Obstacle avoidance planning of robot based on hybrid strategy to improve dung beetle and DWA algorithm

In order to address the path planning challenge of robots in intricate and ever-changing environments, a path planning approach that integrates an enhanced dung beetle algorithm with an improved dynamic window approach (DWA) is introduced. An improved dung beetle algorithm is used to generate the optimal solution path. To address the inherent limitations of the original dung beetle optimization (DBO) algorithm, particularly its susceptibility to premature convergence and suboptimal solution accuracy, this study develops a hybrid initialization framework synergistically combining Fuch’s chaotic mapping with opposition-based learning mechanisms. Furthermore, a dual-distribution mutation operator blending Cauchy and Gaussian distributions is incorporated to enhance the global search capability and convergence characteristics of the enhanced algorithm. Regarding the avoidance of dynamic obstacles, the improved DWA is used for path tracing. In order to solve the problem that the DWA algorithm may collide and get stuck in a local minimum, three enhancement strategies are suggested. The key points are introduced combined with global path planning to avoid local optimum. On this basis, combined with the safe braking distance and dynamic obstacle evaluation function, the safety is improved the capability to deal with changing obstacles is enhanced. The two refined algorithms are merged, and experiments of dynamic and static path planning are set up. Experimental findings indicate that the combined algorithm enables real-time dynamic obstacle avoidance using the globally optimal path, with shorter planning paths, higher safety, and smoother linear velocity changes.