Dual model predictive control with Kalman filter-based observer for automatic carrier landing

This paper addresses the automatic carrier landing problem for carrier-based aircraft operating under challenging conditions, including deck motion, carrier airwake, sensor noise, and actuator saturation. To tackle these challenges, a dual model predictive control (DMPC) method is proposed, offering constraint-handling capabilities and robustness to both external disturbances and sensor noise. The proposed method employs two predictive models: a nominal model to ensure constraint satisfaction and a learned model to optimize performance. This dual-model structure enables a clear separation between the stability and performance objectives. To ensure system stability and robustness, a terminal constraint set and receding optimization constraints are designed. Moreover, a Kalman filter-based observer is designed to enhance control performance. A novel automatic carrier landing system (ACLS) is then developed based on the proposed DMPC method. Furthermore, an adaptive weight adjustment logic is proposed to further enhance ACLS performance. Comparative simulations of the proposed ACLS against alternative methods demonstrate its efficacy and superiority.