Data Collection in Underwater Acoustic Sensor Networks Using Multiple AUVs With Sequential Greedy-QPSO and Fast Marching Algorithms

Multiple autonomous underwater vehicles (AUVs) face challenges such as ocean currents and obstacles when collecting data from underwater acoustic sensor networks (UASNs). This article proposes a novel method to enhance the safety and efficiency of data collection by focusing on waypoint selection and path optimization. Initially, we integrate the sequential greedy quantum-behaved particle swarm optimization (SeqGreedy-QPSO) algorithm for waypoint selection. Subsequently, task points are clustered using balanced k-means clustering and allocated to each AUV using the self-organizing map (SOM) algorithm. Path planning utilizes the time matrices from the weighted fast marching square (FMS) algorithm and the anisotropic fast marching (AFM) algorithm to generate safe and efficient paths, considering the impact of obstacles and ocean currents. Through comparative simulation experiments, our approach demonstrates significant advantages in waypoint selection and path safety. The findings suggest that our method significantly boosts the efficiency and reliability of multi-AUV data collection in complex underwater environments, providing strong technical support for real-world applications.