Digital Twin-Driven Hybrid Control for MADNI Stability in Adverse Conditions

This paper presents a novel integration of Digital Twin (DT) technology with hybrid control systems to improve Unmanned Aerial Vehicle (UAV) stability under adverse conditions like high winds. The DT framework enables real-time simulation, monitoring, and optimization, providing predictive insights for flight stability. Traditional controllers, such as Linear Quadratic Regulator (LQR), are combined with Reinforcement Learning (RL) techniques like Deep Deterministic Policy Gradient (DDPG) to develop hybrid strategies. Results reveal that hybrid approaches balance the adaptability of RL with the stability of LQR, outperforming standalone methods. The LQR with Particle Swarm Optimization (PSO) and the DDPG-LQR-PSO hybrid achieve the lowest gradient of -0.14 radians and settling times of 2 seconds for pitch, roll, and yaw. This research pioneers UAV resilience advancements, enabling superior performance in complex, rapidly changing environments.