基于神经网络和人工势场的滚动规划

Rolling Planning Based on Neural Networks and Artificial Potential Field

针对多障碍动态环境,论文提出了一种新的路径安全评价方法,并在此基础上提出了基于神经网络和人工势场法的滚动规划方法。与以往的全局规划相比,减少了全局中无关信息干扰,增加了规划效率;与以往的滚动规划相比,增加了移动机器人的预测能力,减少了碰撞的可能性。仿真实验表明,这种规划方法能有效地解决多障碍动态变化的时变环境中的路径规划问题。

This paper brings out a novel method to estimate path safety in dynamic environment with multiple obstacles, and then puts forward a rolling planning method based on neural networks and artificial potential field.Compared to previous global path planning,this planning method decreases the disturbance of irrelative information and increases the planning efficiency;compared to previous rolling planning method,it increases the predicting ability of mobile robot and decreases the probability of collision.The simulation results show that the planning method proposed is effective and feasible.