Multi-agent methodology for distributed and cooperative supervisory estimation subject to unreliable information

Abstract

In this work, a novel multi-agent framework for cooperative supervisory estimation of linear time-invariant (LTI) systems is proposed. This framework is developed based on the notion of sub-observers and a discrete-event system (DES) supervisory control and is applicable to a large class of systems. We introduce a group of sub-observers where each sub-observer is estimating certain states that are conditioned on a given input, output, and state information. The cooperation among the subobservers is managed by a DES supervisor. The supervisor makes decisions regarding the selection and configuration of a set of sub-observers to successfully estimate all the system states, while the feasibility of the overall integrated cooperative sub-observers is verified. When certain anomalies (faults) are present in the system, or the sensors and sub-observers become unreliable, the supervisor reconfigures the set of selected sub-observers so that the impacts of anomalies on the estimation performance are minimized to the extent that is possible. The application of our proposed methodology in a practical industrial process is demonstrated through numerical simulations.

Optimal distributed and cooperative supervisory estimation of multi-agent systems subject to unreliable information

“…The state space representation of the system is given by the triplet (A, B, C) where we have For purpose of conducting simulations we assume that o 1 = o 2 = o 3 = u 4 = u 5 = u 6 = u 7 = 0 in order to better demonstrate the effects of an uncertainty (anomaly, fault, or unreliability) on the system. Using the procedures proposed in [13], we design a set of seven sub-observers as follows: Let us begin by transforming the set of sub-observers into the DES framework. According to discussions in Section IVa, the DES event sets are as follows…”

“…It should be noted that designing the sub-observers for the LTI system S in equation (1) and investigating the convergence properties of the overall cooperative set of subobservers have already been developed and investigated in detail in [13] and is beyond the scope of this work. Next, we describe the DES models of the sub-observers and present our optimal supervisory estimation strategy.…”

“…1) Based on the multi-agent system and its configuration characteristics, generate certain sub-observers according to the methodology presented in [13]. 2) Model the constructed sub-observers in the DES framework according to the procedure that is described in Section IV.…”

“…In this work, we propose an optimal supervisory cooperative estimation framework and solution for multi-agent linear time-invariant (LTI) systems in presence of process and/or sensor faults. Our cooperative supervisory estimation approach was recently proposed in [13], in which a group of sub-observers are designed and their cooperation is supervised by a DES. However, in case that certain anomalies (faults) are present, the DES supervisor only checks the existence of a set of sub-observers to successfully estimate all the system states.…”

Distributed Kalman Filtering

“…Decomposition of a linear process model into a cascade of simpler subsystems is given in [282]. Other applications can be seen in [7], [338], [247], [248], [268], [269], [270], [271], [273], [275], [283], [284], [299], [321], [323], and [324] respectively. …”