This module actively fixes, bends, and stretches each joint. The mobile platforms use crawlers to enable the person to travel even on uneven ground. These platforms also enable the user to turn on a rotation board mechanism. Experimental results related to basic and indispensable operations for our daily life confirm the design's effectiveness.

This paper presents a Novel Intelligent Lift-Type Walking-Assist Mobile Robot (NILTWAMOR) developed to support the user's body weight and assist in walking motion. The NILTWAMOR can automatically adjust the hoisting force of the lift and walking velocity according to the user's health level and the intention of the user's lower limbs, and this proposed control system can maintain human posture and walking behavior without falling down. The present walking support system of NILTWAMOR is compared with a conventional lift-type walker machine, and the NILTWAMOR is shown to be superior with respect to flexibility and adaptability through experimental results.

Development of a small and light robot that can excavate by itself is required for planetary exploration. The robot conducts in situ analysis of geological samples and deploys devices for measurement and observation to investigate the underground of a planet. Although excavation robots have been developed, they have satisfied a demand that did not yet exist. Then we focussed on the peristaltic crawling of an earthworm as a locomotion mechanism and developed a small excavation robot. First, we proposed that the excavation robot should consist of two parts: propulsion and excavation. We use an earth auger for the excavation part. The propulsion part moves by peristaltic crawling and it must restrain the unit against the rotation reaction of the earth auger. Therefore, we constructed a peristaltic crawling robot that has a dual-pantograph mechanism. Finally, we measured the basic properties and features of a unit working in dirt. Also, we confirmed that the robot can move vertically, both upwards and downwards, in an acrylic pipe. Good performance was observed in the experiments.

Important tasks of mobile robots are environment recognition and obstacle avoidance to behave on real world. Legged type mobile robots can be adjusted for difficult terrains because robots have a high degree of freedom in order to three dimensional obstacle avoidance. In this paper, we propose an locomotion strategy on terrains having several different stiffness. A possibility of continuous walking is determined by force sensors mounted each leg of a quadruped robot. The presented approach is verified availability from experimental results using a prevalent quadruped robot called TITAN-VIII.

In this paper, control algorithms for kinematically redundant 3-wheeled mobile robots are proposed. One of popular mobile robots in this type is the poly-articulated expandable 3-wheeled planetary rover "Tri-Star3", which was developed by Tokyo Institute of Technology in Japan. Because each wheel of the robot has one kinematic redundancy, the robot is able to make diverse configurations. However, this robot cannot fully utilize the kinematic redundancy because it employs only three actuators and there are some singularities inside the workspace. Thus, it uses discrete control algorithms to control its motion. To enhance the motion capability, a modified wheel mechanism, which has no singularity and is able to utilize the redundancy, is proposed. The performance of this device is compared to that of Tri-Star3.

The following sections are included:

• Introduction

• Arrival Accuracy Improvement on Image Pointing Teleoperation

• Experimental Results

• Conclusions

• Acknowledgments

• References

The subject of this paper is pose control of a differential-drive mobile robot in a laboratory set-up without obstacle. The control approach has been developed with reference to the problem of pose stabilization of nonholonomic systems, which cannot be solved by linear classical methods. The implementation of an improved polar coordinate design approach on the two-wheeled mobile robot Pioneer 3-DX is described. The experimental results obtained are compared with simulation results using open source Blender for robotics.

This paper proposes a method of improving accuracy of self-position estimation with metallic landmark for mobile robot. Many methods of the past self-position estimation researches have been using GPS, Laser Range Scanner and CCD Camera. However, these methods may be unable to get landmark information correctly depending on environment. Metallic landmark is useful in those environments which conventional sensors cannot work well. Accuracy of self-position estimation was increased by combining information from metallic landmark with information from other equipments.

This paper presents a desired trajectory generation of the biped base joint that keeps the CoP position at the constant position when periodic external forces are exerted. For a single sinusoidal external force, the desired trajectory is set based on the frequency transfer function that relates the desired angle and the external force to the ankle torque. For generic periodic external forces, Fourier expansion is introduced, and the Fourier coefficients are estimated based on the framework of an adaptive control. Simulations will demonstrate the effectiveness of this method.

Many researchers have recently been studying micro robot on various kinds of technology such as mechanism, actuators, sensors, control method, and energy supplying method. On the other hand, some contests of micro robot or micro machine are held every year actively. Thus, not only researchers but also general people and students of universities or high schools participating in those contests are having great interests in micro robot or micro machine. Our authors' group also has been studying 1[cm³] volume of micro robot. So far, in our studies, we proposed a structure of an electromagnet type of micro robot which can turn with only one electromagnet coil, and investigated its characteristics in running straight. In this paper, we proposed a control method that enables the micro robot to take a turn with only one electromagnet coil. We confirmed the effectiveness of the proposed control method by some experiments.

