Kousuke Sekiyama

Self-organizing control of urban traffic signal network

This paper proposes a self-organizing control strategy of a signal network for urban traffic flows. The signal network is described using a system of nonlinear oscillators with the nearest neighborhood coupling. The natural frequency, split and coupling constants of the signals are adjusted dynamically according to local information of traffic flows so that the desired network offset patterns are self-organized through mutual entrainment. We provide a preliminary analysis of the coupled oscillator system. Numerical simulations demonstrate the effectiveness of the proposed method under dynamical environmental changes.

Group behavior control for MARS (micro autonomous robotic system)

We present the design of our small robot MARS (micro autonomous robotic system) and basic experiments of group behavior pattern. MARS is small mobile autonomous robot. The robot has sensors, battery and infrared light communication device. In addition, the robot can get programs from the host computer (programmable). We focus on group behavior design (cooperation, competition coordination etc.) to use real robot MARS and we try to realize the phase change of the group behavior pattern. One of our ultimate dream is designing life-like robotics systems, such as ant or fish societies.

Distributed control of flexible transfer system (FTS) using learning automata

This paper proposes a flexible transfer system (FTS) as one of the self-organizing manufacturing systems. The FTS is composed of autonomous robotic modules, which transfer a palette carrying an object. Through the self-organization of a multilayer strategic vector field corresponding to a task, the FTS can generate quasi-optimal transfer path in fully distributed way. We apply the learning automata for path generation algorithm. Simulation is conducted to evaluate the basic performance of the system and the results show the feasibility of application. Also, the developed hardware is explained.

Efficient communication method in the cellular robotic system

Presents a new approach to communication in an autonomous distributed system, focusing on the cellular robotic system (CEBOT). The CEBOT is a self-organizing robotic system composed of a large number of robotic units (cells) with a simple function. In order for cellular robots to implement a task efficiently, cooperation and coordination among cells are of great importance and require efficient communication. Since communication takes place in the process of task execution, it is likely to be intensive and biased by the small number of cells related to the task. This could decrease communication efficiency. To deal with these problems of robot communication, the authors propose demand-based intensive communication (DBIC), and show its effectiveness with simulation and experimental results for the prototype CEBOT Mark IV.

Robust velocity sliding mode control of mobile wheeled inverted pendulum systems

There has been an increasing interest in a kind of underactuated mechanical systems, mobile wheeled inverted pendulum (MWIP) models, which are widely used in the field of autonomous robotics and intelligent vehicles. Robust velocity tracking problem of MWIP systems is investigated in this study. In the velocity control problem, model uncertainties accompany uncertain equilibriums, which make the controller design become more difficult. A sliding mode control (SMC) method based on a novel sliding surface is proposed for the systems, which are capable of handling both parameter uncertainties and external disturbances. By assuming the specially designed sliding surface, the proposed SMC controller is capable of eliminating the steady velocity tracking error. The asymptotical stability of the closed-loop system is achieved through selecting sliding surface parameters in terms of some rules. The effectiveness of the proposed methods is finally confirmed by numerical simulations.

Distributed Intelligent Systems in Cellular Robotics

Recently, as the robotic systems are being improved, these systems are expected to be applied in more global fields. In addition to the conventional research for the improvement of the function of individual robot, distributed robotic systems are being researched. The distributed robotic systems consist of a number of autonomous robots. In the systems, some tasks are carried out through the cooperation, coordination and competition among the robots. The distributed robotic systems will work more efficiently than the sum of the ability of each. Therefore, the distributed robotic systems are expected as an advanced robotic technology. Following advantages can be considered such as flexibility, robustness, fault tolerance, and extendibility of the system. On the other hand, to realize the distributed robotic systems, there are many subjects to research. In recent years, the research works on the distributed robotic systems have been carried out actively and attractively, which include the research on cooperative/coordinate mobile robots [1][2], multi-arm manipulator robots [3] [4], several kinds of robots [5] [6], and group/swarm robotic systems [7] [8]. Related research works also include distributed artificial intelligence [9] [10], distributed computer systems[11], artificial life[12] and so on. The research on distributed artificial intelligence will indicate any ideas to the research on the intelligence of the distributed robotic systems. The distributed computer systems will show the control and communication architecture for the distributed robotic systems. And the research on the artificial life will present how to organize the distributed robotic systems, since the systems as seen in the society of ants or human beings will be considered as one of the distributed autonomous systems. Therefore, the distributed robotic systems will be developed by the coordinate research from viewpoint of the biological fields.

Toward Social Robotics

This paper describes conception of social robot system as self-organizing system. Distributed autonomous robot system is not autonomous as a group, if the system totally depends on the external intervention to maintain its fundamental function. We discuss the meaning of autonomy for group robot system and a concept of collective autonomy as a fundamental framework for design of the social robotic system. It is firstly discussed the difference between cooperative and self-organizing behavior. After brief review of the classical self-organizing system theory, we present the self-referential coupling of serf-organizing systems as a design framework of the collective autonomous system. Through primitive simulation model, we also discuss relation between redundancy and optimality in the system structure.

Modeling and control of a novel narrow vehicle

There is a growing interest on narrow vehicles which can partially solve the parking, congestion and pollution issues associated with urban transportation. A novel narrow vehicle, UW-Car, which is based on a mobile wheeled inverted pendulum (MWIP) platform and a movable seat is proposed in the paper. The three dimensional dynamic model of the underactuated system running on the flat ground is derived by using Lagranges motion equation. Because of the strong nonlinearity of the UW-Car system, a sliding mode control (SMC) scheme is chosen to ensure the state variables converge to desired values. The velocity control of UW-Car is discussed in this paper. By using the proposed controller, a UW-Car can move at a desired velocity and tracking a given yaw angle velocity while keeping the seat to be always upright. Numerical simulations are provided to verify and illustrate the effectiveness of proposed approaches.

Multiple object detection for intelligent robot vision by using growing neural gas

Recently, various types of robots have been researched and developed for supporting our life. Also, the perceptual system for the robot is researched. Visual perception includes a lot of valuable information and it is useful for all intelligent robot system. In this paper, we discuss intelligent robot vision in order to detect multiple object and human. There are many visual sensors and we use the range-imaging camera to detect distance and image data. We propose a method for perceive moving target by using growing neural gas and Genetic algorithm. In the experimental results, we show the potency of our method.

Locomotion Stabilization with Transition between Biped and Quadruped Walk based on Recognition of Slope

The applicative field of activities of robots who have only one locomotion strategy is limited. As means of enhancing the mobile range, it is necessary to have various locomotion modes. Therefore, we focus on dynamic transitions between several kinds of locomotion modes adapting to environmental changes. In this paper, we aim to realize stable locomotion along some unknown test courses consisting of flat and slope with transition between biped and quadruped walks. To achieve this transition, we propose a method to recognize a slope, a design of the transition motion, and a technique for modifying the trajectories of quadruped walk to move between flat and slope. By experiments, we verify that the proposed methods can successfully be applied to both upslope and downslope cases.