Multi-robot path planning based on the improved nutcracker optimization algorithm and the dynamic window approach

Abstract

Multi-robot path planning faces challenges such as conflict avoidance, collaboration, and dynamic environments. This paper proposes a multi-robot path planning algorithm that integrates the improved nutcracker optimization algorithm with the improved dynamic window approach. To address the nutcracker algorithm’s sensitivity to initial conditions and slow convergence, a population initialization strategy is introduced for more diverse initial populations. Additionally, a simplified path node strategy is also designed to shorten paths and reduce steering times. By incorporating a dynamic inertia weight factor w, the balance between global exploration and local optimization is improved. To address the limitations of the dynamic window approach, which is unable to avoid dynamic obstacles instantly and is prone to falling into local optimal solutions, the target distance subfunction, the path evaluation subfunction and the deviation from danger zone subfunction are added to the evaluation function. Finally, the two algorithms were fused together and we conducted four experiments to validate the performance of the MANOA, IDWA, and MANOA-IDWA algorithms, as well as the application of MANOA-IDWA in multi-robot path planning. Results show that MANOA-IDWA significantly increases path planning success rates in dynamic environments, producing shorter and smoother paths, thus enhancing the safety and stability of multi-robot operations.