An Adaptive Local Trajectory Planning Method for DWA Based on the ACT-R Cognitive Mechanism

To address the issues of fixed evaluation weights and the lack of historical experience utilization in the traditional Dynamic Window Approach (DWA) for autonomous vehicles in complex dynamic environments, this paper proposes an adaptive local trajectory planning method based on the ACT-R cognitive mechanism. First, the system constructs an environmental perception based on obstacle clearance, predicted risks from dynamic obstacles, and target progress, and establishes three cognitive planning strategies: Normal, Avoid, and Recover. Second, the memory activation mechanism of ACT-R is utilized to describe the historical usage characteristics of different strategies, and utility feedback learning is employed to reflect the long-term effectiveness of strategies under different environmental states. This allows the results of strategic decisions to be mapped in real time to the multi-objective evaluation function of DWA, thereby achieving adaptive and dynamic adjustment of planning weights. Finally, simulation results show that in mixed static and dynamic obstacle scenarios, while ensuring 100% collision-free arrival, this method achieves an average speed 30.17% higher than traditional DWA, reduces travel time by 18.48%, and improves trajectory smoothness by 46.34%, effectively balancing safety and travel efficiency.