International Journal of Humanoid Robotics

Humanoid robots have become more and more popular, which is illustrated by the increasing number of available platforms and the huge number of high-quality publications in the research area of navigation and motion planning for humanoids. Recently, a lot of progress has been made in the areas of 3D perception, efficient environment representation, fast collision checking, as well as motion planning for navigation and manipulation with humanoids, also under uncertainty and real-time constraints. All these techniques work well for their independent application scenario, however, currently no system exists that combines the individual approaches. Thus, we are still far from the deployment of a humanoid robot in the real world. The goal of this special issue is to identify gaps in the research directions and to discuss which aspect need to be considered for combining the different approaches so as to enable humanoids to reliably act and navigate in real environments for an extended period of time.

We present a novel optimization-based motion planning algorithm for high degree-of-freedom (DOF) robots in dynamic environments. Our approach decomposes the high-dimensional motion planning problem into a sequence of low-dimensional sub-problems. We compute collision-free and smooth paths using optimization-based planning and trajectory perturbation for each sub-problem. The overall algorithm does not require a priori knowledge about global motion or trajectories of dynamic obstacles. Rather, we compute a conservative local bound on the position or trajectory of each obstacle over a short time and use the bound to incrementally compute a collision-free trajectory for the robot. The high-DOF robot is treated as a tightly coupled system, and we incrementally use constrained coordination to plan its motion. We highlight the performance of our planner in simulated environments on robots with tens of DOFs.

Accurate and reliable localization is a prerequisite for autonomously performing high-level tasks with humanoid robots. In this paper, we present a probabilistic localization method for humanoid robots navigating in arbitrary complex indoor environments using only onboard sensing, which is a challenging task. Inaccurate motion execution of biped robots leads to an uncertain estimate of odometry, and their limited payload constrains perception to observations from lightweight and typically noisy sensors. Additionally, humanoids do not walk on flat ground only and perform a swaying motion while walking, which requires estimating a full 6D torso pose. We apply Monte Carlo localization to globally determine and track a humanoid's 6D pose in a given 3D world model, which may contain multiple levels and staircases. We present an observation model to integrate range measurements from a laser scanner or a depth camera as well as attitude data and information from the joint encoders. To increase the localization accuracy, e.g., while climbing stairs, we propose a further observation model and additionally use monocular vision data in an improved proposal distribution. We demonstrate the effectiveness of our methods in extensive real-world experiments with a Nao humanoid. As the experiments illustrate, the robot is able to globally localize itself and accurately track its 6D pose while walking and climbing stairs.

This paper presents a motion planning method which provides a complete whole body trajectory of a humanoid robot traveling in a complex environment with various life items, steps, slopes, gates, walls, etc. A two-stage planning scheme is employed to simplify the problem of discontinuously changing dynamical constraints, where a sequence of double-support postures is planned in the first stage, and a continuous and smooth trajectory interpolates it in the second stage. RRT is utilized in the both stages in order to exploit the whole body with a large degree-of-freedom and perform in such a three-dimensionally intricate field. An inevitable issue of that random-sampling-based approach is that the searched trajectory is necessarily jaggy and detouring. The proposed method includes effective post-processing techniques for each stage, which are mandatory in practice. In the first stage, a necessary condition is that any pair of adjacent postures has to share one supporting foot as the pivot. A key idea is to insert intermediate postures between a pair of non-adjacent milestones with a cross-combination of the support feet to bypass and average a sequence of double-support transitions. The optimal sequence of the possible transitions is found by applying Dijkstra's method. In the second stage, an idea of the center of mass (COM)-enforcement to artificially constrain the trajectory of COM on one based on the mass-concentrated approximation is proposed, expecting that a solution can be found on this manifold. A continuous and smooth trajectory which satisfies the geometric constraint is planned by RRT and NURBS interpolation, and then, the temporal property of the trajectory is adjusted based on the dynamic programming to satisfy the dynamical constraint without breaking the geometric constraint. The performance was examined in a mock field of a living room. Comparing with a naive method only with RRT-connect, it substantially shortened the total travel distance in many cases. It also succeeded to plan dynamically feasible walking trajectories on that complex terrain.

