Robot dynamic path planning based on improved ant colony and DWA algorithm

Path planning technology is an important branch in the field of mobile robot research, which enables the robot to find a relatively optimal path safely and quickly in the multi-obstacles environment. Aiming at the blind search of the ant colony algorithm, easy to fall into local optimization and slow convergence speed in global path planning, and the problem that the dynamic window approad (DWA) is difficult to effectively avoid dynamic obstacles in local path planning, a fusion algorithm of the improved ant colony algorithm and the DWA is proposed. Firstly, the GRRT-Connect algorithm is proposed to allocate initial pheromones unequally to solve the local optimization problem in trap maps. Secondly, the ant colony relay search method is added to solve the self deadlock problem of an ant tabu list, and the slice optimization method is used to optimize the optimal path selection mechanism to obtain the global optimal path. Then, the DWA is run with the key points of the optimal path as the sub-target points, and an adaptive speed adjustment method is proposed for optimal driving. Finally, a pre-calculation method is proposed to avoid dynamic obstacles and achieve the effect of local planning. The simulation results show that compared with the results in the existing literature, the optimal path length of the fusion algorithm is shortened by 10.28 % and the convergence speed is accelerated by 6.55 %. © 2022 Northeast University. All rights reserved.