论文部分内容阅读
针对坦克稳定器这一类非线性和不确定性的复杂控制对象,提出一种神经滑模控制方法。该方法将滑模控制与神经网络相结合,解决了控制系统跟踪性能和鲁棒性能之间的矛盾。系统中的滑模控制器保证了系统的快速跟踪性能;而神经网络具有很强的自学习功能,通过学习能够保证系统的稳定性,同时可对扰动和参数变化进行有效的抑制补偿,从而在不牺牲系统鲁棒性的同时达到削弱抖振的目的。从理论上证明了滑动平面的稳定性,并且通过仿真验证了该结果。仿真结果表明该设计方法优于经典设计,为实际设计提供了一种可行的新方法。
Aiming at the complex nonlinear and uncertain control objects such as tank stabilizer, a neural sliding mode control method is proposed. This method combines sliding mode control with neural network to solve the contradiction between tracking performance and robust performance of control system. The sliding mode controller in the system ensures the fast tracking performance of the system. However, the neural network has a strong self-learning function, which can ensure the stability of the system through learning, and can effectively suppress and compensate the disturbance and parameter changes. Without sacrificing the robustness of the system while achieving the purpose of reducing chattering. The stability of the sliding plane is theoretically proved, and the result is verified through simulation. The simulation results show that the design method is better than the classical design, which provides a feasible new method for practical design.