SHARP: A Hybrid Metaheuristc Approach forIntelligent Robotic Path Planning

Robot path planning in crowded environments is a challenging optimizationproblem that requires short and collision-free trajectories under limited compu-tational budgets. Several robot path planning approaches have been proposed,however the majority of them either (1) depend on manually chosen weightsto combine objectives, so their results can be biased and unstable, or (2) canget stuck early in crowded environments, leading to longer paths. To addressthese challenges, we introduce multi-objective planning variants that exploit thePSN hybrid metaheuristic. The novelty of the proposed approach lies in thisdual multi-objective layer: it enables both preference-aware and preference-freeplanning using the same underlying PSN search mechanism, without requiringPareto-archive management or redesigning the optimizer. To improve executabil-ity, the final waypoint sequence is smoothed using cubic spline interpolation. Theproposed approach is evaluated on dynamic and static 2D workspaces with vary-ing obstacle densities and compared against PSO, Grey Wolf Optimizer, and the Sine Cosine Algorithm. Across six scenarios, the PSN-based variants consistentlyreturn the shortest feasible paths, reducing path length by 8–46% relative tothe best competing method and by about 47% on average relative to standardPSO. A dynamic simulation with moving obstacles and a moving target furtherillustrates the approach’s adaptability for online re-planning.