A distributed motion planning approach to cooperative underwater acoustic source tracking

TL;DR

This paper introduces a distributed model predictive control framework for cooperative underwater acoustic source tracking using AUVs, optimizing their guidance for improved accuracy and robustness under communication and environmental constraints.

Contribution

It presents a novel distributed control approach combining RHC, UT prediction, and multi-agent decision-making for enhanced underwater source tracking performance.

Findings

Outperforms decentralized MPC and PSO in simulations.

Improves tracking accuracy under environmental uncertainties.

Ensures robust coordination despite communication constraints.

Abstract

Accurate tracking of underwater acoustic sources is critical for a variety of marine applications, yet remains a challenging task due to communication constraints and environmental uncertainties. In this regard, this paper addresses the problem of underwater acoustic source tracking using a team of autonomous underwater vehicles (AUVs). The core idea is to optimize the guidance of each agent to achieve coordinated motion planning that leads to optimal geometric configurations with respect to the target, thereby enhancing tracking performance. To tackle this, we propose a Distributed Model Predictive Control (DMPC) framework to improve performance and robustness. The control problem is formulated as a multi-objective optimization task, incorporating geometric observability, proximity to the target, and communication connectivity. A Receding Horizon Control (RHC) approach, coupled with an Unscented Transform (UT)-based prediction scheme, is employed to ensure longterm tracking accuracy while accounting for uncertainties. The optimization is distributed using the sequential multi-agent decision-making framework, combined with the Time-Division Multiple Access (TDMA) communication protocol. The proposed methodology is implemented in a simulation environment that accounts for the constraints of acoustic communication. The approach is compared with existing methods such as decentralized MPC and Particle Swarm Optimization (PSO).