Research On Stabilization Of Event-triggered Dynamic Quantized Feedback Control Systems With Limited Bandwidth

With the rapid development of network communication technology,networked control systems(NCSs)have attracted more and more concentration.Compared with conventional control systems,NCSs can achieve long distant control mission,accommodate larger-scale control systems and obtain higher feedback channel reliability due to the introduced network channels.However,the digital feedback networks also introduce some problems to NCSs while they bring so many advantages to NCSs.The most significant one is the quantization error introduced by the digital characteristic because the sampled signals must be quantized into discrete values with limited bits before transmission.As NCSs become larger and larger,and their structures are more complicated,more network resources are consumed by NCSs than before.So much attention has been paid to how to improve the efficiency of network resource utilization,such as designing a reasonable dynamic quantizer to reduce the number of bits used for transmission or to extend the sampling interval of the systems to reduce network resource occupation.In addition,there are also other problems brought by network communication,such as delay,packet dropout and so on.These problems will have a great impact on NCSs,and are of great significance in designing new control strategies in both theory and practice.In this thesis,model-based event-triggered dynamic quantized control strategies are studied for these problems,and the network delay and system noise are studied in depth.First,a model-based event-triggered quantized feedback control strategy is proposed as the basis of the whole thesis.In order to extend the inter-sampling time,we adopt a model-based event-triggered control strategy which estimates the system states and calculates the control input with a nominal model of the system plant between two sampling instants.In order to ensure the asymptotic stability of the system at a finite quantization bit rate,a dynamic uniform quantizer is adopted.By further taking advantage of event-triggering strategies,a uniform surface quantizer is finally applied.As for the encoding and decoding problems brought by the dynamic quantizer,the system state based triggering threshold is replaced by a model state based one.In order to ensure that the state error does not cause the system to immediately trigger the event again after updating,a sufficient quantization bit number condition is provided.It is proved that there exists a positive minimum inter-sampling time for systems with the event-triggered control strategy provided above,indicating that this strategy will not lead to "Zeno-behavior".The effectiveness of the strategy is confirmed by simulations.Secondly,the influence of bounded channel delay is quantitatively considered in the above event-triggered quantized feedback control systems.In order to make the model states of sensor and controller equal after updating,the "time-stamp"information is adopted,included in every network packet,to update the controller state.With the time-stamp information,the controller can obtain an accurate channel delay and make an appropriate compensation for updating.In addition,the event-triggering condition is modified according to the network delay,and the corresponding lower bound of quantization bits is given.Furthermore,in order to ensure the asymptotic stability of the systems,two constraints are proposed for the channel delay,and the upper bound of tolerable delay is provided based on these two conditions.The simulation results verify the effectiveness of the strategy.Finally,the effects of both system noise and channel delay are considered in model-based event-triggered quantized feedback systems.For the purpose of maintaining that there is a positive minimum inter-sampling time for the system in the presence of bounded system noise,a mixed event-triggering threshold is adopted,and a corresponding lower bound of quantization bits is also provided.An upper bound of the channel delay to ensure the stability of the system under the bounded system noise is given and verified by simulations.