摘要
针对当前路径规划中存在的诸多问题,提出了基于遗传算法的机器人避开多随机障碍物的路径规划方法。首先提出障碍物环境的神经网络模型,并利用该模型建立机器人动态避碰路径与神经网络输出的关系,将需规划路径的二维编码简化成一维编码,并把动态避碰要求和最短路径要求以及边界约束条件融合成一个适应度函数。通过对该算法进行实验仿真,证明该方法具有良好的动态避障性能,是有效和正确的。
Aiming at the problems involved in current path planning,a method for mobile robot path planning to avoid multi-random obstacles based on GA is put forward.First,the nerve network model of the obstacles is established.Then making use of the model,the fitness function is used to meet the requirements of dynamic obstacle avoidance and the shortest route and boundary control,after that,the complex two dimension route coding problem is converted into one-dimension ones.Simulation results indicate that the method is correct and feasible.
出处
《西华大学学报（自然科学版）》
CAS
2007年第1期56-58,62,共4页
Journal of Xihua University:Natural Science Edition
