Path Planning for Mobile Robots Based on the Improved Adaptive Ant Colony Algorithm

This paper presents an improved adaptive ant colony algorithm in response to the drawbacks of the basic ant colony algorithm for robot path planning within grid environments, including low search efficiency, susceptibility to local optima, slow convergence speed, and non-smooth paths. To determine the next state dynamically, an adaptive factor is first introduced by using an adaptive pseudo-random state transition mechanism for path selection. The potential field gravity function and corner guiding factor are introduced to rewrite the original heuristic function. Second, the pheromone update strategy incorporates a kind of dynamic reward and punishment system. We define an adaptive evaporation factor that keeps the system from entering local optima. Redundant nodes in the path are then optimized using pruning. Eventually, a series of comparing and test experiments confirm the efficacy and efficiency of the proposed ant colony algorithm. Experiments show that the improved adaptive ant colony algorithm generates smoother optimal paths, accelerates convergence speed, and significantly improves search efficiency.