摘要
避障控制一直是移动机器人路径规划的难点。提出了一种未知环境下基于神经网络的机器人动态避障方法,同时把混合力/位置控制结构应用到移动机器人的避障控制中。力控制算法是通过在移动机器人和障碍物之间形成虚拟力场,并对其整定,以使它们两者之间能保持期望距离。由于移动机器人的动力学模型和障碍物的不确定性也会对避障控制的性能造成影响,因此采用Elman神经网络来补偿不确定性,同时整定移动机器人和障碍物之间的精确距离。仿真实验表明该动态避障算法是有效的。
Collision avoidance is always difficult in path planning of mobile robot. A dynamic environment of robots based on neural network method of dynamic obstacle avoidance, is presented while the intelligent hybrid force/position control technology is/tpplied to mobile robot obstacle avoidance control areas. Through the force control algorithm is formed between the mobile robot and obstacles vir tual force field, and its setting, so that they can maintain the hope distance between the two. However, in the simulation process, the uncertainty of the mobile robot dynamic model and the obstacles will have impact on the performance of obstacle avoidance. Therefore, Elman neural network tocompensate for the uncertainty caused by the environment, is used while ajusting the exact distance between the mobile robots and the obstacles. Simulation results show that the dynamic obstacle avoidance algorithm is effective.
出处
《控制工程》
CSCD
北大核心
2013年第2期280-285,共6页
Control Engineering of China
基金
河北省科技攻关项目(07213526)
燕山大学博士基金(B168)资助的课题