This work presents a method to handle walking on rough terrain using inverse dynamics control and information from a stereo vision system. The ideal trajectories for the center of mass (CoM) and the next position of the feet are given by a pattern generator. The pattern generator is able to automatically find the footsteps for a given direction. Then, an inverse dynamics control scheme relying on a quadratic programming optimization solver is used to let each foot go from its initial to final position, controlling also the CoM and the waist. A 3D model reconstruction of the ground is obtained through the robot cameras located on its head as a stereo vision pair. The model allows the system to know the ground structure where the swinging foot is going to step on. Thus, contact points can be handled to adapt the foot position to the ground conditions.

The emerging research area of a quantum-inspired computing has been applied to various field such as computational intelligence, and showed its superior abilities. However, most existing researches are focused on theoretical simulations, and have not been implemented in systems under practical environment. For human–robot communication, associative memory becomes essential for multi-modal communication. However, it always suffers from low memory capacity and recall reliability. In this paper, we propose a quantum-inspired bidirectional associative memory with fuzzy inference. We show that fuzzy inference satisfies basic postulates of quantum mechanics, but also learning algorithm for weight matrix in associative memory. In addition, we construct a communication system with robot partner using proposed model. This is the first successful attempt to overcome conventional problems in associative memory model with a robot application.

Traditional satellite recognition usually applies low-level features (e.g., invariant moment) to describe the global information of satellites. Therefore, the local property and the latent high-level semantic concepts of satellites are likely to be ignored. Recently, the topic model has been given more attention in object recognition field, which generally supposes that the local feature can be modeled as words and images are regarded as documents. Based on this assumption, it aims to discover the latent topics existed between words and documents and then utilizes the topics to represent the images. However, topic model often uses bag-of-words (BoW) strategy that each local feature descriptor is represented by only one word. In this paper, we propose a novel method called sparse coding (SC) based probabilistic latent semantic analysis (SC-pLSA) for satellite recognition. Compared with conventional topic models, our method performs sparse coding to explore the potential correlation between the descriptor and multiple words. Consequently, the descriptor can be described by a small set of words. Based on this word-level representation, probabilistic latent semantic analysis (pLSA) model with simplicity and low computational cost is employed for learning the latent topics. Experimental results on the BUAA-SID 1.0 multi-view satellite dataset validate the effectiveness of our proposed method, and especially under the interference of noise, it outperforms the traditional recognition methods: Hu invariant moment, BoW, SC, locality-constrained linear coding (LLC), conventional pLSA and latent Dirichlet allocation (LDA) with three different classifiers: nearest neighbor (NN) classifier, linear SVM classifier, and sparse representation classifier (SRC).

In this paper, we present a novel methodology to obtain imitative and innovative postural movements in a humanoid based on human demonstrations in a different kinematic scale. We collected motion data from a group of human participants standing up from a chair. Modeling the human as an actuated 3-link kinematic chain, and by defining a multi-objective reward function of zero moment point and joint torques to represent the stability and effort, we computed reward profiles for each demonstration. Since individual reward profiles show variability across demonstrating trials, the underlying state transition probabilities were modeled using a Markov chain. Based on the argument that the reward profiles of the robot should show the same temporal structure of those of the human, we used differential evolution to compute a trajectory that fits all humanoid constraints and minimizes the difference between the robot reward profile and the predicted profile if the robot imitates the human. Therefore, robotic imitation involves developing a policy that results in a temporal reward structure, matching that of a group of human demonstrators across an array of demonstrations. Skill innovation was achieved by optimizing a signed reward error after imitation was achieved. Experimental results using the humanoid HOAP-3 are shown.

The paper proposes a synergy-based myoelectric control strategy for prosthetic hands. Synergy is first reviewed in the context of hand movement, then postural synergy-based proportional and simultaneous control has been introduced to prosthetic manipulation via the principal component analysis (PCA) framework. Experiments have been comprehensively carried out on lab-developed prosthetic hand called SJU-5 to evaluate the proposed method. It is evident that the synergy driven myoelectric control achieves the targeted objectives and performs well on the SJU-5 prosthetic hand.