This paper proposes a control strategy for legged robots that jump continuously. This control strategy tries to utilize mechanical elastic elements to reduce actuator torque, and consists of the following four parts. The first part is a design of desired motions for stance phases. The second part is a motion controller for the stance phases. This controller is a resonance-based controller, which is proposed by the authors, and adjusts stiffness of elastic elements installed in each joint of the robots. Then, actuator torque will be reduced while generating the desired motions. The third part is a controller for keeping constant postures in flight phases. The fourth part is a controller for generating landing motions. Simulation results demonstrated the effectiveness of the proposed controller.

In this paper, we propose a motion strategy for a Climbing Parallel Robot (CPR) named TREPA. This motion strategy allows the robot to slide through a metallic orthogonal structure. Three kinematic performance indexes are introduced in order to assess the robot's performance during the execution of the task. By means of this strategy, we found the sequence of movements and trajectories that allow the robot to move from one column to another while avoiding collisions, singular configurations and the robot's joint limits.

This paper presents the development and implementation of a fuzzy logic control algorithm for a tree climbing robotic platform - Kamanbaré. Kamanbaré is a four leg bio-inspired robotic platform, whose main goal is to climb trees for environmental research applications such as gathering botanical specimens and insects, performing climatic and arboreal fauna studies, among others. A fuzzy control algorithm was then developed and implemented to manage the prototype robot, controlling the joint positions of its legs. Simulation results of the control algorithm are presented showing that fuzzy control can, in fact, perform quite well for a particular multilink robot.

In this paper we present a mobile robotic platform designed for experimental analysis of the robot control during sliding at high velocity. We use the developed platform for experimental analysis of the mobile robot performing aggressive 90 degrees steering maneuver at high speed (about 8 m/s) on the highly slippery surface (Coulomb friction coefficient about 0.4). The maneuver is performed in feedforward manner by the controller, which was previously developed using methods of stochastic multiobjective optimization applied to the simplified mathematical model of the robot. The theoretical trajectory of the maneuver assumes significant oversteering associated with large slippage angle (more then 30 degrees), which is kinematically incompatible with no-slipping condition and thus is significantly dependent on actual properties of the wheel-terrain interaction. The experimentally observed trajectory of the robot was qualitatively similar to the one obtained in the model, though the actual angle of turn was less then the desired (about 75 degrees instead of 90 degrees).

Nowadays treadmill's control has become more and more popular in many fields, such as athletic exercise, rehabitation training, computer game, and so on. In the previous treadmill usage, the subject passively follows the speed of treadmill. However, in many application cases, the treadmill control strategy by human will is very desirable. Focusing on this problem, in this paper, by analyzing the formula of center of pressure (COP) and simulation results, a key index which indicates the intended walking speed was found. Based on the experiment data, a new model of intended walking speed was established, and further calibrated by least-squares regression technology. The new treadmill control strategy proposed in this paper was built with the intended walking speed model. The treadmill experiment shows that our approach is able to control the treadmill velocity smoothly, which verifies the validity.

Effect of flexibility in the spine of a quadruped robot on its energy efficiency and stability is studied. The approach is based on learning parameters of cyclic position command to the actuators. The learning system employs a reinforcement learning method to find the parameters that result in stable running while the velocity-energy ratio is minimized. The learning is done on a simulated system and the final results are tested on a situated robot. The results show that a wider range of parameters results in stable running for softer spine. In addition, robot with harder spine consumes more energy.

This paper presents the walking gait design and optimization for the LOCH humanoid. On the basis of walking phase division, gait planning is conducted online phase by phase, and accordingly gait parameters are updated with external inputs. After description of the general idea, gait design is introduced in detail for every phase. Following gait design and parameterization, optimization is performed with genetic algorithm for better stability margin. Simulation and experiments show the design of gait enables the LOCH robot to walk stably and naturally.

This paper considers the problem of controlling both the planar position and the orientation of an underactuated airship with a reduced number of actuators in the presence of a unknown persistent wind disturbance. The airship is a nonholonomic system described by a set of nonlinear equations and the dynamics are subjected to bounded uncertainties. A smooth and time-varying coordinate transformation is proposed to reduce the disturbance rejection problem of the airship to that of a simple linear time-invariant system. A new adaptive robust feedback controller provides global stabilization and asymptotical rejection against the unknown wind disturbance under the plant uncertainties. The proposed design method is simple and straightforward. Simulations are performed to validate the effectiveness of the proposed controller.

A generic learning control (LC) approach for various type robotic systems in point-to-point motion tasks (PTPMT) has been developed. It employs simple, but consistent with the optimal control theory, test control functions. A modular program system with a user-friendly interface is developed where appropriate neural networks (NN) are involved for the first time. The program system was successfully verified using two dynamic models: a two-link robotic manipulator and a mobile robot with two independently driven wheels.

