2026-04-17T02:49:13Zhttps://recercat.cat/oai/request

oai:recercat.cat:2117/4414422025-09-11T09:28:34Zcom_2072_1033col_2072_452951

Design of predictive control approaches for multi-agent systems: application to autonomous robot navigation Borràs Borràs, Josep Àrees temàtiques de la UPC::Informàtica Mobile robots Predictive control Multiagent systems Robots mòbils Control predictiu Sistemes multiagent The deployment of multi-robot systems in applications such as automated warehouses, autonomous exploration, agriculture monitoring, and smart transportation requires scalable control strategies that ensure safety, efficiency, and coordinated behavior in dynamic and uncertain environments. This thesis addresses the navigation planning of multiple non-holonomic mobile robots through a framework based on model predictive control. A centralized architecture is first developed using a multiple-shooting formulation, incorporating features tailored to specific requirements, such as the individual tuning of cost weights via Pareto fronts. This serves as abaseline for the subsequentexploration of three additional architectures—local centralized, decentralized via clustering, and distributed—each enabling cooperative trajectory planning under dynamic and collision-avoidance constraints. The proposed methodologies are implementedinMATLABusingCasADi and evaluatedt hrough simulations involving heterogeneous robot fleets. Performance is quantitatively assessed and compared in terms of computational load and task completion time. 2025-09-11T09:28:34Z 2025-09-11T09:28:34Z 2025-09-11T09:28:34Z 2025-06 Bachelor thesis http://hdl.handle.net/2117/441442 Open Access Universitat Politècnica de Catalunya