DX'17:Papers with Abstracts

Papers
Abstract. Sequences of events describing the behavior and actions of agents or systems can be collected in several domains. An episode is a collection of events that occur in a given partial order. By performing a recognition of recurrent episodes in several sequences and comparing them, it is pos- sible to determine a pattern common to all the se- quences. In this paper, we propose an approach to recognize episodes that are common in a set of event sequences. The method described is applied to the automotive domain for learning diagnostic procedures.
Abstract. In Cyber-Physical Systems (CPSs), physical processes, computational resources and communi- cation capabilities are tightly interconnected. Traditionally, the physical components of a CPS are described by means of differential or difference equations, while the cyber components are modeled by means of discrete dynamics. Therefore, hybrid systems, that are heterogeneous dynamical sys- tems characterized by the interaction of continuous and discrete dynamics, are a powerful modeling framework to deal with CPSs. Motivated by the great importance of security issues for CPSs, we characterize the observability and diagnosability properties for hybrid systems in the general case where the available information may be corrupted by an external attacker. Then, as CPSs are found in a wide range of applications, we demonstrate how to estimate the continuous state by simulating two scenarios: the control of a Direct Current (DC) Microgrid, and the control of a network of Unmanned Aerial Vehicles (UAVs) cooperatively transporting a payload.
Abstract. This paper combines a residual-based diagnosis approach and an unsupervised anomaly detection method to develop a hybrid methodology for monitoring smart buildings for which complete models are not available. The proposed method combines data mining approach and model-based diagnosis to update a diagnosis reference model and improve the overall diagnostics performance. To estimate the likelihood of each potential fault in complex systems like smart buildings, the dependencies between components and, there- fore, the monitors should be considered. In this work, a tree augmented naive Bayesian learning algorithm (TAN) is used for the classification. We demonstrate and validate the proposed approach using a data-set from an outdoor air unit (OAU) system in the Lentz public health center in Nashville.
Abstract. A broad range of real-world systems can be defined using discrete-time hybrid systems,
e.g., chemical process plants and manufacturing systems. We characterize this application domain using a class of discrete-event systems, max-plus linear discrete-event systems, which captures synchronization without concurrency or selection. The model framework of these hybrid systems is non-linear in a conventional algebra, but linear in the max-plus algebra, thereby enabling linear-time inference. We use an observer-based framework for monitoring and diagnosing max-plus diagnostics models, and further improve computational efficiency by searching over only the most-likely space of behaviours. We illustrate our approach using a chemical process-control example.
Abstract. This paper deals with a benchmark-based experimental comparison of three diagnoser-based approaches for fault diagnosis of discrete event systems modeled by Petri nets: the MBRG/BRD approach, the FMG/FMSG approach and the SSD approach. The experiments are performed on a level crossing benchmark, using the respective software tools integrating the approaches. Different features are shown in terms of state-space building (exhaustive or partial), procedure for analyzing diagnosability (based on complete or on-the-fly built state-space) and state-space representation (concrete or symbolic). Based on the obtained experimental results, a comparative discussion is provided particularly regarding memory and time consumption for analyzing diagnosability of the three techniques.
Abstract. In consistency-based diagnosis (CBD), abnormal behavior is sorted out based on de- viation from a normal behavior specification. Probabilities have been added to CBD for quantifying uncertainty on, e.g., the behavior of faulty components. While resulting in more complete models, the requirement of such uncertainty parameters goes in opposition to the original CBD motivation. The conflict measure stands closer to CBD by comput- ing solutions without the need of priors on candidates, however, its results might not be suitable when only partial observations are available. In this paper, we propose a method called the diagnostic coefficient, which better solves the partial observability case, while needing the same parameters as the conflict measure. The diagnostic coefficient is based on the idea that observations are conflicting if the observed outputs are discrepant with respect to alternative outputs that could have been observed. We report experiments with logical circuits where the diagnostic coefficient shows promising results compared to the conflict measure under various settings with missing observations.
