Model-based fault detection and isolation for a gas–liquid separation unit

A complete fault detection and isolation system is designed for a gas–liquid separation unit. It involves the determination and identification of grey box models, the design of a model-based residual generator, and finally the evaluation of the residuals via a set of statistical tests. The latter are cumulative sum (CUSUM) tests which are combined in such a way that both fault detection and fault isolation can be achieved. The performance of the resulting diagnosis system, such as missed alarm rate, wrong isolation rate and mean detection delay, are studied via simulations.     

A fault detection scheme for linear discrete-time systems with an integrated online performance evaluation

This paper is concerned with the design of the fault detection systems, into which a residual generation, evaluation and threshold are integrated, for linear discrete time-varying processes over a finite horizon. In the proposed design scheme, the residual generation is realised in the context of H_∞ fault estimation with a prescribed attenuation level. This attenuation level is minimised by using the Krein-space linear estimation theory and, subsequently, an H_∞ fault estimator with the minimum attenuation level is designed in terms of the solution to a set of Riccati-like recursions. For the residual evaluation and decision making purpose, the false alarm rate and fault detection rate indicators are introduced in the norm-based framework, which is integrated into the decision making procedure. For the online computations of the false alarm rate and fault detection rate indicators, further estimates delivered by the H_∞ fault estimator are applied without additional (online) computations. By means of checking the change in the false alarm rate and fault detection rate indicators, a decision is then made. In this way, the fault detection performance can be significantly improved. Finally, one application example is exploited to demonstrate the application of the proposed integrated fault detection and performance evaluation schemes.     

Local and Nonlocal Discrete Regularization on Weighted Graphs for Image and Mesh Processing

We propose a discrete regularization framework on weighted graphs of arbitrary topology, which unifies local and nonlocal processing of images, meshes, and more generally discrete data. The approach considers the problem as a variational one, which consists in minimizing a weighted sum of two energy terms: a regularization one that uses the discrete p-Dirichlet form, and an approximation one. The proposed model is parametrized by the degree p of regularity, by the graph structure and by the weight function. The minimization solution leads to a family of simple linear and nonlinear processing methods. In particular, this family includes the exact expression or the discrete version of several neighborhood filters, such as the bilateral and the nonlocal means filter. In the context of images, local and nonlocal regularizations, based on the total variation models, are the continuous analog of the proposed model. Indirectly and naturally, it provides a discrete extension of these regularization methods for any discrete data or functions.     

Fault detection in autonomous robots based on fault injection and learning

In this paper, we study a new approach to fault detection for autonomous robots. Our hypothesis is that hardware faults change the flow of sensory data and the actions performed by the control program. By detecting these changes, the presence of faults can be inferred. In order to test our hypothesis, we collect data from three different tasks performed by real robots. During a number of training runs, we record sensory data from the robots while they are operating normally and after a fault has been injected. We use back-propagation neural networks to synthesize fault detection components based on the data collected in the training runs. We evaluate the performance of the trained fault detectors in terms of number of false positives and time it takes to detect a fault. The results show that good fault detectors can be obtained. We extend the set of possible faults and go on to show that a single fault detector can be trained to detect several faults in both a robot’s sensors and actuators. We show that fault detectors can be synthesized that are robust to variations in the task, and we show how a fault detector can be trained to allow one robot to detect faults that occur in another robot.

A novel unsupervised classification approach for network anomaly detection by k-Means clustering and ID3 decision tree learning methods

This paper presents a novel host-based combinatorial method based on k-Means clustering and ID3 decision tree learning algorithms for unsupervised classification of anomalous and normal activities in computer network ARP traffic. The k-Means clustering method is first applied to the normal training instances to partition it into k clusters using Euclidean distance similarity. An ID3 decision tree is constructed on each cluster. Anomaly scores from the k-Means clustering algorithm and decisions of the ID3 decision trees are extracted. A special algorithm is used to combine results of the two algorithms and obtain final anomaly score values. The threshold rule is applied for making the decision on the test instance normality. Experiments are performed on captured network ARP traffic. Some anomaly criteria has been defined and applied to the captured ARP traffic to generate normal training instances. Performance of the proposed approach is evaluated using five defined measures and empirically compared with the performance of individual k-Means clustering and ID3 decision tree classification algorithms and the other proposed approaches based on Markovian chains and stochastic learning automata. Experimental results show that the proposed approach has specificity and positive predictive value of as high as 96 and 98%, respectively.     

