Control architecture of autonomous underwater vehicle for coverage mission in irregular region

Abstract This paper proposes a control architecture of autonomous underwater vehicle (AUV) for coverage mission in irregular region (CMIR) with an improved task planner. The CMIR control architecture is structured in four layers: perception layer, reactive layer, decision making layer and execution layer. To improve the task planner in decision making layer of control architecture, we adopted convex partitioning method and Success-History Based Adaptive Differential Evolution Algorithm (SHADE) including linear population size reduction (L-SHADE) as the computing engine, which make CMIR control architecture more suitable to coverage problems. Through statistical analysis, we obtained a set of optimized control parameters. Simulation experiments were carried out, three task planners were compared in terms of convergence speed, optimized path length and computing time consumption. The results show that the coverage efficiency of the task planner based on L-SHADE is higher than that of task planner based on particle swarm optimization (PSO) and differential evolution (DE). Sea trial was performed with Sailfish-AUV, which proved that the CMIR control architecture is feasible in actual marine environment.