Optimized Self-Triggered Attitude Tracking Control for Quadrotor Uncrewed Aerial Vehicle: An RL-Based Predefined-Time Control Strategy

초록

This article mainly focuses on the optimized adaptive self-triggered attitude control issue for quadrotor uncrewed aerial vehicle via a reinforcement learning (RL)-based predefined-time control scheme. In the procedure of recursive design, the command filtered backstepping control technique is introduced to evade the issues of computational complexity and weaken the effects of filter error on control accuracy. Furthermore, a predefined-time optimized adaptive attitude control solution is proposed by associating with predefined-time stability theory and RL, where interval type-2 fuzzy logic systems are employed to concentrate on unknown nonlinearity. Moreover, the self-triggered mechanism is utilized to substitute for time/event-triggered mechanisms to enhance the utilization of bandwidth resource and the realizability of signal transmission on demand. It is demonstrated by stability results that predefined-time constraint for all closed-loop attitude system signals is attainable via the devised RL-based control strategy. Ultimately, the simulation analysis verifies that the reported control solution possesses favorable effectiveness.

키워드