Search

Benefiting from the notion of Fréchet derivatives, we define Fréchet discrete operators, such as gradient and Hessian, on infinite-dimensional spaces. The Fréchet discrete gradient expands upon the concept of the discrete gradient of Gonzalez (1996) for finite-dimensional spaces. The Fréchet discrete Hessian elevates the property to second-order representations of the Fréchet derivative. By leveraging these operators, we offer an initial exploration of discrete gradient methods for convex optimization in infinite-dimensional spaces. Under mild conditions on the objective functional, we establish the convergence of any sequence generated by the proposed Fréchet discrete gradient method, regardless of the choice of the finite learning rate.

Control of a platoon of connected vehicles with nonlinear longitudinal dynamics subject to failure in receiving communicated and radio data is addressed in this paper. We consider a scenario when due to malfunction of communication and radio devices, required data for cooperative adaptive cruise control may be stochastically unavailable. We develop a control scheme such that under certain conditions, the regulation of the intervehicle distances in desired values can be guaranteed. Specifically, we rigorously show that if the probability of receiving the required data for each vehicle is not zero, then under the proposed control strategy, the intervehicle distances almost surely converge to the desired values. Accurate performance of radars is a key assumption in the existing results in the literature on cooperative adaptive cruise control. Hence, the main contribution of this paper is that under the proposed control strategy, the robustness of the platoon against stochastic data loss in both communication network and vehicles radars is guaranteed. Simulation results verify the acceptable performance of the proposed control strategy.

Cooperative adaptive cruise control (CACC) in leaderless scenario is addressed in this article. The common idea in CACC is that a leader vehicle determines the steady-state speed of the platoon without considering feedback from other vehicles. The main contribution of this article is to propose a control strategy such that each vehicle in a leaderless platoon keeps a desired distance from its preceding vehicle, determined based on the constant time-gap spacing policy, and simultaneously, all the vehicles reach consensus upon their speeds. Moreover, whereas each vehicle needs access to the information of some neighboring vehicles (via a communication network) to achieve speed consensus, we assume that it has access to the relative distance of just one possible preceding vehicle (using a radar/lidar). Simulation results illustrate the performance of the proposed control strategy.

We develop a small-gain theory for systems described by set-valued maps between topological spaces. We introduce an abstract notion of stability unifying the continuity properties underlying different existing concepts, such as Lyapunov stability of equilibria, sets, or motions, (incremental) input–output stability, asymptotic gain properties, and continuity with respect to fast-switching inputs. Then, we prove that a feedback interconnection enjoying a given abstract small-gain property is stable. While, in general, the proposed small-gain property cannot be decomposed as the union of stability of the subsystems and a contractiveness condition, we show that it is implied by standard assumptions in the context of input-to-state stable systems. Finally, we provide application examples illustrating how the developed theory can be used for the analysis of interconnected systems and design of control systems.

This work describes and compares two phase-locked-loop (PLL) algorithms aimed at tracking a biased sinusoidal signal with unknown frequency, amplitude, and phase, with inherent robustness to dc offset. The proposed methods endow quadrature PLLs, renowned for their excellent tracking performance, with frequency-adaptation capability, while providing robust global stability certificates. The large-gain global stability, proved by Lyapunov-like arguments borrowed from adaptive control theory, represents a major benefit when compared to the conventional PLLs, whose convergence instead can be proved only locally by small-signal analysis or small-gain assumptions. In this connection, the proposed algorithms represent the first frequency-adaptive and dc-bias rejecting PLL-type architectures with Lyapunov-certified global stability. When used for signal tracking, the proposed methods are shown to outperform the adaptive observer, especially in noisy conditions. Moreover, they provide more accurate frequency estimates than existent frequency-adaptive PLLs, showing enhanced robustness in facing both phase-noise and measurement perturbations.

