Adaptive Distributed Lyapunov-Based Model Predictive Control for Multi-UAV Formation Tracking with Weighted Directed Graphs

This paper tackles the formation tracking control problem of multiple unmanned aerial vehicles (UAVs) with a weighted directed communication graph. A novel adaptive fully distributed model predictive control (DMPC) framework is proposed, aiming to not only ensure closed-loop stability and feasibility of the local optimization problem, but also moderate the computation and communication consumption. Each local controller comprises an outer translation control loop and an inner rotation control loop. The outer loop employs Lyapunov-based model predictive control (MPC) to determine optimal translation control actions subject to input constraints. The MPC problem is formulated using only the relative neighborhood formation error between itself and its neighbors, thereby substantially reducing the total communication traffic. An adaptive estimator is also incorporated to estimate the future evolution of neighboring vehicles. Through closed-loop analysis, sufficient stability conditions regarding the selection of control parameters are established. Simulation results are provided to demonstrate the effectiveness of the proposed design.