Hybrid Adaptive Control for FPV Drones: A Unified Mrac-SMC Approach With Disturbance Learning

This paper presents a hybrid adaptive control system combining Model Reference Adaptive Control (MRAC) with Super-Twisting Sliding Mode Control for FPV drones under wind disturbances. Key innovations include normalized adaptation laws with eigenvalue constraints ensuring stability, and a pattern-learning disturbance observer for proactive compensation. The system achieves 2.85 m RMSE trajectory tracking under 2.0 m/s wind speeds, representing 61% improvement over conventional PID while maintaining 1.2 kHz computational efficiency. Mathematical stability guarantees and practical performance make this approach valuable for precision drone applications.