We discuss passivity beyond continuous-time systems by highlighting several inconsistencies of the definition currently in use, for instance when applied to discrete-time systems, and by motivating the need for a new and more consistent notion. Hence, we propose a new definition, ϱ-passivity, that addresses the raised issues and that, as a result, is naturally applicable to a larger class of systems including discrete-time systems with non-zero relative degree. We show that, in line with the classical definition, ϱ-passivity is preserved under parallel and negative-feedback interconnection. These are preliminary results aimed at bridging passivity between different time domains, and taking the first step towards a more comprehensive and consistent passivity theory.

Methods based on Reproducing Kernel Hilbert Spaces (RKHS) have proven to be a valuable tool for the identification of linear time-invariant systems in both discrete- and continuous-time. In particular, unlike most other techniques, they enable to systematically confer a priori desirable properties, such as stability, on the estimated models. However, existing RKHS methods mainly target impulse responses and, hence, do not extend well to the context of nonlinear systems. In this work, we propose a novel RKHS-based methodology for the identification of discrete-time nonlinear systems guaranteeing that the identified system is incrementally input-to-state stable (δISS). We model the identified system using a predictor function that, given past input and output samples, yields the output prediction at the next time instant. The predictor is selected from an RKHS by solving a constrained optimization problem that guarantees its δISS properties. The proposed approach is validated via numerical simulations.

This paper studies the role of a class of node systems with adjustable damping coefficients in the node dissipation inequalities, called the active nodes, for achieving passivity and finite-gain stability of network systems with external inputs and outputs. The paper first discusses the relation between the active nodes with passivity-type dissipation inequalities and those with L2-stability-type dissipation inequalities, especially focusing on how to transform one type into the other. Then, the paper proceeds to discuss the graph-theoretic condition on the locations of active nodes to achieve network passivity or finite-gain stability. Thereafter, a method for network reduction by exploiting active nodes is presented. Finally, a human-in-the-loop stabilization problem is solved for a network system by exploiting the results developed in the paper.

This paper proposes a passive methodology for detecting a class of stealthy intermittent integrity attacks in cyber-physical systems subject to process disturbances and measurement noise. A stealthy intermittent integrity attack strategy is first proposed by modifying a zero-dynamics attack model. The stealthiness of the generated attacks is rigorously investigated under the condition that the adversary does not know precisely the system state values. In order to help detect such attacks, a backward-in-time detection residual is proposed based on an equivalent quantity of the system state change, due to the attack, at a time prior to the attack occurrence time. A key characteristic of this residual is that its magnitude increases every time a new attack occurs. To estimate this unknown residual, an optimal fixed-point smoother is proposed by minimizing a piece-wise linear quadratic cost function with a set of specifically designed weighting matrices. The smoother design guarantees robustness with respect to process disturbances and measurement noise, and is also able to maintain sensitivity as time progresses to intermittent integrity attack by resetting the covariance matrix based on the weighting matrices. The adaptive threshold is designed based on the estimated backward-in-time residual, and the attack detectability analysis is rigorously investigated to characterize quantitatively the class of attacks that can be detected by the proposed methodology. Finally, a simulation example is used to demonstrate the effectiveness of the developed methodology.

This paper considers the isolability of sensor faults and malicious cyber attacks in cyber-physical systems. Attacks and faults are characterized by their basis functions, which are utilized for threat isolation and to characterize their isolability. An isolability metric is proposed that exploits the threat basis function to quantify the threat isolability. It is rigorously proved that a threat is non-isolable if and only if this metric is zero. A methodology to analytically calculate the proposed isolability metric is developed by solving a minimax optimization problem. Our findings are validated with numerical simulations.

