Multi-AUV cooperative search method based on dynamic optimal coverage

Abstract

This paper introduces a multi-AUV target search method founded on dynamic optimization coverage, aiming to plan paths that achieve complete coverage without backtracking and with reduced travel costs for participating AUVs in cooperative search missions. The study comprises three key components: 1) Environment Modeling:We present a mixed-type grid approach, capable of accurately modeling concave–convex polygons. 2) Area Allocation:A dynamic cooperative partition strategy based on K-means+ is employed to facilitate optimal area allocation for AUVs. 3) Path Planning: The paper introduces a search path planning algorithm called the Optimal Allocation Minimum Spanning Tree (OA-MST). OA-MST generates globally smooth and non-backtracking scanning paths within irregular task areas using an adaptive spanning tree. Moreover, the study includes multiple simulation experiments to validate the method’s efficacy. OA-MST is compared with the lawnmower strategy and four advanced Minimum Spanning Tree (MST) generation methods under identical area allocation conditions. The results demonstrate that the proposed collaborative search method significantly enhances path smoothness and reduces path costs across diverse scenarios. In addition, sea trials are conducted to verify the feasibility and practicality of implementing the proposed methods in engineering applications.