论文部分内容阅读
针对一类具有未知函数控制增益的非线性系统,利用RBF神经网络的逼近能力,依据滑模控制原理,提出了一种直接自适应神经网络控制器设计新方案。通过引入积分型切换函数及逼近误差自适应补偿项,监督控制用饱和函数代替符号函数,根据李雅普诺夫稳定性理论,证明了闭环系统是全局稳定的,跟踪误差收敛到零。该算法应用于连续搅拌型化学反应器CSTR(Continuous Stirred Tank Reactor),仿真结果显示,该算法能很好地使CSTR跟踪给定的温度信号,表明了该控制策略的有效性。
Aiming at a class of nonlinear systems with control gains of unknown functions, a new scheme of direct adaptive neural network controller design is proposed by using the approximation ability of RBF neural network and sliding mode control principle. According to Lyapunov stability theory, the closed-loop system is proved to be globally stable and the tracking error converges to zero by introducing integral switching function and adaptive compensation term of approximation error. Supervisory control uses the saturation function instead of the sign function. The algorithm is applied to CSTR (Continuous Stirred Tank Reactor). The simulation results show that this algorithm can make the CSTR track the given temperature signal well, which shows the effectiveness of the control strategy.