Path Planning for Robots Based on Adaptive Dual-Layer Ant Colony Optimization Algorithm and Adaptive Dynamic Window Approach

Abstract

To address the limitations of ant colony optimization (ACO) algorithm in terms of convergence speed, search efficiency, local optimal traps, and dependence on high-precision maps, this article proposes adaptive dual layer ACO (ADL-ACO) algorithm and adaptive dynamic window approach (ADWA) for dynamic path planning of robots. The ADL-ACO has a dual-layer structure, which is divided into a path planning layer and a trajectory optimization layer, where the adaptive elite ACO (AEACO) generates collision-free initial paths and the trajectory optimization algorithm (TOA) further optimizes the initial paths. The first layer is the AEACO algorithm for the path planning layer, which accelerates the convergence speed and enhances the global search capability through adaptive parameter tuning and pseudorandom state transition rules. The second layer is the TOA for the trajectory planning layer, which optimizes the initial path in terms of length, number of turns, safety, and smoothness, and utilizes the segmented B-spline technique to enhance the smoothness of the paths. Moreover, the ADWA is proposed for dynamic obstacle avoidance in dynamic environments to enhance the adaptability of the algorithm in complex environments. The simulation results indicate that ADL-ACO reduces the length of optimal path, average path length, execution time, optimal path turning point, and smoothness in comparison to other algorithms, and ADWA improves the robot’s obstacle avoidance efficiency and safety. The experimental trials in real-world indoor and outdoor conditions validate the algorithm’s efficacy in this study. The method presented in this research provides a novel way to address the path planning for mobile robots.