Design and Performance Evaluation of Nonlinear Model-Predictive Control for 3-D Ground Target Tracking With Fixed-Wing UAVs

Abstract

This study presents the design of a nonlinear model-predictive controller (NMPC) for a fixed-wing uncrewed aerial vehicle (UAV) to circumnavigate a ground target. First, a nonlinear 3-D target tracking system model is presented. Subsequently, an NMPC is designed and formulated as a nonconvex optimal problem. To derive sufficient stability conditions for a nonlinear closed loop, a linear controller with bounded disturbance is analyzed in a specific terminal region. The controlled trajectory is attracted to the terminal region in the vicinity of the system reference, thereby enabling the use of convex model-predictive control tools for the proposed NMPC. Consequently, the NMPC closed-loop system is proven to reach the terminal region in a fixed prediction horizon, and consequently, the UAV can track the ground target. During the course, an initialization technique is used for optimization to prevent stability compromise by suboptimality. System stability is met for three different speed references with variations in the weighting factors. Extensive simulations are conducted to validate the proposed approach. Experimental results are included, providing insights into the field tests and verifying the control development. The results show that the UAV system is successfully steered to the target reference while effectively remaining within its confines.