NON-COOPERATIVE CONTROL OF A SHUTTLE DRONE FOR MANNED DRONE CAPTURE

• Autor(a) principal: Rodrigues, Patrícia Alexandra Gonçalves
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10362/155443
• topic: Human-in-the-loop; Model Predictive Control; Non-linear MPC; Differential games; Pursuit-Evasion Games; Unmanned Aerial Vehicle; Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática

Drones have become increasingly popularized in recent years, be it in military use or as a hobby for the general public. The increase in usage of drones poses a challenge to security of zones where flying is monitored and prohibited for everyone, such as airports. Unmanned aerial vehicles show great promise in tackling problems such as this, where they must capture a drone in a set environment. This thesis addresses this problem by studying pursuit-evasion games, human-in-theloop as well as non-linear and predictive control in order to create an unmanned aerial vehicle capable of pursuing and capturing a manned drone. The full models for the ground vehicle and the RUAV were done, as well as two simple versions of the fixed-wing UAV, and one simplified version of the RUAV. For the RUAV two simple internal controllers were created, in order to produce the necessary forces and torques. Next, a NMPC controller was created for each vehicle, then a simulation of both vehicles with NMPC was done for joint-tracking, and finally for the differential game. Lastly the human-in-the-loop was added to the NMPC of the fixed-wing, and then to the differential game between the fixed-wing and the RUAV. The results show that a NMPC controller is effective at handling dynamic models and aggressive changes in trajectory while still accounting for the physical constraints imposed. The differential game used provided the pursuit-evader scenario desired, where it was possible to see various interactions between vehicles. Finally, the human-in-theloop was integrated with the fixed-wing model.