基于混合蛙跳算法优化人工势场的路径规划方法研究

Study on path planning algorithm based on artificial potential field optimized by shuffled frog leaping algorithm

针对传统人工势场在路径规划时易陷入局部极小、目标不可达和狭窄通道左右摆动的问题,提出了一种改进人工势场算法。该改进算法将目标距离引入斥力函数解决目标不可达问题:将斥力分解为由障碍指向机器人的分量和垂直于机器人与障碍连线的分量,分别用于阻止机器人接近障碍和引导机器人离开障碍。针对蛙跳算法迭代速度慢的问题,提出了改进混合蛙跳算法来提高蛙跳算法的进化速度和精度,并用于优化人工势场的参数以提高路径规划性能。仿真试验验证了所提出算法的有效性。

Abstract： To solve the problems of goal nonreachable with obstacle nearby （GNRON） , local minimum and swinging in narrow roads in the use of traditional artificial potential field, a modified artificial potential field al- gorithm was proposed. In this modified algorithm, the target distance was introduced into the repulsion function to solve the problem of GNRON, and the repulsive force was decomposed into a component force from obsta- cles to the robot and a vertical component force, which were used to prevent the robot close to obstacles and guide the robot to leave obstacles respectively. To improve the iterative speed and accuracy of the shuffled flog leaping algorithm, the modified shuffled flog leaping algorithm was proposed to optimize the parameters of arti- ficial potential field to effectively improve the path planning performance of the artificial potential field algo- rithm. The effectiveness of the proposed algorithm was verified by the simulation experiments.