Decentralized control of network‐based interconnected systems: A state‐dependent triggering method

Abstract

This paper investigates decentralized control for a class of interconnected system. Different from the traditional systems, the considered system has the following features: (i) its subsystems are connected through a communication network subject to transmission delays and packet losses; (ii) the subsystems' multi-actuators are subjected to random faults; and (iii) the subsystems are subject to probabilistic nonlinear disturbances, the inner variation information of the nonlinearities, as well as their bounds information, is utilized to analyze the nonlinearities. Furthermore, in order to reduce the network bandwidth burden, a decentralized state-dependent triggering scheme is proposed. Considering aforementioned characteristics and using the state-dependent triggering scheme, new type of network-based interconnected system model is built. By using the Lyapunov functional approach, sufficient conditions for the mean square stability and stabilization of the network-based interconnected systems are obtained. Then reliable controllers, as well as the triggering matrices of the local subsystems, can be co-designed by using a cone complementary linearization algorithm. Two simulation examples are given to illustrate the effectiveness and application of the proposed method. 1127 decentralized reliable guaranteed cost control of interconnected system subject to actuator faults. In [13], a decentralized adaptive approximation design method is proposed for the fault tolerant control of a class of interconnected nonlinear system. However, when considering all these three issues at the same time, the related research is not enough thus still need further investigation.It is well known that limited communication capacity is one of the main sources of the system performance degradation in the networked control systems, which can deteriorate the first two issues described in the previous texts. As to the limited network bandwidth, the following two methods are usually used to reduce the network bandwidth load: (i) the first is to reduce the size of the data. For example, in [14,15], the authors investigate the quantized feedback stabilization of interconnected systems. Generally speaking, the quantized method can only reduce the size of the data while can not reduce the quantity of the data, then the throughput capacity of the network can not be eased effectively, especially when the number of subsystems is large. (ii) The second method is to reduce the quantity of the released data. State-dependent triggering scheme (SDTS) (also called eventtriggering scheme or self-triggering scheme) has been proved to be an efficient way to reduce the data transmission in the network [16][17][18][19][20], which can relieve the burden of network bandwidth occupation when compared with traditional periodic sampling method. In the periodic sampling scheme, the sampling interval is determined before the system is deployed and which is computed to ensure the system performance of the worst conditions. Consequently, the obtained sampling interva...

Decentralized event-triggering communication scheme for large-scale systems under network environments

“…If all state information can be measured at the same time by the sensor, the following centralized event-triggered communication (CETC) scheme proposed in [23] can be used to trigger the transmission…”

“…To satisfy the requirement of the multi-packet transmission and asynchronous measurements for the system under consideration, our main idea is to design a DETC to simulate a CETC, that is, decentralized implement of CETC for making use of the advantages of CETC and avoiding the disadvantages of traditional DETC. Compared with CETC adopted in [21,22], the requirement of signal-packet transmission among the distributed subsystems is no longer required; Compared with the existing DETC [17,23], the requirement of synchronized measurements among the distributed sensors are not necessary. Moreover, the state estimation and DET parameters heuristic algorithms presented in this work have the novel role to reduce the occupancy of the communication bandwidth.…”

“…However, the synchronized measurements among the local event-triggered communication schemes are also needed since the system's evolvement depends on other subsystem's real-time state. Therefore, although some theoretical results including those in [16,23,24] are important and have some significant technical merits, there are still major difficulties in their applications to a large-scale system under the network environments and with the multi-packet transmissions, asynchronous measurements and unknown controller gains. This issue motivates the present study.…”

“…Moreover, if the controller gains are not known a priori, the above methods cannot be used again. In [23], a state-dependent decentralized triggering method is presented for network-based interconnected systems with actuator faults. However, the synchronized measurements among the local event-triggered communication schemes are also needed since the system's evolvement depends on other subsystem's real-time state.…”

“…The second method is decentralized event-triggered communication (DETC) scheme [5,8,23]. Different from CETC, DETC depends on the local states to determine whether or not to transmitted the current sampled data.…”