This article proposes a stealthy integrity attack generation methodology for a class of nonlinear cyber-physical systems. Geometric control theory and stability theory of incremental systems are used to design an attack generation scheme with stealthiness properties. An attack model is proposed as a closed-loop dynamical system with an arbitrary input signal. This model is developed based on a controlled invariant subspace that results from geometric control theory and is decoupled with the system outputs and the nonlinear function. The presence of the arbitrary signal in the attack model provides an additional degree of freedom and constitutes a novel component compared with existing results. The stealthiness of the attack model is rigorously investigated based on the incremental stability of the closed-loop control system, and the incremental input-to-state stability of the anomaly detector. As a result, a sufficient condition in terms of the initial condition of the attack model is derived to guarantee stealthiness. Finally, a case study is presented to illustrate the effectiveness of the developed attack generation scheme.

“An exciting journey has started for me!” In this way, I began my inaugural “President’s Message” column [1].

Welcome to my second column of 2022, following the exciting “President’s Message” invited column by Bob Shorten (Imperial College, United Kingdom) that dealt with the potential impact of control theory on society, highlighting the challenges in front of us (see the April 2022 issue of IEEE Control Systems).

This paper proposes a sensor additive switching watermark methodology for detecting physical faults and replay attacks, and then identifying the occurring threat type: either the physical fault or the replay attack. The sensor watermark methodology includes a switching watermark generator in the plant side and a switching watermark remover in the control and monitoring side. The switching protocols of the generator and the remover and their common watermark seeds are specifically designed, guaranteeing the switch synchronization in both the nominal and fault cases, and allowing the switch asynchronization in the replay attack case. The latter is used for discriminating the two considered threat cases. The detectability and discrimination ability of the proposed watermark approach, characterizing the class of attacks and faults that can be detected and discriminated, is rigorously investigated. A simulation example is presented to illustrate the effectiveness of the proposed approach.

The problem of detecting replay attacks in linear time-invariant discrete-time systems is considered in this paper. In the same spirit of watermarking techniques that apply a distinctive signature to the plant\x92s signals, we propose an event-triggered control scheme, that is purposely designed to generate a unique sequence of switching intervals, by computing an appropriate input value to be held constant while the communication is not triggered. We provide a detailed undetectability characterization in the time domain and design a controller that achieves the desired behavior. Our proposed method results in a control scheme that makes it hard for an attacker to satisfy undetectability conditions, and, as a result, a standard observer-based residual generator can be employed to reveal replay attacks. We finally validate the method using a numerical example.

Presents changes to the CSS Constitution and Bylaws.

Presents the President’s message for this issue of the publication.

The role of laboratory activities in engineering education is fundamental to enhance the students engagement and learning. In this paper we discuss the use of a laboratory test-bed for education and research developed to replicate the dynamic behavior and the control design challenges of an underactuated multirotor Unmanned Aerial Vehicle (UAV), which is becoming a game-changing technology in several application fields. In view of recent challenges related to online teaching activities, the setup has been endowed with a remotization kit to allow students to access the system from anywhere with an internet connection and to perform experiments with satisfactory results. The experience gathered in integrating laboratory activities with the proposed test-bed into the undergraduate Control Systems module at Imperial College London is reported. The paper presents the development and organization of the remote laboratory lectures and provides an overview of the obtained results.

Presents the President’s message for this issue of the publication.

This paper focuses on developing a distributed leader-following fault-tolerant tracking control scheme for a class of high-order nonlinear uncertain multiagent systems. Neural network-based adaptive learning algorithms are developed to learn unknown fault functions, guaranteeing the system stability and cooperative tracking even in the presence of multiple simultaneous process and actuator faults in the distributed agents. The time-varying leader's command is only communicated to a small portion of follower agents through directed links, and each follower agent exchanges local measurement information only with its neighbors through a bidirectional but asymmetric topology. Adaptive fault-tolerant algorithms are developed for two cases, i.e., with full-state measurement and with only limited output measurement, respectively. Under certain assumptions, the closed-loop stability and asymptotic leader-follower tracking properties are rigorously established.

