CRGWO-DWA: A structured global–local collaborative planning framework for smooth UAV navigation in dynamic environments

UAV path planning in dynamic environments requires balancing global path quality, local obstacle avoidance, and motion smoothness. To address the premature convergence and unsmooth paths of the Grey Wolf Optimizer (GWO) and the local-optimum tendency and evaluation imbalance of the Dynamic Window Approach (DWA), this paper proposes CRGWO-DWA, a structured global-local collaborative planning framework. In the global planning stage, CRGWO enhances population diversity and search performance through piecewise linear chaotic initialization, periodic regeneration of the control parameter, and adaptive leadership weighting, while cubic spline interpolation is used to generate a smooth reference path. In the local planning stage, DWA is improved using min-max normalized evaluation and redesigned weighting to enhance trajectory discrimination and responsiveness under dynamic obstacle interference. Experimental results show that CRGWO achieves the best or tied-best performance in most reported benchmark metrics, reduces path length and redundant turning nodes in static environments, and produces more compact and smoother paths in three-dimensional scenarios. In dynamic multi-obstacle environments, CRGWO-DWA enables safe and smooth UAV navigation with stable velocity profiles. These results indicate that the proposed framework provides an effective solution for UAV path planning in dynamic environments.