Uncertainty-aware Dual-mode Planning for an Underwater Walking Robot

This study presents an uncertainty-aware dual-mode mission planning framework for an underwater walking robot capable of walking and swimming. Unlike conventional deterministic approaches, the proposed method considers stochastic detection outcomes and uncertainty in traversal time. A chance-constrained formulation ensures that mission completion time satisfies a probabilistic bound, while fuel consumption is limited by a predefined capacity. The objective maximizes the expected mission reward with a time penalty, allowing adaptive early termination via optimal prefix selection. To evaluate performance, a brute-force method provides optimal benchmarks for small-scale problems. Based on these, a computationally efficient heuristic combining greedy insertion and 2-opt refinement is developed for scalability. A limit on swimming actions is also imposed to reflect operational constraints. Simulation results show that the proposed method maintains robust performance under uncertainty conditions while significantly reducing computational complexity compared with exhaustive searches, highlighting the importance of modeling uncertainty and mode-selection trade-offs.