The paper proposes a methodology to effectively address the increasingly important problem of distributed fault-tolerant control for large-scale interconnected systems. The approach dealt with combines, in a holistic way, a distributed fault detection and isolation algorithm with a specific tube-based model predictive control scheme. A distributed fault-tolerant control strategy is illustrated to guarantee overall stability and constraint satisfaction even after the occurrence of a fault. In particular, each subsystem is controlled and monitored by a local unit. The fault diagnosis component consists of a passive set-based fault detection algorithm and an active fault isolation one, yielding fault-isolability subject to local input and state constraints. The distributed active fault isolation module–thanks to a modification of the local inputs–allows to isolate the fault that has occurred, avoiding the usual drawback of controllers that possibly hide the effect of the faults. The Active Fault Isolation method is used as a decision support tool for the fault-tolerant control strategy after fault detection. The distributed design of the tube-based model predictive control allows the possible disconnection of faulty subsystems or the reconfiguration of local controllers after fault isolation. Simulation results on a well-known power network benchmark show the effectiveness of the proposed methodology.

Control of a platoon of vehicles subject to jamming attacks is addressed in this paper. Because of jamming attacks, some communicated information and radio data are assumed to be lost or delayed in a stochastic manner. By considering the constant time-gap spacing policy, we propose a control strategy which under certain conditions guarantees the almost sure regulation of the vehicles in desired relative distances. Accordingly, depending on the control gains, the robustness of the platoon against a wide range of jamming attacks is guaranteed. The main contribution of the paper is that the proposed control scheme is robust against jamming attacks on both the communication network and vehicles radars. Simulation results illustrate the performance of the proposed control strategy.

State estimation for a class of nonlinear systems using a network of distributed observers is dealt with in this paper. We propose a network of observers, each of which has its own local measurements which may not be sufficient for observability of the system, while the joint measurements of the sensors in the network guarantees the observability. We derive sufficient conditions on the proposed observers such that the estimated states of each observer asymptotically converge to the states of the system using local measurements and the estimates of a subset of observers in its neighborhood. Accordingly, it is assumed that each observer has access to the estimates of its neighbors via a communication network. The main problem of existing distributed state estimation strategies in the literature is their limitation to linear systems or to nonlinear systems with observable nonlinearities (by considering nonlinearities as states to be estimated). However, based on the proposed strategy, distributed state estimation of a more general class of nonlinear systems can be devised. A numerical example is provided to evaluate the accuracy of the obtained results.

The detection of stealthy integrity attacks for nonlinear cyber-physical systems is a great challenge for the research community. This paper proposes a backward-in-time detection methodology to enhance the anomaly detector against stealthy integrity attacks for a class of nonlinear cyber-physical systems. It uses the virtual value of the state at a time instant prior to the occurrence time of the attacks for detecting stealthy attacks. The definition of stealthy integrity attacks is formulated in the context of nonlinear plants such that they are undetectable with respect to traditional anomaly detectors. A H∞fixed-point smoother is developed for estimating the analytical virtual values of the states at a prior time to the attack occurrence time, and then, the backward-in-time detection schemes are proposed based on the smoother. Based on the prior estimates, attack residual generation and threshold generation schemes are designed. Finally, a simulation is presented to illustrate the effectiveness of the enhanced anomaly detector.

This paper deals with a topology for a class of interconnected systems, referred to as a highly interconnected system, consisting of interconnected plants and local controllers. We address the respective cyber attack surfaces as well as a design approach for detection and isolation of covert cyber attacks. For each pair of plant and controller, a cyber attack is implemented by a malicious agent, and its detection and isolation are achieved by associating the controller with two observers. These observers estimate the states of the plant, and compare the estimated states to determine if a neighbouring plant is under a covert cyber attack. The paper presents the modelling of the topology, the analysis of the covertness of cyber attacks, the design approach for the detection and isolation as well as a required existence condition. Simulation results are provided for the application of the design approach to interconnected pendula systems that are subject to a covert cyber attack.