The new method of the synthesis of multi-dimensional adaptive control system with reference model self-adjustment for the centralized control of the spatial motion of autonomous underwater vehicles is developed in this paper. The conditions of the self-adjustment process stability with the presence of essential dynamic reciprocal effect between all control channels are obtained and strictly proved. The application of synthesized control laws allow to provide the high control quality at any variations of the object parameters within the given ranges. The efficiency of synthesized control system is confirmed by numerical simulation results.

The paper presents the performance of a control system designed to maintain a desired negative pressure to support the payload of a wall climbing robot with a Vortex machine on different climbing surfaces. The control objective is to provide fast response to changes in pressure demands and air leakage when a vacuum chamber is dragged over different climbing surfaces by a wheeled robot. Experiments have been performed on a single vacuum chamber in which a negative pressure is created by a very cheap AC motor and control is applied by changing the speed of the motor. A pole-placement control system is shown to decrease the time constant of the Vortex system from an open-loop value of 2.85 second to a closed-loop value of 0.15 second.

Functional Electrical Stimulation (FES) is a promising method to restore mobility in paraplegia. Development of a proper control strategy for the FES depends on an accurate model of the muscle. The muscle model consists of relatively well known time-invariant passive properties and uncertain time-variant active properties. A model structure comprising muscle contraction and activation of the quadriceps muscle is formulated. The fuzzy model thus formulated is optimized using genetic optimization, and validated against experimental data. Then, the performance of the model is compared with that of well known Riener's mathematical model in terms of accuracy of the dynamical characterization of quadriceps muscle.

This paper exposes the experimental evaluation of a new technique for the estimation of the instantaneous helical axis of movement of human anatomical joints. The measurement technique, using a six degrees of freedom spatial electro-goniometer, is tested onto a simple revolute joint and onto a subject's knee. A motion capture system with active optical markers is used at the same time in order to validate the measurement results.

Spinal cord injury (SCI) can be due to either traumatic or nontraumatic disorders, such as lateral or multiple sclerosis and tumours. Functional Electrical Stimulation (FES) is the application of a controlled electrical stimulus to the intact peripheral nervous system in order to provide muscular contraction and to produce a functionally useful movement. Muscle fatigue is an important issue while applying electrical stimulation and it restricts the performance of the muscle. Leg extension exercise can be used on its own to increase muscle strength or as preparation for the SCI subject to proceed to other FES based activity. Applying an appropriate control technique is thought to be an effective way of reducing muscle fatigue. The aim of this paper is to investigate the impact of using fuzzy control and PID control through a leg extension exercise in a modeled SCI subject.

This paper presents the effectiveness of spring brake orthosis (SBO) for functional electrical stimulation (FES)-assisted paraplegic walking with wheel walker. The work is a first effort towards restoring natural like swing phase in paraplegic gait through a new hybrid orthosis, referred to as spring brake orthosis (SBO). This mechanism simplifies the control task and results in smooth motion and more-natural like trajectory produced by the flexion reflex for gait in spinal cord injured subjects. The study is carried out with a model of humanoid with wheel walker using the Visual Nastran (Vn4D) dynamic simulation software. Fuzzy logic control (FLC) is developed in Matlab/Simulink to regulate the muscle stimulation pulse-width required to drive FES-assisted walking gait and the computed motion is visualised in graphic animation from Vn4D. The results show that SBO can reduce torque and stimulation pulses required compared with stimulation without SBO for FES-assisted paraplegic walking with wheel walker.

This paper presents the optimal design of the functional electrical stimulation (FES)-assisted indoor rowing exercise (FES-rowing) with spring orthosis using multi objective genetic algorithm (MOGA). The indoor rowing exercise is introduced as a total body exercise for rehabilitation of function of lower extremities through the application of FES. The spring orthosis is used to assist the paraplegics rowing exercise by reducing the required electrical stimulations for activation of quadriceps and hamstring muscles. The FES-rowing performance largely depends on suitable selection of controller and spring parameters. Single objective optimisation techniques can hardly provide good solution in such cases. Multi-objective GAs with fitness sharing technique is used to find optimal set of solutions which trade off between the conflicting objectives. The performance of the FES-rowing with spring orthosis is assessed in terms of the knee trajectory and electrical stimulation required by the muscles to perform the FES-rowing. The rowing model is developed using Visual Nastran Software (vN4D). Fuzzy logic control (FLC) is implemented to control the knee and elbow trajectories during the rowing manoeuvre.

