Narrow Results

This paper gives an overview of the current and forthcoming research projects of the Collaborative Research Center 588 "Humanoid Robots — Learning and Cooperating Multimodal Robots." The activities can be divided into several areas: development of mechatronic components and construction of a demonstrator system, perception of user and environment, modeling and simulation of robots, environment and user, and finally cooperation and learning. The research activities in each of these areas are described in detail. Finally, we give an insight into the application scenario of our robot system, i.e. the training setup and the experimental setup "household."

This paper presents a recognition method for human actions in daily life. The system deals with actions related to regular human activity such as walking or lying down. The main features of the proposed method are: (i) simultaneous recognition, (ii) expressing lack of clarity in human recognition, (iii) defining similarities between two motions by utilizing kernel functions derived from expressions of actions based on human knowledge, (iv) robust learning capability based on support vector machine. Comparison with neural networks optimized by a back propagation algorithm and decision trees generated by C4.5 proves that the accuracy of recognition in the proposed method is superior to others. Recognizing actions in daily life robustly is expected to ensure smooth communication between humans and robots and to enhance support functionality in intelligent systems.

We propose a hand-eye coordination system for a humanoid robot that supports bimanual reaching. The system combines endpoint closed-loop and open-loop visual servo control. The closed-loop component moves the eyes, head, arms, and torso, based on the position of the target and the robot's hands, as seen by the robot's head-mounted cameras. The open-loop component uses a motor-motor mapping that is learnt online to support movement when visual cues are not available.

This paper presents three basic bipedal walking gait adjustment modes: step-frequency, step-length and biped lower extremities' pattern adjustments. All the adjustment modes are based on a simple Fourier Series formulation named the Truncated Fourier Series (TFS) model which is newly proposed as a walking pattern generator. Making use of these three gait adjustment modes, bipedal walking can be modified in real-time according to the environment changes. In this paper, the developed gait adjustment modes have been studied by dynamic simulations. The results obtained show that stable walking on uneven terrains as well as human-like walking behaviors can be achieved.

Teachability has been extensively studied under the context of making industrial robots to be programmable and reprogrammable. However, it is only recently that the artificial intelligence (AI) research community is accelerating the research works with the objective of making humanoid robots and many other robots to be teachable under the context of using natural languages. We human beings spend many years learning knowledge and skills despite our extraordinary mental capabilities of being teachable with the use of natural languages. Therefore, if we would like to develop human-like robots such as humanoid robots, it is inevitable for us to face the issue of making future humanoid robots teachable with the use of natural languages as well. In this paper, we present the key details of a top-down design for achieving a teachable mind which consists of two major processes: the first one is the process that enables humanoid robots to gain innate mental capabilities of transforming incoming signals into meaningful crisp data, and the second one is the process which enables humanoid robots to gain innate mental capabilities of undertaking incremental and deep learning with the main focus of associating conceptual labels in a natural language to meaningful crisp data. These two processes consist of the two necessary and sufficient conditions for future humanoid robots to be teachable with the use of natural languages. In addition, this paper outlines a very likely new finding underlying the human brain’s neural systems as well as the obvious mathematics underlying artificial deep neural networks. These outlines provide us with a strong reason to separate the study of the mind from the study of the brain. Hopefully, the content discussed in this paper will help the AI research community to venture into the right direction which is to make future humanoid robots, non-humanoid robots, and many other systems to achieve human-like self-intelligence at the cognitive level with the use of natural languages.

The paper discusses the relationship between learning, innovation and (institutional) reflexivity. It is often held that reflexivity is a crucial factor for learning and innovation processes. However, a rather formalistic approach to reflexivity is predominant. We propose to overcome this limitation and to develop a more meaningful concept of reflexivity which "reflects" the contingent, relational, dynamic and complex character of organizational environments and reality. Based on this broadened understanding it appears that reflexivity is imminently a dialectic category and, under specific circumstances, it can also inhibit innovation. This is especially the case when reflexive tools are abused to push performance only. In order to illustrate our concept and hypotheses we added two case studies which highlight the conflicting counterparts of reflexivity and innovation and pointed us to important cultural "success factors".

In this paper, we analyze the technology adoption problem of firms when relevant information about a new technology is dispersed among them. Developing a continuous time model in which imperfectly and differently informed multiple firms determine strategically when to adopt a new technology, we show that the phenomena of an economically inefficient initial delay of adoption, a staggered adoption, and an inefficiently early mass adoption can arise in equilibrium, particularly in the form of strategic delay, informational learning, and herding behavior, respectively. We also address the incentive scheme that helps to achieve efficient collective adoption of the new technology under dispersed information and show under what conditions, if any, such an incentive scheme can work well.

