一种基于神经网络的任意模型参考自适应控制

摘要

针对一般模型参考自适应控制方法在解高阶非线性模型时参考模型阶数较高的不足,采用一种任意模型参考自适应控制降低了参考模型的难度.利用隐层神经网络对模型进行逼近,对线性化时由不确定因素导致的误差进行补偿,并利用直接Lyapunov稳定性理论证明了跟踪误差有界,最后将其应用到飞行器纵向非线性模型的自动着陆下滑控制设计中.仿真结果表明,所设计的控制器能够使飞行器较好地跟踪理想着陆轨迹,从而验证了方法的有效性.%In order to eliminate the disadvantage of high order which comes from the common model reference adaptive control method when solving the high order nonlinear system problem, the output feedback arbitrary reference model adaptive control is introduced: and the single hidden layer neural network is used to approach the reference model so as to compensate the error which comes from linearization and uncertain; the direct Lyapunov stability theory is used to prove the boundary of the error. Finally, the method is used to solve the automatic aircraft glide-landing controller design based on longitudinal nonlinear model. The simulation results show that the designed controller can meet the requirements of the aircraft which has been designated mission.