Hybrid gain performance‐based random event‐triggered filter of positive semi‐Markovian jump systems with intermittent sensor faults

Abstract

This article proposes a new random event-triggering scheme to design filter of positive semi-Markovian jump systems with intermittent sensor faults. To obtain a balance of saving network resources and improving system performance, a random event-triggering mechanism consisting of time-and event-triggering modes is constructed. A Bernoulli distribution-based stochastic variable is introduced to orchestrate the switching of two triggering modes. A concept of hybrid gain performance is presented to guarantee that the filter can attenuate the affect from the disturbance and sensor faults. Then, a random event-triggered filter is designed for positive semi-Markovian jump systems subject to intermittent sensor faults. Under the designed filter, hybrid gain performance is achieved. Furthermore, the proposed approach is extended to the systems with stochastically incorrect sensor measurement. A linear programming method is employed for describing and computing the presented conditions. Finally, the validity of the results is verified via two examples.

Hybrid dynamic event‐triggered fault control of positive Markovian jump systems

“…The dynamic ETM of positive systems is different from general systems. It was shown in References 12,24, and 26 that 1‐norm is more suitable for positive systems. Therefore, the 1‐norm is used in (6) and (7).…”

“…Remark The hybrid dynamic ETM has some advantages over the static ETM. In References 24, 26, and 27, linear event‐triggered ETMs were proposed for the fault detection and filter design of PMJSs. The event‐triggering conditions in References 24, 26, and 27 were dependent on a fixed constant.…”

“…In References 24, 26, and 27, linear event‐triggered ETMs were proposed for the fault detection and filter design of PMJSs. The event‐triggering conditions in References 24, 26, and 27 were dependent on a fixed constant. Such a static ETM reduces the flexibility of the design.…”

“…These new frameworks mainly contain stochastic CLFs and LP computation method, which have essential differences from general systems. The linear event‐triggered frameworks in References 24‐27 further illustrate the effectiveness of CLFs and LP in solving the corresponding issues of PMJSs. The dynamic ETM is dependent on a dynamic variable and has a better design flexibility with respect to the static ETM.…”

“…The static ETM is described in the form of inequality and it provides the relation between the state and the error of the sampling and the real states. [16][17][18] New linear frameworks were presented for the event-triggered fault detection, 24 event-triggered filter, 25,26 and event-triggered control 27 of PMJSs. These new frameworks mainly contain stochastic CLFs and LP computation method, which have essential differences from general systems.…”