Dynamic Event-Triggered-Based Asynchronous Controland Filtering For Singular Semi-markov Jump Systems

In engineering,dynamic systems are often affected by internal and external environments such as sudden failures of internal components,sudden changes of environments and random malicious attacks.These factors will cause sudden changes in system structure and parameters.As a kind of special hybrid systems,Markov jump systems(MJSs)can be used to describe this kind of dynamic systems.Compared with MJSs,the more general semi-Markov jump systems(S-MJSs)have greater advantages in modeling engineering practice,whose sojourn-time follows more conventional distributions such as Gaussian distribution and Weber distribution.Therefore,the comprehensive analysis of S-MJSs has aroused extensive research by scholars.In addition,singular systems have become another research hotspot,which are composed of slow subsystems and fast subsystems.And dynamic characteristics of slow subsystems are described by differential equations or difference equations,while the static characteristics of fast subsystems are described by algebraic equations.Generally speaking,singular systems are more representative of the actual systems than the general ones,so they can be widely used in modeling of aerospace,economic systems,neural network systems,etc.With the demand of practical engineering,the simple model can no longer meet the demand of industrial modeling.Singular semi-Markov jump systems(SS-MJSs)have been paid more attention because they can describe more features of practical physical systems.In recent years,in order to realize the convenient interaction between human and the physical world,Cyber-Physical Systems(CPSs)came into being.However,due to the deep integration between the information systems and the physical systems,CPSs are vulnerable to various covert attacks from adversaries,which will lead to the system performance degradation.The most common cyber-attacks are denial of service(Do S)attacks and deception attacks.At the same time,theoretical analysis and experiments have shown that dynamic event-triggered transmission strategies and dynamic-memory event-triggered transmission strategies can effectively reduce data transmission rates and computational burden.Therefore,by using Lyapunov function theory and linear matrix inequalities(LMIs)technique,the analysis of control and filtering for SS-MJSs under the cyber-attacks and dynamic event-triggered transmission strategies,becomes the main research content of this paper.Firstly,the problem of dissipative asynchronous control for singular stochastic semiMarkov jump systems(SSS-MJSs)with general conditional probabilities under multiple cyber-attacks is investigated.A unified framework is constructed to formulate the occurrences of deception attacks and Do S attacks.Considering the effects of the dynamicmemory event-triggered transmission strategy and multiple cyber-attacks,a new type of SSS-MJSs model is fairly established.Since the mode information of the SSS-MJSs can not be obtained accurately,the fact of asynchronous patterns between the initial SSS-MJSs and the designed asynchronous controller is intentionally expressed as a hidden Markov model(HMM).Furthermore,an integral inequality is well constructed to further decrease the conservatism of the system,which is regarded as a generalization of Wirtinger-based inequality.Then,based on the technique of LMIs,the criteria of exponential stable with certain dissipative performance for the resulting closed-loop SSS-MJS are provided.Lastly,a practical example of the catalytic cracking process system is provided to verify the effectiveness and practicability of the developed approach.Then,the design of asynchronous dissipative filter for SS-MJSs with general conditional probabilities under multi-channel deception attacks is investigated.To better depict the practical communication situations,the mutually independent random variables that subject to the Bernoulli distribution are adopted to formulate the occurrence of deception attacks.Considering the effects of the dynamic event-triggered transmission strategy and multi-channel deception attacks,the fact of asynchronous patterns between the initial SS-MJSs and the designed asynchronous filter is intentionally expressed as an HMM.Then,based on the technique of LMIs,the design method of asynchronous dissipative filter is given.Finally,a numerical example is provided to illustrate the effectiveness of the developed approach.