Intelligent Vector Field Histogram based collision avoidance method for AUV

Abstract

Intelligent Vector Field Histogram (IVFH) is proposed to accomplish the collision avoidance of Autonomous Underwater Vehicle (AUV) in underwater three-dimensional dynamic space. The directed cuboid static obstacle environment model is introduced to improve the efficiency of collision avoidance decision-making. The candidate direction set is composed of relative heading and relative pitch. According to the obstacle distribution of local environment, the traveling fitness of each candidate direction is calculated from the perspectives of safety and rapidity, so as to realize the collision avoidance of AUV in three-dimensional dynamic environment. IVFH method can adaptively adjust the parameters that affect the collision avoidance action amplitude according to the obstacle distance and size. The collision avoidance velocity is adjusted by fuzzy logic according to the environment conditions, so as to ensure that AUV can fully perceive the environment and then complete the collision avoidance decision-making. The simulation results show that IVFH method can overcome the threshold sensitivity of traditional VFH method, escape from the trap area, and realize collision avoidance in three-dimensional dynamic complex obstacle environment. The proposed IVFH collision avoidance method has been verified by simulations, as well as lake and sea experiments, demonstrating its effectiveness and practicability.