Integrating acoustic communication and distributed optimization for cooperative AUV control

In the context of cooperative underwater robotics, this thesis proposes a distributed framework for the cooperative coordination of Autonomous Underwater Vehicles (AUVs) that is based on integrating underwater acoustic communication with distributed optimization techniques. The objective is the development of a modular and scalable system aiding autonomous missions in subsea marine environments, elevated communication latency, and heavily constrained communication bandwidth. The developed cooperative coordination scheme implements an ad-hoc aggregative algorithm wherein all agents contribute to the cooperative resolution of the problem. Communication among the AUVs happens through an acoustic infrastructure that is built upon specialized protocols, including Time Division Multiple Access (TDMA) for deterministic channel access, and Dynamic Compact Control Language (DCCL) for efficient message compression. The entire communication system has been developed in the Rust programming language and integrated on embedded hardware platforms. For the implementation of the optimization algorithm, its integration with acoustic communication and simulation of AUV behavior were used containerized software tools, ROS2, Choirbot and Gazebo. The tests have demonstrated the framework capability to support multi-robot missions in the presence of communication delays, while preserving informational coherence among agents and maintaining the collective behavioral stability. The obtained results confirm the validity of the proposed infrastructure, however, future testing in real operational scenarios will be necessary. In such contexts, mechanisms for dynamic adaptation of communication parameters and packet loss tolerance will need to be studied and implemented in order to ensure the robustness and reliability of the system.