This paper presents an algorithm for estimating the parameters of a biased sinusoidal signal. The proposed method uses the output signals of a second order generalized integrator without adaptation on its resonant frequency to derive a linear regression equation where the unknown parameters are a nonlinear combination of bias and frequency of the input signal. The global stability of the method is proven. Remarkably, the proposed method represents the minimum-order estimator known for the problem under consideration, being implementable by a 4th-order adaptive system. Simulation results and comparisons with existing methods show the accurate estimation capability of the proposed approach.

The paper deals with the key problem of detecting cyber-attacks in the context of large-scale systems (LSS). As these systems grow in size and complexity, cyber-attacks may target limited parts of the LSS, leading to tackling the problem in a decentralized way. We analyze the properties of distributed detection schemes under both local and interconnection attacks, and show that they are vulnerable to attacks that exploit the structure of the interconnections between subsystems. We also provide conditions and strategies that may be adopted to make the distributed control architecture regulating the LSS robust to this class of cyber-attacks.

In this paper, we propose a novel distributed fault detection method to monitor the state of a—possibly large scale—linear system, partitioned into interconnected subsystems. The approach hinges on the definition of a partition-based distributed Luenberger-like estimator, based on the local model of the subsystems and that takes into account their dynamic coupling. The proposed methodology computes—in a distributed way—a bound on the variance of a properly defined residual signal. This bound depends on the uncertainty affecting the state estimates computed by the neighboring subsystems and it allows the computation of local fault detection thresholds, as well as the maximum false-alarm rate. The implementation of the proposed estimation and fault detection method is scalable, allowing Plug & Play operations, and the possibility to disconnect the faulty subsystem after fault detection. Theoretical conditions on the convergence properties of the estimates and of the estimation error bounds are provided. Simulation results on a power network benchmark show the effectiveness of the proposed method.

The problem of rejection of a sinusoidal disturbance with unknown frequency acting on an unknown single-input-single-output stable linear system is addressed in this paper. We present a new hybrid approach that does not require knowledge of the frequency response of the transfer function over the range of frequencies of interest. The proposed controller reposes upon a switching strategy within a family of linear controllers based on the adaptive feedforward / internal model control methodology. A dead-beat frequency estimation method is integrated in the controller. The method also accounts for the presence of bounded sensor noise as well as imprecise frequency estimation. It is shown that, within a finite number of switchings, the regulation error is ultimately bounded by a function of the norm of the noise that depends on the choice of the controller and the estimator gains.

Presents the President's message for this issue of the publication.

Welcome to my third column of 2022. It follows the great “President’s Message” invited column by Giovanni Cherubini (IBM Research Zurich, Switzerland) that focused on the impact on big data storage systems of advanced control techniques for magnetic tape drives (see the August 2022 issue of our magazine [1]). Here, I would like to build a virtuous future connection between the Outreach Program of the IEEE Control Systems Society (CSS) and the “new normal” of our conferences I alluded to in my June 2021 “President’s Message” column [2].

The problem of replay attacks in the communication network between Distributed Generation Units (DGUs) of a DC microgrid is examined. The DGUs are regulated through a hierarchical control architecture, and are networked to achieve secondary control objectives. Following analysis of the detectability of replay attacks by a distributed monitoring scheme previously proposed, the need for a watermarking signal is identified. Hence, conditions are given on the watermark in order to guarantee detection of replay attacks, and such a signal is designed. Simulations are then presented to demonstrate the effectiveness of the technique.

In this preliminary work, we consider the problem of detecting cyber-attacks in a linear system equipped with a Model Predictive Controller, where the feedback loop is closed over a non-ideal network, and the process is subject to a random Gaussian disturbance. We adopt a model-based approach in order to detect anomalies, formalizing the problem as a binary hypothesis test. The proposed approach exploits the analytical redundancy obtained by computing partially overlapping nominal system trajectories over a temporal sliding window, and propagating the disturbance distributions along them. The recorded data over such window is then used to define a probabilistic consistency index at each time step in order to make a decision about the presence of possible attacks. Preliminary simulation results show the effectiveness of the proposed attack-detection method.

