Fumi Seto

REAL-TIME SELF-COLLISION AVOIDANCE SYSTEM FOR ROBOTS USING ROBE

In this paper, we propose a real-time self-collision avoidance system for robots which cooperate with a human/humans. First, the robot is represented by elastic elements. The representation method is referred to as RoBE (Representation of Body by Elastic elements). Elastic balls and cylinders are used as the elements to simplify collision detection, although elements of any shape could be used for RoBE. When two elements collide with each other, a reaction force is generated between them, and self-collision avoidance motion is generated by the reaction force. Experiments using the mobile robot with dual manipulators, referred to as MR Helper, illustrate the validity of the proposed system.

Evaluating human motor function of lower limbs in periodic motion during pedaling exercise

The purpose of this study is to establish a quantitative method that can evaluate the motor function of human lower limbs in dynamic situations. As an index of motor function, we focus on the mechanical impedance characteristics of the lower limbs. By controlling these characteristics, humans can improve motion stability and adapt their movement to disturbances in the environment. In this paper, we propose a method that can estimate the impedance characteristics of the lower limbs of a human subject applying torque perturbations during a pedaling exercise. To improve the accuracy of the proposed estimation method, the effects of inertia and gravity in the measurement data were considered by representing the lower limbs by a multi-link model. Experiments were performed in two participants to estimate the stiffness and viscosity at different phases of the periodic movement. The distributions of these parameters showed symmetric properties for both legs in a pedaling cycle-a tendency that was observed in both participants. This agreement suggests that our method could extract a component of human motor control strategy that adjust mechanical impedances during pedaling exercises.