Mani M. Tousi

Brief paper: Optimal hybrid fault recovery in a team of unmanned aerial vehicles

This paper introduces and develops an optimal hybrid fault recovery methodology for a team of unmanned vehicles by taking advantage of the cooperative nature of the team to accomplish the desired mission requirements in presence of faults/failures. The proposed methodology is developed in a hybrid framework that consists of a low-level (an agent level and a team level) and a high-level (discrete-event system level) fault diagnosis and recovery modules. A high-level fault recovery scheme is proposed within the discrete-event system (DES) supervisory control framework, whereas it is assumed that a low-level fault recovery designed based on classical control techniques is already available. The low-level recovery module employs information on the detected and estimated fault and modifies the controller parameters to recover the team from the faulty condition. By taking advantage of combinatorial optimization techniques, a novel reconfiguration strategy is proposed and developed at the high-level so that the faulty vehicles are recovered with minimum cost to the team. A case study is provided to illustrate and demonstrate the effectiveness of our proposed approach for the icing problem in unmanned aerial vehicles, which is a well-known structural problem in the aircraft industry.

A hybrid fault diagnosis and recovery for a team of unmanned vehicles

In this paper, a hybrid fault detection, isolation, and recovery (FDIR) methodology is developed for a team of unmanned vehicles which takes advantage of the cooperative nature of the system to accomplish the desired mission in presence of failures. The proposed methodology is hybrid and consists of a low level (agent level) and a high level (team level) FDIR. The high level FDIR is formulated in the discrete-event system (DES) supervisory control framework, whereas the low level FDIR uses the classical control techniques. By properly integrating the two FDIR components, a larger class of faults can be detected and isolated when compared to results in the literature. A reconfiguration strategy is also designed so that the team is recovered from faults. Simulation results are provided to elucidate the efficacy of the proposed approach.

A distributed and cooperative supervisory estimation of multi-agent systems - Part I: Framework

In this work, we propose a framework for supervisory cooperative estimation of multi-agent linear time-invariant (LTI) systems. We introduce a group of sub-observers, each estimating certain states that are conditioned on given input, output, and state information. The cooperation among the sub-observers is supervised by a discrete-event system (DES) supervisor. The supervisor makes decisions on selecting and configuring a set of sub-observers to successfully estimate all states of the system. Moreover, when certain anomalies are present, the supervisor reconfigures the set of selected sub-observers so that the impact of anomalies on the estimation performance is minimized. This framework is applicable to any multi-agent system including large-scale industrial processes. In this paper (Part I), our proposed framework for supervisory estimation is developed based on the notion of sub-observers and DES supervisory control. In the companion paper (Part II), a DES-based combinatorial optimization method for selection of an optimal set of sub-observers is presented, the feasibility of the overall integrated sub-observers is validated, and the application of our proposed method in a practical industrial process is demonstrated through numerical simulations.

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

In this work, a novel framework for optimal cooperative supervisory estimation of multi-agent linear time-invariant (LTI) systems is proposed which is applicable to a large class of multi-agent systems. This framework was recently developed by the authors based on the notion of sub-observers and a discrete-event system (DES) supervisory control. Each sub-observer estimates certain states that are conditioned on given inputs, outputs, and states information. Moreover, the cooperation among the sub-observers is managed by a DES supervisor. In this work, our proposed supervisory estimation framework is extended to the combinatorial optimization domain. When certain anomalies (faults) are present in the system, or the sensors and sub-observers become unreliable, the proposed optimal DES supervisor makes decisions regarding the selection and reconfiguration of sets of sub-observers to estimate all the system states, while simultaneously a performance index that incorporates the communication cost, computation cost, and reconfiguration cost, and the number of invalid state estimates is minimized. The application of our proposed methodology in a practical industrial process is demonstrated through numerical simulations.

A distributed and cooperative supervisory estimation of multi-agent systems - Part II: Verification and case study

A framework for supervisory cooperative estimation in multi-agent linear time-invariant (LTI) systems is presented in the companion work (Part I). We introduced a set of sub-observers such that each estimates some states with a given set of input, output, and state information. A discerete-event system (DES) supervisory control framework is used for cooperation among the sub-observers. The supervisor selects a set of sub-observers to successfully estimate all states of the multi-agent system. In addition, in presence of a fault in the system, the supervisor reconfigures the set of selected sub-observers to minimize the fault impact on the estimation performance. Our general framework can be applied to any multi-agent system including industrial processes. In the companion paper (Part I), our proposed framework for the supervisory estimation is developed based on the notion of subobservers and DES supervisory control. In this paper (Part II), a DES-based combinatorial optimization method for selection of an optimal set of sub-observers is presented, the feasibility of the overall integrated sub-observers is validated, and the application of our proposed method in a practical industrial process is demonstrated through numerical simulations.

A hybrid fault detection and isolation strategy for a team of cooperating unmanned vehicles

In this paper, a hybrid fault detection and isolation (FDI) methodology is developed for a team of cooperating unmanned vehicles. The proposed approach takes advantage of the cooperative nature of the team to detect and isolate relatively low-severity actuator faults that are otherwise not detectable and isolable by the vehicles themselves individually. The approach is hybrid and consists of both low-level (agent/team level) and high-level [discrete-event systems (DES) level] FDI modules. The high-level FDI module is formulated in the DES supervisory control framework, whereas the low-level FDI module invokes classical FDI techniques. By properly integrating the two FDI modules, a larger class of faults can be detected and isolated as compared to the existing techniques in the literature that rely on each level separately. Simulation results for a team of five unmanned aerial vehicles are also presented to demonstrate the effectiveness and capabilities of our proposed methodology.

Decentralized pole-placement using generalized sampled-data hold functions

This paper presents a novel sequential technique for pole-assignment in linear time-invariant (LTI) decentralized control systems. Generalized sampled-data hold functions (GSHF) are used as local controllers to place the modes of the equivalent discrete-time closed-loop system in the desired locations in the z-plane. These locations are assumed to be obtained by using a proper mapping from the continuous-time domain. The GSHFs are obtained one at a time, in a sequential fashion. In other words, each local controller is designed for the equivalent discrete-time closed-loop model associated with the previously designed controllers. While no bound is provided on the intersample ripple, the convergence of the samples to zero ensures that the intersample values will also approach zero as time increases. The main characteristic of the proposed method is that unlike conventional pole-placement algorithms, the design complexity here does not increase after each local controller is obtained. A numerical example is provided which confirms the efficacy of the proposed pole-placement technique.

An application of control system in strategic management

The problem of productivity evaluation and systematic decision making in strategic management on an organization is tackled in this paper by employing control system methodologies. A novel approach is proposed to transform a performance evaluation problem in management domain into a fault diagnosis problem in control system domain. Then, a solution for the resulting fault diagnosis problem is provided by taking advantage of control system theories for fault detection and isolation.

Supervisory control of switching control systems

In this paper, the problem of designing a switching policy for an adaptive switching control system is formulated as a problem of supervisory control of a Discrete-Event System (DES). Two important problems in switching control are then addressed using the DES formulation and the theory of supervisory control under partial observation. First, it is verified whether for a given set of controllers, a switching policy satisfying a given set of constraints on the transitions among controllers exists. If so, then a minimally restrictive switching policy is designed. Next, an iterative algorithm is introduced for finding a minimal set of controllers for which a switching policy satisfying the switching constraints exists. It is shown that in the supervisory control problem considered in this paper, limitations on event observation are the factors that essentially restrict supervisory control. In other words, once observation limitations are respected, limitations on control will be automatically satisfied. This result is used to simplify the proposed iterative algorithm for finding minimal controller sets.