Autonomous UAV Landing via Visual Servoing and Semi-Markov Decision Processes

This paper presents an innovative approach to autonomous UAV landing on moving platforms by integrating a Semi-Markov Decision Process (SMDP) framework with a visual servoing control strategy. The proposed method models the UAV's position relative to the landing target using a Semi-Markov state space, accounting for varying visibility conditions and the precise positioning required for a successful landing. By incorporating a reward function, the SMDP framework enables the UAV to evaluate trade-offs between different actions and dynamically optimize its trajectory. This approach enhances the UAV's ability to adapt to platform movement, handle uncertainty, and improve landing performance in unpredictable environments. The contribution of this work lies in the novel application of SMDPs combined with visual servoing, offering a robust, adaptable solution to the challenges of autonomous landing in real-world, dynamic conditions.