Image-Based Distributed Predictive Visual Servo Control for Cooperative Tracking of Multiple Fixed-Wing UAVs

This letter proposes a novel approach combining the distributed model predictive control (DMPC) and the image-based visual servoing (IBVS) for cooperative tracking problem of multiple fixed-wing Unmanned Aerial Vehicles (UAVs) equipped with pan-tilt cameras. In particular, the target is unknown and can only be perceived by cameras. Different from existing methods that require the target localization and the positions of the UAVs, our proposed IBVS scheme achieves cooperative tracking of multiple UAVs by directly utilizing image feature points. It is also applicable to scenarios where UAV positioning information is unavailable. The approach integrates the control of both the UAVs and pan-tilt cameras, enabling the UAVs to circle around the target and keep it near the image center at the same time. Moreover, the UAVs are strategically distributed around the target to ensure optimal observations. Due to the dynamic constraints of the fixed-wing UAVs and the perception limitations imposed by cameras, MPC is utilized to maintain the target in close proximity to the UAVs for improved observations. To meet real-time optimization requirements, a DMPC solution based on soft constraints is proposed, and the stability of the distributed system is proven using the Lyapunov method. Finally, a hardware-in-the-loop (HIL) simulation platform is established and experimental results verify the feasibility and effectiveness of our proposed approach.