This paper develops a neural-network-based model reference adaptive control (MRAC) scheme for a rotorcraft in the presence of input saturation. Such a control scheme provides acceptable tracking performance for the rotorcraft in a wide range of flight conditions. Combined with hyperbolic tangent functions, the MRAC scheme is capable of tracking the reference signals without violating input constraints. A modified projection operator is utilized to prevent system from parameter drift due to the strong nonlinearity and uncertainty of the rotorcraft mathematical model. Stability of the proposed MRAC scheme is proved based on Lyapunov stability theory. The performance of the resulting controller is tested by conducting numerical simulations for an autonomous rotorcraft in various flight missions.

本文为存在输入饱和的旋翼机开发了一种基于神经网络的模型参考自适应控制 (MRAC) 方案。这种控制方案为旋翼机在广泛的飞行条件下提供了可接受的跟踪性能。结合双曲正切函数，MRAC 方案能够在不违反输入约束的情况下跟踪参考信号。由于旋翼飞行器数学模型的强非线性和不确定性，采用修正的投影算子来防止系统参数漂移。基于李雅普诺夫稳定性理论证明了所提出的 MRAC 方案的稳定性。通过在各种飞行任务中对自主旋翼机进行数值模拟来测试所得控制器的性能。