论文部分内容阅读
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。
Obstacle avoidance control has always been a challenge for mobile robot path planning. A dynamic obstacle avoidance method based on neural network in unknown environment is proposed. At the same time, the hybrid force / position control structure is applied to obstacle avoidance control of mobile robot. Force control algorithm is through the formation of a virtual force field between the mobile robot and the obstacle, and its setting, so that they can maintain the desired distance between the two. Because the dynamics model of mobile robot and the uncertainty of obstacles also affect the performance of obstacle avoidance control, Elman neural network is adopted to compensate for the uncertainty and to set the exact distance between mobile robot and obstacle. Simulation experiments show that the dynamic obstacle avoidance algorithm is effective.