Local Path Planning for Mobile Robots Based on Fuzzy Control DWA

The Dynamic Window Approach (DWA) is a core algorithm for local path planning in mobile robots, but it faces challenges regarding the adaptability of the fixed-weight evaluation function in dynamic and complex environments. In traditional DWA algorithms, the weight parameters for goal orientation, obstacle avoidance, and speed are typically fixed, which can lead to issues such as path oscillations and obstacle avoidance failures in dense or highly dynamic environments. To address this, this paper proposes an adaptive weight adjustment mechanism based on a fuzzy logic controller, which dynamically optimizes the evaluation function's weights by perceiving environmental features in real time. Specifically, a fuzzy inference system is designed with obstacle density and target distance as inputs to adjust the priority of weights for orientation, obstacle avoidance, and speed, achieving a multi-objective (safety, efficiency, smoothness) collaborative optimization. Simulation results demonstrate that the proposed algorithm reduces collision frequency by 18.8%, decreases path curvature by 12%, and shortens target arrival time by 15% compared to the traditional DWA in dynamic obstacle environments. Practical tests on the TurtleBot3 platform further validate the algorithm's effectiveness in narrow passages and sudden obstacle scenarios. This research provides a highly interpretable and adaptable solution for local path planning in complex environments, with broad application potential in fields such as warehouse logistics and service robotics.