Many injures during childhood are related to the use of playground equipment. Until recently, scientific data of how children actually use playground equipment were scarce. Childhood injury cases were not examined thoroughly from the perspective of how equipment can be modified for improving safety without ruining its attraction to children. To design age-appropriate and safer playground equipment, it is essential that scientific data on the interaction between children and this equipment be accumulated. Herein we report on studies to develop new playground equipment by applying sensor technology to examine the science behind children's interaction with playground equipment. We developed a rock-climbing wall equipped with force sensors to record the physical behavior of children while on the wall, thus allowing measurement of these behaviors in a more natural environment. Fifty force sensors installed in the developed rock-climbing wall are able to collect a large amount of data while children are playing with the equipment. The behavior data of 623 children were recorded in the present study. Herein, we also report on a child behavior prediction model created from the collected data.

One of the problems existing in the area of maintenance of systems for the transport of mass and/or energy is to examine the integrity of the service lines into the cities and industries infrastructure. To carry out this task is required count on help of appropriate technological tools; in these cases it is important to the use of mechatronics systems, more specifically robotic inspection systems. The pipeline inspection robots, have several purposes during the travel into the pipeline, like: cleaning the pipe, removal of liquids, separation of products and inspection, among others. This paper aims to present the main robotic systems used for pipe inspection, and the simulation of the movement of an inspection robot within the flow, taking into consideration parameters such as: Operation pressure, pipeline properties, fluid properties, and physical constrains of a typical inspection robot.

An autonomous cleaning robot is proposed so as to move between floors in a high-rise building. Presently, many of cleaning robots are not considered to move on places between floors. In human living environments, it is often the case that the cleaning area is three-dimensional space such as a high-rise building. Therefore, the cleaning robots can not clean the area such as stairs. In this paper, a mechanism and a control method are described for the robot climbing down and cleaning stairs. The usefulness of the proposed robot is demonstrated through the simulation.

This research proposes the development of a Throw and Collect inspecting device composed by a throwing mechanism, a drawing mechanism, and a child machine, aiming the enhancement of the accessibility to disaster sites. Past reports concentrated on the throwing mechanism, enabling to throw the child machine high in the air with a pneumatically driven novel unit called Magnetic Brake Cylinder [1]. In this report, methods to throw a tethered child machine are discussed in order to obtain higher throws while using the same available pneumatic energy as when throwing a wireless child machine. Simulation of these methods is compared to experimental results, and a difference of only 10% of height between wireless and tethered throws is verified.

Autonomous walking robots represent a special category of robots, characterized by the fact that the power supply and technological equipment are on-board the platform. The weight power supply and technological equipment are important parts of the total load the walking machine can transport. The performance of modular walking systems is closely related to the adopted gait. The movement of the mobile systems can be divided in two types: conditioned by the static stability; conditioned by the dynamic stability. The loss of static or quasi-static stability may be produced in two cases, namely: the vertical projection of the center of gravity is outside of the support polygon, one or more tangential components of the reaction forces from the support points are greater than the friction forces. The movement of the walking robot under dynamical stability conditions may be realized when the forward average speed is greater than a certain limit. Different methods of leg adjustments and body adjustments are integrated into the strategy. This paper explores the various configurations of robots, having experience with modular robot MERO.

The paper presents an update on the ISO robot standardization activities which are developing safety standards for personal care robots (non-medical) and for medical robots. The Committee Draft of the non-medical personal care robots (ISO/CD 13482) safety standard has been accepted via international balloting and work is underway to produce the Draft International Standard. These new type of robots have been classified into three categories, namely, mobile servant robots, person carrier robots and physical assistant robots. The work for formulating the medical robot safety standard has recently been started by the creation of an ISO Study Group and an update on this work is also provided.

This paper presents the mechanical construction of a climbing robot with wheeled locomotion and adhesion through permanent magnets. This machine is intended to be used in the inspection of different types of ferromagnetic structures, in order to, for instance, detect weaknesses due to corrosion, particularly in fuel tanks, ship hulls, etc. The vehicle will have a semi-autonomous behaviour, allowing a remote inspection process controlled by a technician, this way reducing the risks associated with the inspection of tall structures and ATEX places. The distinguishing characteristic of this robot is its dynamic adjustment system of the permanent magnets in order to assure the machine adhesion to the surfaces, even when crossing irregular and curved surfaces.

Welds used in shipbuilding, pressure vessels, nuclear reactors and petrochemical industry utilize multiple pass techniques to ensure complete weld fusion. Any problems found at the late stage require potential removal of multiple weld passes and high-volume weld assembly production, the delay between welding and inspection result in hundreds or even thousands of out-of-tolerance welded parts. Due to high temperature of the weld pool (above 1000°C) and harsh welding environment, manual inspection during the welding process is impossible. So an automated weld inspection system is essential for the welding industry. This paper discusses an automated welding inspection system developed by the London South Bank University and its application to the melt weld pool inspection. The system is capable of providing near real-time inspection of the weld pool concurrent with the welding process. This paper presents both manual and automated test results obtained from the melt weld pool during the welding processes in the field trial, and mentions the parameters need to be concerned for the melt weld pool inspection.

AUTHOR INDEX