Adaptive Output Feedback Control of Nonlinear Systems With Mismatched Uncertainties Under Input/Output Quantization

It is still an open problem to synthesize control with input and output quantizations for nonlinear systems subject to mismatched parametric uncertainties via backstepping design. This is because 1) the output becomes discontinuous and nondifferentiable after quantization, making the conventional recursive backstepping design method unsuitable as the differentiation of the virtual control signals does not exist, while the existing backstepping-based quantized control method is only applicable to systems modeled in normal form or systems without mismatched uncertainties; 2) it is difficult to deal with the impacts of errors caused by output quantization when the virtual control laws contain time-varying estimates of unknown parameters; and 3) since only the quantized output signal is available, the effects of certain items related to unknown parameters in currently available observers become difficult to be explicitly addressed. In this article, we present a solution to this problem by employing a dynamic filtering technique to avoid differentiating virtual control signals in backstepping design, using the parameter projection technique to design the parameter estimator and proposing a new form of state observer to facilitate the development of output feedback control. It is shown that, with the derived adaptive backstepping control involving both input and output quantizations, all the closed-loop signals are ensured bounded and the control performance in terms of the mean square error is adjustable through properly choosing certain design parameters. The benefits and effectiveness of the proposed scheme are also validated by numerical simulation.