Spatial–spectral locality constrained elastic net hypergraph for hyperspectral image clustering

Hypergraph is an effective method used to represent the contextual correlation within hyperspectral imagery for clustering. Nevertheless, how to discover the closely correlated samples to form hyperedges is the key issue for constructing an informative hypergraph. In this article, a new spatial–spectral locality constrained elastic net hypergraph learning model is proposed for hyperspectral image clustering (i.e. unsupervised classification). In order to utilize the spatial–spectral correlation among the pixels in hyperspectral images, first, we construct a locality-constrained dictionary by selecting K relevant pixels within a spatial neighbourhood, which activates the most correlated atoms and suppresses the uncorrelated ones. Second, each pixel is represented as a linear combination of the atoms in the dictionary under the elastic net regularization. Third, based on the obtained representations, the pixels and their most related pixels are linked as hyperedges, which can effectively capture high–order relationships among the pixels. Finally, a hypergraph Laplacian matrix is built for unsupervised learning. Experiments have been conducted on two widely used hyperspectral images, and the results show that the proposed method can achieve a superior clustering performance when compared to state-of-the-art methods.     

Integrating multi-objective genetic algorithm based clustering and data partitioning for skyline computation

Skyline computation in databases has been a hot topic in the literature because of its interesting applications. The basic idea is to find non-dominated values within a database. The task is mainly a multi-objective optimization process as described in this paper. This motivated for our approach that employs a multi-objective genetic algorithm based clustering approach to find the pareto-optimal front which allows us to locate skylines within a given data. To tackle large data, we simply split the data into manageable subsets and concentrate our analysis on the subsets instead of the whole data at once. The proposed approach produced interesting results as demonstrated by the outcome from the conducted experiments.     

Finite-frequency fault detection for two-dimensional Fornasini–Marchesini dynamical systems

This paper is concerned with the fault detection problem for two-dimensional (2-D) discrete-time systems described by the Fornasini–Marchesini local state-space model. The goal of the paper is to design a fault detection filter to detect the occurrence of faults in finite-frequency domain. To this end, a finite-frequency H_? index is used to describe fault sensitivity performance, and a finite-frequency H_∞ index is used to describe disturbance attenuation performance. In light of the generalised Kalman–Yakubovich–Popov lemma for 2-D systems and matrix inequality techniques, convex conditions are derived for this fault detection problem. Based on these conditions, a numerical algorithm is put forward to construct a desired fault detection filter. Finally, a numerical example and an industrial example are given to illustrate the effectiveness of the proposed algorithm.     

Optimal and online preemptive scheduling on uniformly related machines

We consider the problem of preemptive scheduling on uniformly related machines. We present a semi-online algorithm which, if the optimal makespan is given in advance, produces an optimal schedule. Using the standard doubling technique, this yields a 4-competitive deterministic and an e≈2.71-competitive randomized online algorithm. In addition, it matches the performance of the previously known algorithms for the offline case, with a considerably simpler proof. Finally, we study the performance of greedy heuristics for the same problem.     

