取各障碍物顶点连线的中点为路径点，相互连接各路径点，将机器人移动的起点和终点限制在各路径点上，利用Dijkstra算法来求网络图的最短路径，找到从起点P1到终点Pn的最短路径，由于上述算法使用了连接线中点的条件，不是整个规划空间的最优路径，然后利用遗传算法对找到的最短路径各个路径点Pi (i=1,2,…n)调整，让各路径点在相应障碍物端点连线上滑动，利用Pi= Pi1+ti×(Pi2-Pi1)（ti∈[0,1] i=1,2,…n）即可确定相应的Pi，即为新的路径点，连接此路径点为最优路径。-Take the midpoint of each obstacle vertex point for the path, the path of each connection point, move the robot start and end points of each path on the limited use Dijkstra algorithm to find the shortest path network diagram to find the starting point P1 from the Pn the shortest path to the end, because the algorithm uses the midpoint of the conditions of cable, not the entire optimal path planning of space, and then use genetic algorithm to find the shortest path of each path points Pi (i = 1,2, ... n ) adjusted to the path of obstacles in the corresponding endpoint connection point slide, using Pi = Pi1+ ti × (Pi2-Pi1) (ti ∈ [0,1] i = 1,2, ... n) to determine the appropriate of Pi, is the new path point, the connection point for the optimal path for this path.