Research On Model Free Adaptive Control With Data Quantization

With the rapid development of science and technology,the actual industrial production process is becoming more and more complex,and it is very difficult to establish an accurate mathematical model for these process systems,so data-driven control theory has become a research hotspot in the field of control in recent years.On the other hand,most of the actual control systems need to run and work in the network environment.Networked systems bring data sharing and information processing convenience,but also bring data quantification,data packet loss,time delay,bandwidth constraints and other issues due to its own structural parameters.In this paper,we consider the stability analysis and improved algorithm design of two data-driven control theories and methods in the presence of quantized data.(1)The control problem of a class of nonlinear discrete single-input single-output(SISO)systems with data quantization is studied.A data-driven control method with uniform quantizer is proposed.The dynamic linearization of the controlled nonlinear system and the model linearization of the original nonlinear system are realized by using pseudo partial derivatives.Then a method of model-free adaptive control with uniform quantizer(MFACUQ)is constructed,in which the output data of the system is used as quantization signal.Through theoretical derivation and analysis,the tracking error of the system is proved to be convergent and bounded under MFACUQ method.(2)In order to achieve better error convergence effect,the proposed MFACUQ algorithm is improved by adding a coding and decoding quantization mechanism.A model-free adaptive control algorithm with coding and decoding quantization mechanism(MFACUQ-E)is proposed,and the feasibility of convergence condition and the tracking effect after data quantization are discussed.The results show that the tracking effect of MFACUQ-E method is better than that of MFACUQ method,and the tracking error can converge to zero under certain conditions.Simulation results demonstrate the effectiveness of the proposed MFACUQ-E method.(3)The control problem for a class of discrete single-input single-output nonlinear systems with repetitive tasks is studied.The uniform quantizer is used to process the data quantization process.A model-free adaptive iterative learning control(MFAILCUQ)algorithm with coding and decoding quantization mechanism is proposed.Theoretical analysis and simulation results show that the proposed MFAILCUQ method can track the desired trajectory with zero error under the condition of considering data quantization.(4)The control problem of a class of discrete single-input-single-output nonlinear systems with repetitive tasks is studied,and a model-free adaptive iterative learning control(MFAILCLQ)algorithm with logarithmic quantizer is proposed.The tracking error of the system can converge to zero quickly under the algorithm.Simulation results demonstrate the effectiveness of the proposed MFAILCLQ algorithm.