一类不确定非线性系统的鲁棒自适应轨迹线性化控制

针对一类不确定非线性系统,研究了一种新的鲁棒自适应轨迹线性化控制方案.利用径向基神经网络的在线逼近能力以及被控对象分析模型的有用信息设计一种径向基神经网络干扰观测器来估计系统中存在的不确定性.观测器输出用于设计补偿控制律抵消不确定性对系统性能的影响,鲁棒自适应控制律用于克服逼近误差.采用Lyapunov方法严格证明了在自适应调节律作用下闭环系统所有误差信号最终有界.最后利用倒立摆系统验证了新方法的有效性.

Robust adaptive trajectory linearization control for a class of uncertain nonlinear systems

A novel robust adaptive trajectory linearization control(RATLC)method is presented for a class of uncertain nonlinear systems.A radial-basis-function neural network disturbance-observer(RDO)is developed by combining the universal approximation property of radial-basis-function neural network with the useful information of the controlled plant.The RDO output is then used to design a compensation control law to cancel the system uncertainties,and a robust adaptive term is also introduced to eliminate the reconstruction error.By Lyapunov's direct method,a rigorous proof demonstrates that this adaptive law can guarantee the ultimate boundedness for all error signals in the closed-loop system. Finally,the application to control an inverted pendulum shows the effectiveness of the proposed method.