Knowledge and learning mechanisms are investigated in a multiple and inter-related context depicted by the definition of an innovation ecosystem [Lusch and Nambisan (2015)]; by questioning knowledge, the paper opens up practice-based learning studies [Gherardi (2000)]. Action research offered the opportunity to focus on knowledge and learning mechanisms in an emerging innovation ecosystem linked to a project supported by the Italian Ministry of Research. The paper identifies three different knowledge practices — connecting knowledge dots, integrating knowledge, and authoring and disseminating knowledge — to describe learning and knowledge mechanisms in a networking innovation context, leading to an innovation ecosystem.

This work discusses the generally accepted idea that public–private cooperation only generates unilateral transfers of knowledge from public research and development (R&D) institutions to industrial firms and suggests that there are collaborative patterns in which the generation of knowledge is joint and it is characterized by the presence of bidirectional flows of knowledge. Through a multi-case study, this aspect is investigated in three public–private partnerships in the Argentine biopharmaceutical sector, focusing in particular on the knowledge flows and the identification of possible learning processes that the public part develops when cooperating with firms. The main findings and contributions of the work are that knowledge in the hands of the industry flows towards the public part, strengthening its R&D capabilities, that all the benefits received by the public part can be understood in terms of learning and that science–industry cooperation is compatible with the enrichment of the public research agenda.

This paper introduces a real-time video surveillance system which can track people and detect human abnormal behaviors. In the blob detection part, an optical flow algorithm for crowd environment is studied experimentally and a comparison study with respect to traditional subtraction approach is carried out. The different approaches in segmentation and tracking enable the system to track persons when they change movement unpredictably in occlusion. We developed two methods for the human abnormal behavior analysis. The first one employs Principal Component Analysis for feature selection and Support Vector Machine for classification of human behaviors. The proposed feature selection method is based on the border information of four consecutive blobs. The second approach computes optical flow to obtain the velocity of each pixel for determining whether a human behavior is normal or not. Both algorithms are successfully developed in crowded environments to detect the following human abnormal behaviors: (1) Running people in a crowded environment; (2) falling down movement while most are walking or standing; (3) a person carrying an abnormal bar in a square; (4) a person waving hand in the crowd. Experimental results demonstrate these two methods are robust in detecting human abnormal behaviors.

This paper sets out to analyze social entrepreneurship in the Central America Learning Alliance, in the context of recent literature on entrepreneurship and learning. Drawing on a recent and rapidly growing literature that describes entrepreneurship as a process that is inherently dynamic and experimental, with learning as a core component, we focus on social entrepreneurship in development as a catalyst of social transformation. A case study of a multi-stakeholder network focused on promoting processes of rural enterprise development, known as the Central America Learning Alliance, is used to illustrate social entrepreneurship in the context of the framework for innovation and learning that is developed in the first part of the paper. We conclude that the key concepts underlying entrepreneurial learning have important implications for social entrepreneurship in the context of building dynamic livelihoods for the poor of Central America.

This paper will report the results of empirical, web-based research on the knowledge processes that take place in a virtual environment which has been created by the convergence between the telecommunications and information technologies. The analysis of some virtual knowledge networks (VKNs) is presented; the 34 analyses focus on the network properties of both nodes and links. These properties give information on a new general model of VKNs that describes the multilevel structure of virtual networks. An interpretation of VKNs is proposed, which makes use of the complexity theory.

In today's economy, the key ingredients in success and survival are adaptability and the capacity to learn and change. Recent progress in the theory of complex systems provides a new basis for our understanding of how this may actually occur, and the factors on which it depends. Complex systems thinking shows what assumptions underlie the reduction of some part of reality to a mechanical model. They demonstrate that the simplicity and "knowledge" derived from such representations can lead to an understanding that entirely misses the most important, strategic changes that may occur. Complex systems models reveal the key processes that underlie "learning", and recognise the limits to knowledge and the inherent reality of uncertainty. They demonstrate the fundamental importance of internal, microdiversity within systems, as the source of exploration that drives learning. These ideas are explained and presented in a simple model of emergent co-evolution, where the exploration of internal diversity leads to the formation of a complex, with synergetic attributes. The paper describes and models briefly the uncertainties inherent in the definition and development of a new product or service. A further model involving complex products is briefly described which shows the importance of "search" in "knowledge generation" for the success of adaptive industrial networks and clusters. All this leads to the statement of a "law of excess diversity" which states that the long-term survival of a system requires more internal diversity than appears requisite at any time.

This article analyses key characteristics pertaining to the evolution of nascent technologies and enterprises. More specifically, we analyse the development of two new renewable energy technologies — solar photovoltaic (PV) and wave power (WP) — within the Norwegian energy sector. There are mainly two reasons for selecting these technologies. First, they may facilitate the transition to more sustainable energy supplies. Second, they display strikingly different development dynamics, from which important lessons can be learned for the management of innovation in the energy sector. The main causes of the Norwegian PV industry's success are its ability to exploit a global niche market, and the effective matching of technological capabilities with market opportunities into strategies for learning and legitimacy. The constrained development of WP technology, on the other hand, is not only attributed to technical difficulties, but also to the lack of niche markets, research rivalry, decreasing public support, and insufficient organisational capacity.

