Adaptive Dynamic Window Approach for Mobile Robot Motion Control Using Particle Swarm Optimization

Autonomous navigation for mobile robots in dynamic and unknown environments requires robust and adaptable path planning algorithms. Among the most widely used local path-planning algorithms is the Dynamic Window Approach (DWA), known for its reactive obstacle avoidance and local trajectory planning. The DWA is widely used for real-time obstacle avoidance; however, it extensively relies on the proper tuning of the weights in its objective function, which manages the trade-off between reaching the goal and avoiding collisions with obstacles. This paper proposes an adaptive DWA for the navigation of mobile robots in static and dynamic environments. The proposed approach is based on the integration of an online Particle Swarm Optimization (PSO) based framework, which dynamically adjusts these weights to improve adaptability and efficiency. The proposed online framework was compared with the classical DWA baseline and an adaptive variant based on genetic algorithm (GA-DWA) as a comparative optimizer, which resulted in a substantial reduction in travel time by 30% and increased the success rate by 20%, with a 10% increase in average velocity over the classical DWA, and consistently outperformed GA-DWA, especially in dense and dynamic environments.