Improvement of Robot Sensor by Integrating Information Using Neural Network

摘要

Recently, the improvement of robotics makes us feel the robot to be more familiar to human activities. For example, robot's motion similar to human movement is also developed. However, in order for a robot to play an active part in the medical and welfare field, the sensors of the robot will be important to have effective functions for robot's safety actions. Our approach is applying intelligent processing technology for the sensors by using neural networks. Our purpose is making recognition system with several sensors. We developed a sensory integrating system by the neural network and apply the system to the robotics. This robot can track the target by applying integrated neural system. We extend the robot system to avoid obstacles, which is based on a subsumption architecture.