Abstract. The twin plant method is central in every research whose focus is checking the diagnosability of discrete-event systems (DESs). Although the property of diagnosability has been extended over time, and several proposals have been advanced to perform a distributed analysis, diagnosability checking still relies on the exploitation of the twin plant method. However, the twin plant structure is redundant, which is a drawback, above all if the considered DES observation is uncertain: in such a case, several distinct twin plants have to be built in order to check the diagnosability for increasing levels of uncertainty. A higher uncertainty level requires a twin plant of larger size. The paper first gives some preliminary thoughts to the reduction of the twin plant size. Next, on the ground that no contribution in the literature has altered the original state-based representation of the twin plant, the paper shows how to transform such a representation into a transition-based one. Finally, it reports some investigations aimed at reducing the effort needed to produce each twin plant: a twin plant inherent to a higher uncertainty level can be produced by incrementing the twin plant relevant to the lower level.
Abstract. Fault diagnosis is an essential part in the Health Management of autonomous vehicles. Within these vehicles the traction subsystem is a critical component, especially in those exploring planetary surfaces. Recent advances in brushless DC motors has raised the interest in new models and control configurations to integrate them in those vehicles due to their low energy consumption high torque/- mass ratio and low maintenance requirements. In this work we develop a full Bond Graph model of this subsystem, including the brushless motor and the control blocks needed for proper and efficient operation. These models will allow us to perform fault diagnosis with Bond Graph Possible Conflicts as the unifying formalism. We derive the Bond Graph-Possible Conflicts of the system, discussing the viability of the proposal.
Abstract. We conduct a comparative study between two approaches for combining signals from several MPCs designed for different fault scenarios. The first is MPC switching where a switch dictates which of the MPC controllers is currently active. The second is MPC mixing where all MPCs are running concurrently and their outputs are blended in proportion to the current estimate of fault state. We demonstrate results using a gravity drained multi-tank system. Our empirical results show that the mixing approach responds more quickly to faults than the switching approach. Further, we show that the speed and accuracy of fault isolation has a critical impact on fault tolerance.
Abstract. Verifying behavioral or safety properties of hybrid systems, either at design stage such as state reachability and diagnosability, or on-line such as fault detection and isolation is a challenging task. We are concerned here with abstractions oriented towards hybrid systems diagnosability checking. The verification can be done on the abstraction by classical methods developed for discrete event systems extended with time constraints, which provide a counterexample in case of non-diagnosability. The absence of such a counterexample proves the diagnosability of the original hybrid system. In the presence of a counterexample, the first step is to check if it is not a spurious effect of the abstraction and actually exists for the hybrid system, witnessing thus non-diagnosability. Otherwise, we show how to refine the abstraction, guided by the elimination of the counterexample, and continue the process of looking for another counterexample until either a final result is obtained or we reach an inconclusive verdict. We make use of qualitative modeling and reasoning to compute discrete abstractions. Abstractions as timed automata are particularly studied as they allow one to handle time constraints that can be captured at a qualitative level from the hybrid system.
Abstract. Decentralized diagnosis of discrete event systems consists in detecting faults in discrete event systems by using decentralized architectures. In particular, inference-based diagnosis is a decentralized architecture of interest, since it is more general than several other decentralized architectures. In this paper, we first propose a method that realizes a diagnosis objective D by an arborescent architecture (or tree). Each leaf of the tree is a decentralized diagnosis, and each node n is a disjunction or con- junction of the diagnosis decisions of the two children of n. Then, we show that if inference-based diagnosis is applicable to D, then all the leafs of the obtained tree are basic decentralized diagnosers. This implies that every inference-based diagnosis is realizable by a combination of basic decentralized diagnosers.
Abstract. This paper presents an approach to applying model-based diagnosis to the task of interpreting in- formation from a wide variety of sources: text, video, meta-data, audio, etc. Much of the information contained in the sources is contradictory, incomplete, purposely deceptive or biased. People make critical decisions based on such murky information. By automating the construction of alternatives, we can design systems that support intelligence analysts and ordinary citizens in understanding the world. We have developed a preliminary version of our HCDX tool (hypothesis construction through diagnosis). We plan to distribute this tool as open source.
Abstract. Complex technical systems usually show a dynamic behavior that is often conveniently represented with a discrete event model. Such a behavior is the result of dynamic components which interact with each other. Due to the complexity of technical systems faults are not totally avoidable. In order to deal with such faults diagnosing the system at run-time is of great interest. To perform such a diagnosis it is common to use fault models. Such models are in practice often hard to obtain. To address this problem we show a diagnosis approach for discrete event systems which uses the model of the nominal behavior only. In order to perform this diagnosis we adopt the well known idea of consistency based diagnosis.
