A hybrid-enhanced walrus optimizer for mobile robot path planning

Abstract

Path planning is essential to autonomous navigation in mobile robots, and its performance directly affects task execution efficiency and operational safety. To address the limited convergence accuracy, susceptibility to local optima, and limited late-stage exploitation capability of the walrus optimizer (WO) in path planning, this paper proposes a hybrid-enhanced walrus optimizer (HEWO). HEWO introduces an adaptive differential evolution update into the migration phase, employs historical elite experience exchange to reuse high-quality information across generations, and incorporates Student’s t-driven thermal conduction search to strengthen late-stage local exploitation, thereby enhancing global exploration and population diversity while mitigating stagnation at local optima. Benchmark and statistical tests on CEC2017 demonstrate the overall optimization advantage of HEWO. Using a composite fitness function integrating multiple path-quality objectives, HEWO is further evaluated for global path planning on two-dimensional grid maps. Across six diverse grid maps, HEWO achieves average reductions of 16.87%, 10.58%, and 33.69% in mean composite fitness, mean path length, and mean smoothness cost, respectively, relative to the best competing result for each metric, while maintaining a 100% feasible path output rate. On three complex 50 × 50 grid maps, HEWO obtains lower best composite fitness values than A*. Simulations based on the robot operating system show that, compared with A*, HEWO achieves a comparable mean path length, 33.3% fewer major turns, and a 4.5% shorter mean travel time. These results demonstrate HEWO’s favorable optimization capability and potential for global path planning in mobile robots.