仿生六足机器人传感信息处理及全方向运动控制  

Sensor signal processing and omnidirectional locomotion control of a bio-inspired hexapod robot

作　　者：陈伟海[1] 刘涛[1] 王建华[1] 任冠佼[1] 吴星明[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100191

出　　处：《浙江大学学报（工学版）》2015年第3期430-438,共9页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(61175108);北京市自然科学基金资助项目(4142033)

摘　　要：针对传统CPG控制方法无法很好地控制机器人足端轨迹的问题,提出基于中枢神经模式发生器(CPG)的控制方法.在CPG控制方法中加入用于足端轨迹控制的轨迹发生器模块,通过调节参数值可以实现机器人的全方向运动控制.为降低传统CPG控制中传感信息反馈以及参数调整的复杂程度,设计2种用于躲避前方和两侧障碍物的传感信号处理的神经网络.该模块实现多传感信号的融合,生成用以控制机器人运动行为的各个参数值,实现机器人的自主避障.设计一个仿生六足机器人样机,将其分别放置在墙角和狭窄空间中进行自主避障行走实验,结果证明了机器人全方向运动控制算法和自主避障算法的可行性.Control method based on a central pattern generator（CPG）was proposed to deal with the issue of intelligent locomotion control of bio-inspired hexapod robots.The traditional CPG-based control was unable to control the foot trajectory of robots.For the proposed CPG control method,trajectory generators were added to the control scheme to control foot trajectory.Thus,omnidirectional walking control could be realized by simply adjusting the value of control parameters.Two artificial neural networks for sensor signal processing were developed to overcome complexity in sensory feedback and parameter tuning in CPG control.The neural networks accomplished the fusion of multi-sensory signals,generating the values of the control parameters for robot behavior control.In this way,the robot realized autonomous obstacle-avoiding.A hexapod robot prototype was designed to conduct two real robot experiments in different situations.The experimental results proved that the effectiveness of proposed omnidirectional locomotion control algorithm and obstacle-avoiding algorithm.

关 键 词：仿生六足机器人 运动控制 传感信息 中枢神经模式发生器 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]