英文论文范例机器人分布式控制系统概述.docx

Distributed Control System Design for Humanoid RobotsAbstract—A humanoid robot with multiple sensors needs a control system with severe requests like real-time control and friendly man-computer interaction. In this paper, we propose a distributed control system consisting of double PC control sub-system based on CAN bus for humanoid robot. The control system is divided into three layers. As the main control layer, double PC control architecture is formed with two sub-systems. Man-machine interface sub-system under Windows OS aims at controller debugging and information observation, meanwhile real-time motion control sub-system under RT-Linux OS is developed for real-time motion control realization. The communication layer based on CAN bus assures the reliable communication between real-time computer and joint controllers. The actuating layer is responsible for joint servo control and sensory data acquisition. At last, the experiment is implemented on newly-build humanoid robot MIH-2. Joint motor testing under Windows OS confirms the effectiveness of humanoid robot debugging and accuracy of data acquisition. Real-time control system testing based on RT-Linux OS demonstrates that the system can provide stationary time-lag ensuring the stability during the walking pattern control for humanoid robot. The experiment results show that the presented control system can meet all the requests during the humanoid robot walking control.Ⅰ. INTRODUCTION With the development and maturity of robot technology, humanoid robot with multi-sensors has gradually become another hotspot in this field [1]. Generally, a humanoid robot has more than thirty DOFs (degree of freedom) to be servo-actuated and should deal with multiple sensor information, which requires reliable and stable control system with high computing speed. Great breakthrough such as well-known humanoid robots P3 [2], ASIMO and HRPs has been achieved thanks to the efforts of many institutes and researchers [3]. Many of the