基于PG-RRT算法的移动机器人路径规划

Path Planning of Mobile Robot Based on PG-RRT Algorithm

路径规划问题是移动机器人领域的重点问题,也是发展移动式机器人智能工厂的基础。快速扩展随机树算法(RRT算法)由于其良好的求解性,广泛应用于移动机器人路径规划。针对RRT算法面对复杂地图时随机采样效率低、路径重复性差的问题,提出一种基于模拟植物生长引导的RRT移动机器人路径规划算法(PG-RRT算法),提升了路径寻优的稳定性和效率。利用植物生长遵循的三大原则(向光性原则、遮挡物影响原则、负向地性原则),结合变步长技术、膨胀技术快速得到用于RRT算法采样的PG膨胀引导域,并得到最终路径。多组不同障碍物地图的仿真实验表明:相比于传统RRT算法和单一PG算法,PG-RRT算法减少了迭代次数,获得了更优的路径距离,而相比于A~*算法,该算法则大大缩短了计算时间。最后通过基于ROS系统机器人平台的实车测试,验证了PG-RRT算法的实用性。

The path planning problem is a vital problem in the field of mobile robots and is the basis for the development of smart factories.Rapidly-expanding random tree algorithm (RRT algorithm) is widely used in path planning because of its excellent solving performance.Aiming at the problem of low execution efficiency and poor path repeatability of the RRT algorithm when faced with complex maps,a plant growth guidance based RRT path planning algorithm (PG-RRT algorithm) for mobile robot was proposed to improve the stability and efficiency of path optimization.By using three principles followed by plant growth (Phototropism,Obstacle influence characteristics and Negative geotropism),combing variable step technique and inflation technique,the PG dilation guide field used for RRT algorithm can be obtained.Finally,the ideal path is obtained by using the RRT algorithm with random sampling characteristics.Abundant simulations show that the PG-RRT algorithm reduces the number of iterations and obtains better path distance compared to the traditional RRT algorithm and the single PG algorithm.It is noteworthy that the search efficiency of the presented path planning algorithm is improved compared with the A^* algorithm.Moreover,the actual vehicle test of robot verifies the practicability of the PG-RRT algorithm.