Most of current biped robots are active walking platforms. Though they have strong locomotion ability and good adaptability to environments, they have a lot of degrees of freedom (DoFs) and hence result in complex control and high energy consumption. On the other hand, passive or semi-passive walking robots require less DoFs and energy, but their walking capability and robustness are poor. To overcome these shortcomings, we have developed a novel active biped walking robot with only six DoFs. The robot is built with six 1-DoF joint modules and two wheels as the feet. It achieves locomotion in special gaits different from those of traditional biped robots. In this paper, this novel biped robot is introduced, four walking gaits are proposed, the criterion of stable walking is addressed and analyzed, and walking patterns and motion planning are presented. Experiments are carried out to verify the locomotion function, the effectiveness of the presented gaits and to illustrate the features of this novel biped robot. It has been shown that biped active walking may be achieved with only a few DoFs and simple kinematic configuration.

This paper presents control strategies for speed regulation on a running kneed biped with underactuation degree two. The robot under consideration has passive compliant knees and a unique hip actuator. In order to achieve variable speed running, a set of desired gaits (each at a different speed) is designed with the controlled Lagrangian method for the under-actuated robot. Then an event-based control law is applied to stabilize the running gait according to the running speed specification, and an energy shaping control is added to enlarge the domain of attraction as well as improve the convergence rate. At last a transitional control law is developed for smooth transitions between different running gaits. The effectiveness of our approach is verified by numerical simulations.

A proportional controller is compared to a nonlinear backstepping controller with four different grasps for a dexterous anthropomorphic hand. A bioinspired grasp-dependent control scheme which autonomously modulates the grip force using wrist velocity feedback to prevent grasped object slip is also introduced. Four different grasp types are evaluated to illustrate how the wrist velocity feedback architecture must differ depending upon the manner in which objects are grasped. The backstepping controller can successfully increase grip force with wrist velocity in a robustly stable bioinspired fashion. Experimental results show that the developed backstepping controller improves the position tracking abilities for multiple periodic inputs, as well as reduces step input overshoot. The slip prevention capabilities of the backstepping controller are also demonstrated and compared to the proportional control scheme. Results of the slip prevention experiments show that both the grasp type and manipulator orientation with respect to gravity are significant factors in the performance of the controllers. The backstepping control scheme significantly improves slip prevention of grasped objects for multiple grasps and in two different orientations with respect to gravity.

This paper presents a new biped mechanism, which has been designed with main features for low-cost and easy-operation. The biped mechanism is supposed to have the same basic motion capabilities as conventional multi-actuator biped robots, such as straight walking, changing direction, overcoming obstacle, descending and ascending stairs. Furthermore, with two wheels installed on the shank, the biped mechanism has two basic walking modes, which are legged mode and wheeled mode. Walking with proper mode for different environments makes the biped mechanism possess the advantages of both legged and wheeled robots.

While robotic hands have been developed for tasks such as manipulation and grasping, their potential as tools for evaluation of engineered products — particularly compliant structures that are not easily modeled — has not been broadly studied. In this research, a low-cost anthropometric robotic hand is introduced that is designed to characterize glove stiffness in a pressurized environment. The interaction with the compliant pressurized glove provides unique performance requirements and design constraints. The anthropometric robotic hand was designed to mimic the human hand in a configuration corresponding to the neutral position in zero gravity, including the transverse arch, longitudinal arch, and oblique flexion of the rays. The resulting robotic hand also allows for realistic donning and doffing of the prototype glove, its pressurization, and torque testing of individual joints. Solid modeling and 3D printing enabled the rapid design iterations necessary to work successfully with the compliant pressure garment. An instrumentation and data processing method was used to calculate the required actuator torque at each finger's knuckle joint. The performance of the robotic hand was experimentally demonstrated with a prototype spacesuit glove at different levels of pressure, followed by a statistical repeatability analysis. The reliable measurement method validated the pressure-induced stiffening. The resulting robotic design and testing method provide an objective and systematic way of evaluating the performance of compliant gloves.