Distributed optimization strategies and toolboxes for cooperative multi-robot systems

The widespread adoption of multi-robot systems in areas such as farming, smart warehouses, cooperative exploration, and environmental monitoring highlights their potential to enhance efficiency, scalability, and adaptability in real-world applications. However, integrating intelligent cooperative strategies and transitioning from numerical simulations to experimental deployment are significant opened challenges towards the alignment of academic advances with the modern industrial objectives. This thesis addresses the above challenges by developing novel distributed control and optimization strategies for multi-robot systems that are ready for deployment. These strategies are integrated with distributed computing paradigms and virtualization tools, thus enabling multi-robot systems to accomplish complex tasks. We start by developing a multi-robot software platform that supports the implementation of distributed algorithms by leveraging a high-fidelity simulation engine for mixed-reality deployments. These toolboxes also include software-containerization mechanisms for distributed computing. Subsequently, we propose novel distributed optimization strategies by leveraging the aggregative optimization framework to accomplish cooperative tasks. In this approach, we address challenges related to realistic communication over a sparse network, robot heterogeneity, safety, and scalability. Finally, these strategies are validated through several case studies, including cooperative surveillance scenarios, target encirclement and underwater surveillance. In each scenario, we first validate the proposed algorithms in virtual experiment and then deploy them on real robots, including mixed-reality setups. Through these contributions, this thesis provides a comprehensive set of solutions that integrates novel distributed optimization techniques with distributed computing and virtualization tools, thus enabling the deployment of multi-robot systems in practical applications.