Robust fault isolation for a class of non-linear input?output systems

The design and analysis of fault diagnosis methodologies for non-linear systems has received significant attention recently. This paper presents a robust fault isolation scheme for a class of non-linear systems with unstructured modelling uncertainty and partial state measurement. The proposed fault diagnosis architecture consists of a fault detection and approximation estimator and a bank of isolation estimators. Each isolation estimator corresponds to a particular type of fault in the fault class. A fault isolation decision scheme is presented with guaranteed performance. If at least one component of the output estimation error of a particular fault isolation estimator exceeds the corresponding adaptive threshold at some finite time, then the occurrence of that type of fault can be excluded. Fault isolation is achieved if this is valid for all but one isolation estimator. Based on the class of non-linear systems under consideration, fault isolability conditions are rigorously investigated, characterizing the class of non-linear faults that are isolable by the proposed scheme. Moreover, the non-conservativeness of the fault isolability conditions is illustrated by deriving a subclass of nonlinear systems and faults for which this condition is also necessary for fault isolability. A simulation example of a simple robotic system is used to show the effectiveness of the robust fault isolation methodology.     

Design of a bilinear fault detection observer for singular bilinear systems

A bilinear fault detection observer is proposed for a class of continuous time singular bilinear systems subject to unknown input disturbance and fault. By singular value decomposition on the original system, a bilinear fault detection observer is proposed for the decomposed system via an algebraic Riccati equation, and the domain of attraction of the state estimation error is estimated. A design procedure is presented to determine the fault detection threshold. A model of flexible joint robot is used to demonstrate the effectiveness of the proposed method.     

Zonotopic fault detection observer design for Takagi–Sugeno fuzzy systems

This paper considers zonotopic fault detection observer design in the finite-frequency domain for discrete-time Takagi–Sugeno fuzzy systems with unknown but bounded disturbances and measurement noise. We present a novel fault detection observer structure, which is more general than the commonly used Luenberger form. To make the generated residual sensitive to faults and robust against disturbances, we develop a finite-frequency fault detection observer based on generalised Kalman–Yakubovich–Popov lemma and P-radius criterion. The design conditions are expressed in terms of linear matrix inequalities. The major merit of the proposed method is that residual evaluation can be easily implemented via zonotopic approach. Numerical examples are conducted to demonstrate the proposed method.     

Robust fault detection for discrete-time Markovian jump systems with mode-dependent time-delays

This paper investigates a fault detection problem for a class of discrete-time Markovian jump systems with norm-bounded uncertainties and mode-dependent time-delays. Attention is focused on constructing the residual generator based on the filter of which its parameters matrices are dependent on the system mode, that is, the fault detection filter is a Markovian jump system as well. The design of fault detection filter is reduced to H-infinity filtering problem by using H-infinity control theory, which can guarantee the difference between the residual and the fault (or, more generally weighted fault) as small as possible in the context of enhancing the robustness of residual to modeling errors, control inputs and unknown inputs. Sufficient condition for the existence of the above filters is established by means of linear matrix inequalities, which can be readily solved by using standard numerical software. A numerical example is given to illustrate the feasibility of the proposed method.     

Homogeneous String Segmentation using Trees and Weighted Independent Sets

We divide a string into k segments, each with only one sort of symbols, so as to minimize the total number of exceptions. Motivations come from machine learning and data mining. For binary strings we develop a linear-time algorithm for any k. Key to efficiency is a special-purpose data structure, called W-tree, which reflects relations between repetition lengths of symbols. For non-binary strings we give a nontrivial dynamic programming algorithm. Our problem is equivalent to finding weighted independent sets with certain size constraints, either in paths (binary case) or special interval graphs (general case). We also show that this problem is FPT in bounded-degree graphs.     

