Quantized Feedback Controller Design For Real-Time Network Systems

In networked control systems, the data amounts of measurement signals and control signals sent by transmitters and controllers are limited, owing to the communication constrains from networks, which further limits the number of physic plants and controllers collected to the networks. For conventional control systems, however, it's unnecessary to keep the transmitted information with high frequency or high data accuracy. Under certain data precision and sampling frequency, the physic plants can achieve enough control performance. With the purpose of improving the utilization efficiency and adding more physical plants into the network, quantized feedback control algorithms are broadly proposed, which will reduce data accuracy, namely, the data amounts for each transmission.In this paper, we firstly introduce event-driven control into quantized feedback control systems, to potentially reduce the negative influences of communication constrains on control performances of networked control systems. To this end, we design a quantized feedback control algorithm based on event-driven strategy, including the event detector, event-driven quantization encoder and decoder and event-driven controller. We adopt the fixed event threshold and time-varying one separately to achieve different kinds of control stability. For both fixed and time-varying approaches, we address the proofs of related theorems for the relationship between event threshold and quantization length, stabilities and existence of the minimum inter-event interval. It is noteworthy that the proposed quantized control methods only need to design very few parameters, rather than the plant's specific model.The introduction quantization strategies, in fact, will influence system performances. Therefore, another main research content in this paper is to design an algorithm to estimate and actively compensate these influences. For the quantized input feedback control system, we firstly treat quantization errors as discontinuous disturbances, while considering the closed-loop system, the quantization errors are modelled as states of an extended system; then we develop the impulsive extended state observer (IESO) to estimate the quantization errors. With the proposed impulsive extended state observer, quantization errors can be estimated accurately, and the system performances are improved as well. On this basis, this paper analyzes both quantized input state feedback control and quantized input output feedback control, and gives proofs of related theorems for impulsive extended state observer. In addition, we also analysis the stability condition of the closed-loop system.We make simulations for all the algorithms mentioned before, the results are compared with traditional approaches, which turn out to be better.