融合间隙与DWA的移动机器人动态避障规划

Dynamic obstacle avoidance planning for mobile robots incorporating gap and DWA

针对传统动态窗口法(DWA)前瞻性不足以及动态环境下避障能力不足的问题,提出一种融合间隙与障碍物运动信息的改进算法。首先,利用间隙分析方法提取目标间隙,并计算来自间隙的目标方向,提高算法的环境感知能力;其次,采用几何方法,在重定向阶段增加碰撞预测,实现对移动障碍物的快速反应;然后,为进一步优化运动轨迹,设计动态虚拟目标点,以距离作为依据,通过评价因子的替换,解决转角处的不合理偏转;最后,通过Gazebo仿真平台进行实验验证。结果表明,改进后的算法可以解决原算法通过引入全局路径规划仍然无法处理的前瞻性问题,在与障碍物保持合理间距的同时,能够稳定地通过稠密区域,在执行时间和路径长度上分别缩短23.7%和25.4%,并有效提高了对动态环境的适应能力。

To address the problems of insufficient foresight of the traditional dynamic window method(DWA)and insufficient obstacle avoidance capability in dynamic environments,an improved algorithm that fuses gap and obstacle motion information is proposed.Firstly,the gap analysis method is used to extract the target gap and calculate the target direction from the gap to improve the environment perception capability of the algorithm.Secondly,the geometric method is used to add collision prediction in the redirection phase to achieve fast response to moving obstacles.Then,to further optimize the motion trajectory,the dynamic virtual target point is designed to solve the corner with the distance as the basis and the replacement of evaluation factor to solve the unreasonable deflection.Finally,the experimental validation is carried out by Gazebo simulation platform,and the results show that the improved algorithm can solve the forward-looking problem that the original algorithm still cannot handle by introducing global path planning,and can pass through the dense area stably while maintaining reasonable spacing with obstacles,shortening 23.7%and 25.4%in execution time and path length respectively,and effectively improving the dynamic environment adaptation capability.