H∞ fault detection with randomly occurring nonlinearities and channel fadings

In this paper, the H_∞ fault detection problem is investigated for a class of discrete-time stochastic systems with both channel fadings and randomly occurring nonlinearities. Due to Doppler effect and multi-path delays, channel fadings are inevitable and also cause unpredictable dynamic behaviour. The Lth Rice fadings model, which is accounted for both channel fadings and time delays, can be employed to describe this phenomenon. Meanwhile, by using a Bernoulli distributed white sequence, a kind of non-linear disturbance appearing in a random way is also considered in the H_∞ fault detection issue. The purpose of the addressed problem is to design a fault detection filter such that, in the presence of channel fadings, the overall fault detection dynamics is stochastically stable and, at the same time, the error between the residual (generated by the fault detection filter) and the fault signal is made as small as possible. By utilising the Lyapunov stability theory associated with the intensive stochastic analysis techniques, sufficient conditions are established under which the addressed H_∞ fault detection problem is recast as solving a convex optimisation problem via the semi-definite programme method. Finally, a simulation example is exploited to show the effectiveness of the method proposed in this paper.     

Particle swarm optimizer for variable weighting in clustering high-dimensional data

In this paper, we present a particle swarm optimizer (PSO) to solve the variable weighting problem in projected clustering of high-dimensional data. Many subspace clustering algorithms fail to yield good cluster quality because they do not employ an efficient search strategy. In this paper, we are interested in soft projected clustering. We design a suitable k-means objective weighting function, in which a change of variable weights is exponentially reflected. We also transform the original constrained variable weighting problem into a problem with bound constraints, using a normalized representation of variable weights, and we utilize a particle swarm optimizer to minimize the objective function in order to search for global optima to the variable weighting problem in clustering. Our experimental results on both synthetic and real data show that the proposed algorithm greatly improves cluster quality. In addition, the results of the new algorithm are much less dependent on the initial cluster centroids. In an application to text clustering, we show that the algorithm can be easily adapted to other similarity measures, such as the extended Jaccard coefficient for text data, and can be very effective.     

Classification-based wheel slip detection and detector fusion for mobile robots on outdoor terrain

This paper introduces a signal-recognition based approach for detecting autonomous mobile robot immobilization on outdoor terrain. The technique utilizes a support vector machine classifier to form class boundaries in a feature space composed of statistics related to inertial and (optional) wheel speed measurements. The proposed algorithm is validated using experimental data collected with an autonomous robot operating in an outdoor environment. Additionally, two detector fusion techniques are proposed to combine the outputs of multiple immobilization detectors. One technique is proposed to minimize false immobilization detections. A second technique is proposed to increase overall detection accuracy while maintaining rapid detector response. The two fusion techniques are demonstrated experimentally using the detection algorithm proposed in this work and a dynamic model-based algorithm. It is shown that the proposed techniques can be used to rapidly and robustly detect mobile robot immobilization in outdoor environments, even in the absence of absolute position information.

High performance parallel k-means clustering for disk-resident datasets on multi-core CPUs

Nowadays, clustering of massive datasets is a crucial part of many data-analytic tasks. Most of the available clustering algorithms have two shortcomings when used on big data: (1) a large group of clustering algorithms, e.g. \(k\) -means, has to keep the data in memory and iterate over the data many times which is very costly for big datasets, (2) clustering algorithms that run on limited memory sizes, especially the family of stream-clustering algorithms, do not have a parallel implementation to utilize modern multi-core processors and also they lack decent quality of results. In this paper, we propose an algorithm that combines parallel clustering with single-pass, stream-clustering algorithms. The aim is to make a clustering algorithm that utilizes maximum capabilities of a regular multi-core PC to cluster the dataset as fast as possible while resulting in acceptable quality of clusters. Our idea is to split the data into chunks and cluster each chunk in a separate thread. Then, the clusters extracted from chunks are aggregated at the final stage using re-clustering. Parameters of the algorithm can be adjusted according to hardware limitations. Experimental results on a 12-core computer show that the proposed method is much faster than its batch-processing equivalents (e.g. \(k\) -means++) and stream-based algorithms. Also, the quality of solution is often equal to \(k\) -means++, while it significantly dominates stream-clustering algorithms. Our solution also scales well with extra available cores and hence provides an effective and fast solution to clustering large datasets on multi-core and multi-processor systems.