Acoustic communication-aware coordination algorithm for multi-robot systems in underwater exploration

Coordinating Marine Multi-Robot Systems (MMRS) is particularly challenging due to the limitations of underwater acoustic communication, including high latency, signal attenuation, and low bandwidth. This paper presents a novel coordination framework for a heterogeneous MMRS comprising multiple Autonomous Underwater Vehicles (AUVs) and an Autonomous Surface Vehicle (ASV) serving as a mobile communication relay. The framework integrates two decision aggregation methods – Ordered Weighted Averaging (OWA) and the Adaptive Response Threshold Model (ARTM) – to support coordination based on heterogeneous criteria such as communication quality, spatial distribution, and task urgency. These methods, rarely applied in multi-robot coordination and never before in underwater domains, enable more flexible and robust decision-making. A key contribution of this work is the incorporation of experimentally derived acoustic communication characteristics into the decision-making process. By leveraging real-time signal quality data, the ASV can adaptively prioritise data collection, enhancing communication efficiency, preserving connectivity, and reducing latency. This is the first known approach to combine aggregation-based decision strategies with real-world underwater communication metrics. Simulations under realistic marine conditions show that this communication-aware coordination significantly improves performance and mission success, offering a new direction for intelligent decision-making in underwater multi-robot systems.Highlights A novel coordination framework is presented for marine multi-robot systems (MMRS), integrating real-world acoustic signal quality characterisation into the aggregation-based decision-making process.The method addresses the dual challenge of maximising data throughput and minimising latency in underwater environments, by incorporating acoustic performance metrics directly into coordination logic.Two decision aggregation strategies, ARTM and OWA, are implemented and evaluated under realistic conditions, providing the first comparative analysis of these techniques in the context of underwater communication-constrained robotics.Simulation results using a communication-aware MMRS framework show that OWA offers greater robustness and flexibility, while ARTM provides more consistent latency, offering design guidance for latency-sensitive missions.An open-source, ROS-based MMRS architecture is provided, enabling seamless transition from simulation to real-world deployment, and supporting reproducibility and innovation in collaborative marine robotics.