当移动机器人在行进过程中使用传统人工势场法(artificial potential field method,APF)进行路径规划时,通常会陷入局部最优困境,无法顺利到达目标点.为解决这一问题,首先,对APF算法规划路径失败原因进行分析,其次设置情况判断条件,判断当机器人陷入局部最小值时,通过在合适位置增加临时引导点的方法,引导其跳出局部极小值点;其次,引入分数阶微积分思想方法结合APF算法,提出混合阶次的分数阶梯度下降法进行位置信息迭代,优化算法的收敛速度和收敛精度;最后,用MATLAB软件对该方法进行仿真,实验结果表明提出的方法可以有效解决局部最小值问题,而且在速度、精度上都有明显的提高,且能适应较为复杂的多障碍物环境,验证了改进方法的有效性、正确性.
Improving the APF to Solve Local Minima Research on Path Planning
When mobile robots use the traditional APF for path planning during their journey,they often fall into a local optimal dilemma and cannot reach the target point smoothly.To solve this problem,first an-alyze the reasons for the failure of the APF to plan the path,and then set the condition judgment condition to determine that when the robot falls into the local minimum value.By adding temporary guidance points at appropriate position,it is guided to jump out of the local minimum point.Then bring in fractional calcu-lus thought method combined with the APF,propose mixed order fractional gradient descent method for po-sition information iteration,and optimize the rate of convergence and convergence accuracy of the algo-rithm.Finally,the method was simulated using MATLAB software,and the experimental results showed that the proposed method can effectively solve the local minimum problem,and has significant improve-ments in speed and accuracy.It can also adapt to complex multi obstacle environments,verifying the effec-tiveness and correctness of the improved method.
NETL
NSTL
万方数据
