Flatness-Based Predictive Control with EKF for UAV Target Tracking

Unmanned Aerial Vehicles (UAVs) require robust perception and control strategies for autonomous tracking of moving targets. This work integrates Extended Kalman Filter (EKF)-based trajectory prediction with a flatness-based Model Predictive Control (MPC) scheme for vision-based tracking. The EKF smooths noisy measurements and predicts short-term target motion, ensuring continuity during occlusions. These predictions serve as references for an MPC controller formulated on the quadrotor’s differentially flat outputs, enabling a simplified linear predictive model while respecting acceleration and input constraints. Real-time simulations demonstrate accurate and smooth tracking of the EKF-predicted trajectories, confirming the effectiveness of the combined EKF-MPC framework for vision-guided UAV target tracking.