基于最小二乘算法及神经网络的非线性离散系统的自适应控制

Adaptive control of nonlinear discrete-time systems based on least-squares algorithm and neural networks

针对有关文献所设计的控制律 ,在一较弱的条件下 ,去掉了符号运算部分 ,证明该控制律可避免“零除”问题 ,提高了运算速度。对于这种基于神经网络和 L S算法的自适应控制问题 ,证明了系统状态落入一紧集中 ,闭环系统的所有信号都是有界的 ,且系统输出和参考输出之间的跟踪误差收敛于以零为原点的某一有界球中。

Under weaker assumption, the part of signal operation is removed form the control law designed in earlier research proved that the problem of zero divider can be avoided and the rate of operation is improved. For the problem of adaptive control based neural networks and LS algorithm, it is proved that the system state lies in a bounded set. All signals of the closed-loop systems are bounded , and the tracking error between the plant output and the reference output converges to a bounded ball centered at origin.