LLM-DWA: A Hybrid Path Planning FrameworkCombining Large Language Models with the Dynamic Window Approach

This research addresses the local minima problem in the Dynamic Window Approach (DWA) algorithm. The conventionalDWA, which does not incorporate prior environmental knowledge, often shows degraded goal-reaching performance incomplex scenarios, such as environments with U-shaped obstacles, and even when it reaches the goal, the path planningtime is relatively long. To overcome this limitation, we propose an efficient DWA by using Large Language Models (LLMs).Leveraging the reasoning capabilities of LLMs, prior environmental information is interpreted, and appropriate intermediatewaypoints are generated. Experimental results in both 2D grid environments and 3D simulation platforms demonstrate thatthe proposed LLM-based hybrid method achieves higher efficiency and shorter goal-reaching times in U-shaped obstaclescenarios compared to the conventional DWA. These findings highlight the effectiveness of combining the reasoning capabilitiesof LLMs with DWA to improve navigation performance in complex environments. A video demonstration is available athttps://youtu.be/Otn53HS4KC4.