Path Planning of Multi-Modal Underwater Vehicle for Adaptive Sampling Using Delaunay Spatial Partition-Ant Colony Optimization

Multi-modal underwater vehicle (MUV) is a newly ocean observation platform with the capability to switch between Argo mode and glider mode. This paper presents an efficient and effective path planner to generate trajectory that exploits ocean energy and makes use of favorable ocean currents to propel and lead the MUV to collect ocean measurements located in high scientific interest areas. The proposed strategy employs MUV mode transition within path planner to enable efficient operation of MUV to drift with ocean currents for measurements and transform to operate as an underwater glider for locomotion. The novel hybrid scheme is the first to integrate tournament selection based Delaunay triangulation into ant colony optimization (ACO), herein referred to as DSP-ACO path planner to facilitate more efficient searching for adaptive ocean sampling problem (AOSP). Simulation results show that the proposed DSP-ACO path planner is capable to take advantage of ocean currents and obtain a more optimal trajectory with quicker convergence rate than conventional ACO path planner. Meanwhile, MUV transforms operation between Argo mode and glider mode is more energy efficient to take measurements than underwater glider. Monte Carlo simulation is run to assess the robustness of the DSP-ACO path planner and demonstrates the inherent superiority of the proposed DSP-ACO path planner.