Abstract. This paper investigates the problem of pattern diagnosis of systems modeled as bounded labeled Petri nets that extends the diagnosis problem on single fault events to more complex behaviors. An effective method to solve the diagnosis problem is proposed. It
relies on a matching relation between the system and the pattern that turns the pattern diagnosis problem into a model-checking problem.
Abstract. In this work we present strategies for (optimal) measurement computation and selection in model- based sequential diagnosis. In particular, assuming a set of leading diagnoses being given, we show how queries (sets of measurements) can be computed and optimized along two dimensions: expected number of queries and cost per query. By means of a suitable decoupling of two optimizations and a clever search space reduction the computations are done without any inference engine calls. For the full search space, we give a method requiring only a polynomial number of inferences and guarantee- ing query properties existing methods do not provide. Evaluation results using real-world problems indicate that the new method computes (virtually) optimal queries instantly independently of the size and complexity of the considered diagnosis problems.
Abstract. In this paper, we propose a method to diagnose faults in a discrete event system that only relies on past observed logs and not on any behavioural model of the system. Given a set of tagged logs produced by the system, the first objective is to extract from them a set of fault signatures. These fault signatures are represented with a set of critical observations that are the support of the diagnosis method. We first propose a method to compute the fault signatures from an initial log journal and follow with detail on how the signatures can then be updated when new logs are available.
Abstract. We address the problem of intermittent fault diagnosis as an instance of discrete signal estimation, in the context of fault management in autonomous systems and vehicles. We propose an estimation approach based on constrained optimization using conditional preference theories. We show that in some cases, our estimator can fail to find an estimation for the system. We provide a way to detect and eliminate these cases at design time.
Abstract. The diagnosis model is certainly a key element for any model-based diagnosis process. Experience shows though that in practice we often have no such model available for one or the other reason, so that in many projects we cannot draw on diagnosis processes when tackling problems. In this paper, we thus show how to improve on available automated processes for deriving a diagnostic model from standard simulation models as usually created during development. We delve in particular into the question how research in the context of combinatorial testing and fault injection can help in this respect, and consider several questions that arise.
Abstract. When diagnosing a faulty system one is often confronted with a large number of possible fault hypotheses. Sequential Diagnosis (SD) techniques aim at the localization or identification of the ac- tual fault with minimal cost or effort. SD can be viewed as an Active Learning (AL) task where the learner, trying to find some target hypothesis, formulates sequential queries to some oracle, thereby e.g. requesting additional system measurements. Several query selection measures (QSMs) for de- termining the best query to ask next have been proposed for AL. To date, few of them have been translated to and employed in SD. In this work, we account for this and analyze various QSMs wrt. to the discrimination power of their selected queries within the diagnostic hypotheses space. As a result, we derive superiority and equivalence relations between these QSMs and introduce improved versions of existing QSMs to overcome identified issues. The obtained picture gives a hint about which QSMs should preferably be used in SD to choose a query from a pool of candidates. Moreover, we deduce properties optimal queries wrt. QSMs must satisfy. Based on these, we devise an efficient heuristic search for optimal queries. As (preliminary) evaluation results indicate, the latter is especially beneficial in applications where query generation is costly, e.g. involving logical reasoning, and hence a pool of query candidates is not (cheaply) available.
Abstract. Model-Based Diagnosis (MBD) is a principled approach to fault localization in any type of system that can be described in a formal structured way. Knowledge Base Debugging (KBD) draws on concepts from MBD to find faults in a monotonic knowledge base. We show that KBD is a generalization of MBD in that any MBD problem can be reduced to a KBD problem and solutions of the former can be directly extracted from solutions of the latter. Moreover, we find that the sequential MBD problem is a special case of the sequential KBD problem in that the latter allows a user to provide more types of measurements. As a consequence of these results, KBD approaches can be applied to all systems amenable to MBD.
Abstract. Functional safety analysis (FSA), that is checking whether a designed artifact will perform safely even under the presence of failing components, has gained significant importance in different areas, including aeronautic and automotive systems. The same applies to failure-modes-and-effects analysis (FMEA) and fault-tree analysis (FTA) as the major contributing processes. FSA is labor- and time-consuming as well as error- prone, and would benefit from computer-based tool-support. Work on qualitative model-based systems has developed principled solutions, particularly to FMEA, but did not achieve the step to industrial practice. Rather than novel technical contributions, this paper discusses reasons for this fact and describes the qSafe* project, which aims at overcoming the obstacles and at making a major step towards producing tools that can support current practice.