In times of increasing use of project-based structures, the capability of managing and organising projects becomes critical for competition. Previous research has documented the problems and possibilities of cross-project learning and various mechanisms that organisations can use to stimulate and facilitate learning. Moreover, research on project competence and project capabilities has positioned these capabilities within a knowledge-based theory of the firm. This paper tries to integrate these streams of research and attempts to broaden our current conceptual frameworks of how firms develop project competence. Based on an exploratory multiple-case study of six firms, it is suggested that a more fine-grained analysis of competence dynamics is required. We identify three different learning processes that contribute to the competence dynamics operating in project-based organisations. The first one labelled "shifting" revolves around the major shifts in the project operations of the firm. It is suggested that such major shifts play an important role in laying the foundation and rejuvenating the challenges of project organising. The second learning process identified, labelled "adapting", focuses on the continuous learning that takes place within project operations of the firm, between project generation, project organising, project leadership and project teamwork. The third and final learning process — "leveraging" — emphasises the role of knowledge transfer across projects; across similar projects, across different types of projects. It is suggested that empirical research into competence dynamics in project-based organisations should consider all three types of learning processes and further develop our understanding how these processes are linked to each other.

What can cause five postponements and a delay of two years in introducing a relatively simple ERP system that usually takes only a few months to be implemented? We find the answer to this puzzle by highlighting the context of use of this high technology IT capital good, an issue so far overlooked because the literature on complex product system (CoPS) focuses on the intrinsic dimensions of the product and the provider. We rely on an extensive case study of the ERP system Movex at the furniture manufacturer Edsbyn and on literature on user-related innovations, organisational studies and inter-firm relationships to extract a series of additional user-related complexity dimensions. These include the importance of the capital good for the user, the user's perception of its complexity and the strength and complexity of the routines to be changed at the using organisation. We conclude the paper with implications for complex systems providers.

This paper explores the kind of learning that creates capabilities needed in incremental and radical innovation development. The empirical evidence is based on four innovation-related development projects, implemented through enterprise-student team collaboration. It seems that the richness of innovation and learning processes affects the diverseness of developed capabilities. The goal of innovation development drives learning and capability building, while improved capabilities help adopt challenging goals that stimulate a new level of learning. The process emphasising analysis, planning and analytical knowledge creates capabilities resulting in incremental inventions. The innovation process that focuses on experimentation and utilisation of other knowledge types facilitates capability building resulting in inventions that are radical in nature. These two process types are not contrary but complementary options to develop solutions to satisfy the expressed and unexpressed needs of customers. This paper presents views on how to interconnect these processes characterised by divergent rhythms.

Knowledge-based competition is leading to collaboration with partners and even competitors as firms pursue appropriate knowledge for innovation which has become a strategic imperative. Inbound open innovation helps increase the innovativeness of the firm by monitoring the operating environment and enabling it to source knowledge from collaborative partners. On the basis of in-depth interviews with senior managers and the knowledge-based view of the firm, this study examines the extent to which inbound open innovation activities contribute to collaborative innovation. Then, using a sample of 224 surveys representative of a cross-section of medium to large firms involved in collaborative ventures, the theoretical model is empirically examined. The results show that collaborative creativity, learning and knowledge stock are critical core inputs of collaborative innovation, with the support of formal coordination mechanisms and internal search processes, such as structural centralisation, formality and absorptive capacity.

This paper investigates the ability of knowledge intensive business firms (KIBS) to engage in co-innovation with client firms. Co-innovation is related to the competitive advantage of KIBS as knowledge creators and sources of innovation. We apply a knowledge-based perspective where knowledge-related resources and learning capabilities explain why certain KIBS firms are able to co-innovate. We couple our theoretical expectations with qualitative evidence on three best practices in the Dutch market for environmental services.

This study conducts an empirical analysis on the relationship between innovation and the type of partner based on the assumption that the knowledge and information acquired from partners would vary depending on their type from the perspective of learning through technology cooperation. It further expands the discussion by looking at the relationship between geographic distance between partners and innovation as well as absorptive capacity, a variable that moderates it. The knowledge required for product development is classified into explicit and implicit knowledge, and based on such knowledge type, the form of learning and innovation is categorized into STI (Science, Technology and Innovation) and DUI (Doing, Using and Interacting). Accordingly, technology cooperation partners are divided into STI and DUI partners. The study analyzes the effect of the cooperation partner type on radical and incremental innovation. Unlike the hypothesis, cooperation with a STI partner had a positive effect on incremental innovation while a DUI partner had such effect on radical innovation. The geographical distance between partners had a negative effect on incremental innovation and the moderating effect of appropriability was not verified.