The paper deals with the problem of defining the optimal topology for a distributed fault detection architecture. A partition-based distributed fault detection method is considered based on the model of the system. The system is decomposed into subsystems and each subsystem is monitored by one local diagnoser. Non-overlapping decompositions are considered. A novel cost function is proposed to measure the detectability properties of a distributed fault detection method, depending on the topology of the detection framework. Different objective functions are taken into account and compared in order to analyze the influence of the decomposition on fault detection performance. Preliminary numerical results show that the minimization of the coupling between subsystems could not be always the best choice for the fault detection performance, and that the proposed cost function minimization allows the reduction of the detection time with the considered fault detection method.

In this paper, we first analyze the possible limitations of a model-based fault detection method grounded on a partition-based distributed Luenberger observer. The corresponding fault detection test consists of comparing, for each time instant, the output prediction error with a suitable bound, computed analytically in a distributed and scalable way As a result, we highlight the presence of an often restrictive tradeoff between false-alarm and missed-detection rates. To overcome this significant drawback, we resort to a method based on the analysis of moving averages of residuals. Tests on an academic case study show the effectiveness of this approach.

In this paper, a distributed adaptive fault-tolerant control (FTC) scheme for a class of nonlinear uncertain input-output multi-agent systems is developed. A decentralized local nonlinear observer is employed for estimating the state variables of each agent, and a distributed local FTC component is designed for each agent in the distributed system by utilizing local output measurements and state estimation information exchanged between neighboring agents. The time-varying leader’s command is only communicated to a small portion of follower agents through directed links, and each follower agent exchanges measurement information only with its neighbors through a bidirectional but asymmetric topology. Under certain assumptions, the closed-loop stability and leader-follower consensus properties of the distributed system are rigorously established.

The problem of deadbeat state reconstruction for non-autonomous linear systems has been solved since several decades, but all the architectures formulated since now require either high-gain output injection, which amplifies measurement noises (e.g., in the case of sliding-mode observers), either state augmentation, which yields a non-minimal realization of the deadbeat observer (e.g., in the case of integral methods and delay-based methods). In this context, the present paper presents, for the first time, a finite-time observer for continuous-time linear systems enjoying minimal linear-time-varying dynamics, that is, the observer has the same order of the observed system. The key idea behind the proposed method is the introduction of an almost-always invertible time/output-dependent state mapping which allows to recast the dynamics of the system in a new observer canonical form whose initial conditions are known.

The paper presents a distributed Sensor Fault Diagnosis architecture for Industrial Wireless Sensor Networks monitoring HVAC systems, by exploiting a recently proposed distributed clustering method. The approach allows the detection and isolation of multiple sensor faults and considers the possible presence of modeling uncertainties and disturbances. Detectability and isolability conditions are provided. Simulation results show the effectiveness of the proposed method for an HVAC system.

This paper presents a novel observer-based method for estimating the parameters of an exponentially damped sinusoid, including the frequency, amplitude, phase and the damping factor. A second-order sliding-mode-based adaptation law is designed for estimating the frequency and the damping factor, while amplitude and phase are obtainable from the said parameters by straightforward algebra. The stability and robustness analysis in presence of an additive disturbance proves the existence of a tuning parameter set for which the estimator’s dynamics are asymptotic stable and the estimation error asymptotically converges to a residual set whose size depends on the amplitude of the disturbance. Numerical and comparative examples show the effectiveness of the proposed estimation approach.

