In this paper, we investigate the coverage path planning (CPP) scheme for autonomous underwater vehicles (AUVs) cooperative environment detection in integrated underwater acoustic communication and detection networks (UCDNs), where multiple AUVs detect unexplored oceanic environments and avoid obstacles. Firstly, we present the detection range prediction model related to oceanic environmental parameters and propose the detection and communication scheme in UCDNs. Secondly, to conduct the cooperative environment detection mission, we formulate the CPP problem as a mixed combinatorial and sequential quadratic optimization problem to maximize the coverage ratio and minimize the path length of AUVs. To solve this problem, we investigate the multi-agent proximal policy optimization (MAPPO)-based CPP scheme. In specific, the CPP problem is modeled as a partially observable Markov decision process (POMDP). Since the path planning of the AUVs is not only related to the local information but also the other AUVs' information, the information should be shared among AUVs based on the UCDNs. Furthermore, we introduce the MAPPO-based algorithm under the centralized training with decentralized execution (CTDE) architecture. Extensive simulations are carried out to demonstrate the strength of the proposed scheme.
Coverage Path Planning for AUVs Cooperative Environment Detection in Integrated Underwater Acoustic Communication and Detection Networks
2024-06-24
9066490 byte
Conference paper
Electronic Resource
English
Improved underwater acoustic communication for AUVs
Tema Archive | 1990
|Underwater acoustic telemetry for AUVs
IET Digital Library Archive | 1996
|Real-time target detection using acoustic sensors on AUVs
Tema Archive | 2005
|AUVS' Dynamics Modeling, Position Control, and Path Planning Using Neural Networks
British Library Conference Proceedings | 2002
|