Neural-based adaptive control for nonlinear systems with quantized input and the output constraint

Highlights：

• The restrictive assumptions on quantization parameter in the existing literature are removed, a quantizer control design method, which is independent of the quantization parameters, is introduced to compensate for the effect of input quantization. Besides, in order to design a suitable controller, an intermediate virtual control signal is introduced in the last step during the control design.

• Unlike the existing literature , unknown nonlinear function terms are approximated by utilizing radical basic function NNs in this study. Moreover, there is only one adaptive law, which greatly reduces the computational burden. Together with the backstepping technique and stability theory, the approximation-based adaptive tracking control scheme is generalized which the output constraint and quantized input are considered simultaneously.

摘要

•Based on the decomposition of hysteresis quantizer which is introduced in the existing literature, the problem caused by quantized input is solved. Compared with the existing literature, the system considered in this study is more general due to the existence of the output constraint. To this end, a log-type Barrier Lyapunov function is applied in this study.•The restrictive assumptions on quantization parameter in the existing literature are removed, a quantizer control design method, which is independent of the quantization parameters, is introduced to compensate for the effect of input quantization. Besides, in order to design a suitable controller, an intermediate virtual control signal is introduced in the last step during the control design.•Unlike the existing literature , unknown nonlinear function terms are approximated by utilizing radical basic function NNs in this study. Moreover, there is only one adaptive law, which greatly reduces the computational burden. Together with the backstepping technique and stability theory, the approximation-based adaptive tracking control scheme is generalized which the output constraint and quantized input are considered simultaneously.