In this paper, a novel framework is proposed for deadbeat distributed Fault Detection and Isolation (FDI) of large-scale continuous-time LTI dynamic systems. The monitored system is composed of several subsystems which are linearly interconnected with unknown parameterization. Each subsystem is monitored by a local diagnoser based on the measured local output, local inputs and the interconnection variables from the neighboring subsystems. The local FDI decision is based on two non-asymptotic state-parameter estimators using Volterra integral operators which eliminate the effect of the unknown initial conditions so that the estimates converge to the true value in a deadbeat manner and therefore the fault diagnosis can be achieved in finite time. Moreover, the unknown interconnection parameters and the unknown fault parameters are simultaneously estimated. Numerical examples are included to show the effectiveness of the proposed FDI architecture.

This paper develops an integrated filtering and adaptive approximation-based approach for fault diagnosis of process and sensor faults in a class of continuous-time nonlinear systems with modeling uncertainties and measurement noise. The proposed approach integrates learning with filtering techniques to derive tight detection thresholds, which is accomplished in two ways: 1) by learning the modeling uncertainty through adaptive approximation methods and 2) by using filtering for dampening measurement noise. Upon the detection of a fault, two estimation models, one for process and the other for sensor faults, are initiated in order to identify the type of fault. Each estimation model utilizes learning to estimate the potential fault that has occurred, and adaptive isolation thresholds for each estimation model are designed. The fault type is deduced based on an exclusion-based logic, and fault detectability and identification conditions are rigorously derived, characterizing quantitatively the class of faults that can be detected and identified by the proposed scheme. Finally, simulation results are used to demonstrate the effectiveness of the proposed approach.

This paper deals with the problem of modelling a chemical reactor for the Direct Reduction of Iron ore (DRI). Such a process is being increasingly promoted as a more viable alternative to the classic Blast Furnace for the production of iron from raw minerals. Due to the inherent complexity of the process and the reactor itself, its effective monitoring and control requires advanced mathematical models containing distributed-parameter components. While classical approaches such as Finite Element or Finite Differences are still reasonable options, for accuracy and computational efficiency reasons, an algebraic approach is proposed. A full multi-physical, albeit one-dimensional model is addressed and its accuracy is analysed.

This paper proposes a novel Plug-and-Play (PnP) dynamic approach for the monitoring of Large-Scale Systems (LSSs). The proposed architecture exploits a distributed Fault Detection and Isolation (FDI) methodology for nonlinear LSS in a PnP framework. The LSS consists of several interconnected subsystems and the designed FDI architecture is able to manage plugging-in of novel subsystems and un-plugging of existing ones. Moreover, the proposed PnP approach performs the unplugging of faulty subsystems in order to avoid the propagation of faults in the interconnected LSS. Analogously, once the issue has been solved, the disconnected subsystem can be re-plugged-in. The reconfiguration processes only involves local operations of neighboring subsystems, thus allowing a distributed architecture.

The paper deals with the problem of defining the optimal topology for a distributed fault detection architecture for non-linear large-scale systems. A stochastic modelbased framework for diagnosis is formulated. The system structural graph is decomposed into subsystems and each subsystem is monitored by one local diagnoser. It is shown that overlapping of subsystems allows to improve the detectability properties of the monitoring architecture. Based on this theoretical result, an optimal decomposition design method is proposed, able to define the minimum number of detection units needed to guarantee the detectability of certain faults while minimizing the communication costs subject to some computation cost constraints. An algorithmic procedure is presented to solve the proposed optimal decomposition problem. Preliminary simulation results show the potential of the proposed approach.

This paper presents an adaptive fault-tolerant control (FTC) scheme for a class of nonlinear uncertain multi-agent systems. A local FTC scheme is designed for each agent using local measurements and suitable information exchanged between neighboring agents. Each local FTC scheme consists of a fault diagnosis module and a reconfigurable controller module comprised of a baseline controller and two adaptive fault-tolerant controllers activated after fault detection and after fault isolation, respectively. Under certain assumptions, the closed-loop system's stability and leader-follower consensus properties are rigorously established under different modes of the FTC system, including the time-period before possible fault detection, between fault detection and possible isolation, and after fault isolation.