Detection Of Anomalous Events Research Articles

Anomaly Detection for Asynchronous Multivariate Time Series of Nuclear Power Plants Using a Temporal-Spatial Transformer.

Industrial process monitoring is a critical application of multivariate time-series (MTS) anomaly detection, especially crucial for safety-critical systems such as nuclear power plants (NPPs). However, some current data-driven process monitoring approaches may not fully capitalize on the temporal-spatial correlations inherent in operational MTS data. Particularly, asynchronous time-lagged correlations may exist among variables in actual NPPs, which further complicates this challenge. In this work, a reconstruction-based MTS anomaly detection approach based on a temporal-spatial transformer is proposed. It employs a two-stage temporal-spatial attention mechanism combined with a multi-scale strategy to learn the dependencies within normal operational data at various scales, thereby facilitating the extraction of temporal-spatial correlations from asynchronous MTS. Experiments on simulated datasets and real NPP datasets demonstrate that the proposed model possesses stronger feature learning capabilities, as evidenced by its improved performance in signal reconstruction and anomaly detection for asynchronous MTS data. Moreover, the proposed TS-Trans model enables earlier detection of anomalous events, which holds significant importance for enhancing operational safety and reducing potential losses in NPPs.

Detection and pre-localization of anomalous consumption events in water distribution networks through automated, pressure-based methodology

Anomalous water-consumption events (AEs) can significantly impact the functioning of water distribution networks, and their prompt identification can improve the service provided by water utilities. This study proposes a new methodology for AE detection and pre-localization in water distribution networks relying exclusively on pressure-data collected in the field, which are exploited to evaluate differential-pressure trends for all possible pressure-sensors couples located in the WDN. In greater detail, AEs are detected and pre-localized by analysing differential-pressure trends over time. The level of deviation of these trends from the standard is considered to provide information about (i) AE alert levels and (ii) the area of the network where the AE is most likely to occur. The application of the methodology to two real case studies featuring different characteristics in terms of residential and industrial users demonstrated method effectiveness in detecting and pre-localizing individual and simultaneous AEs of different magnitude and occurring at different times of the day, providing useful information about the presence of AEs without the need for hydraulic models, and allowing the evaluation of their effects in terms of piezometric head alteration in the different areas of the system.

Clustering Aided Weakly Supervised Training to Detect Anomalous Events in Surveillance Videos.

Formulating learning systems for the detection of real-world anomalous events using only video-level labels is a challenging task mainly due to the presence of noisy labels as well as the rare occurrence of anomalous events in the training data. We propose a weakly supervised anomaly detection system that has multiple contributions including a random batch selection mechanism to reduce interbatch correlation and a normalcy suppression block (NSB) which learns to minimize anomaly scores over normal regions of a video by utilizing the overall information available in a training batch. In addition, a clustering loss block (CLB) is proposed to mitigate the label noise and to improve the representation learning for the anomalous and normal regions. This block encourages the backbone network to produce two distinct feature clusters representing normal and anomalous events. An extensive analysis of the proposed approach is provided using three popular anomaly detection datasets including UCF-Crime, ShanghaiTech, and UCSD Ped2. The experiments demonstrate the superior anomaly detection capability of our approach.

Remote Monitoring and Control System of a Water Distribution Network using LoRaWAN Technology

The problems related to the proper management and control in the distribution of potable water affect environmental sustainability generated by leaks and breaks in the infrastructure, causing leaks and loss of water. According to reports from the National Superintendence of Sanitation Services of Peru, more than 50% of complaints about the water service are related to billing problems and water leaks. It is for this reason that technologies such as the Internet of Things technology contribute to generating solutions for the automatic acquisition of data in residences and houses. That is why this paper aims to use long-range and low-power wireless communication systems to improve the service-oriented to the control of the water distribution network, monitoring of vandalism, and detection of anomalous events, reducing response time and economic losses. The paper's development methodology considers the implementation of a water controller node with flow control sensors and solenoid valves and a gateway with Lora communication. In addition, a solenoid valve control circuit and a remote visualization and control system are implemented. The results indicate that the implemented nodes allow adequate monitoring and control in real-time of the water flow, contributing to the adequate management of its consumption and supporting the detection of anomalous events using a Web application.

Early anomaly detection in time series: a hierarchical approach for predicting critical health episodes

The early detection of anomalous events in time series data is essential in many domains of application. In this paper we deal with critical health events, which represent a significant cause of mortality in intensive care units of hospitals. The timely prediction of these events is crucial for mitigating their consequences and improving healthcare. One of the most common approaches to tackle early anomaly detection problems is through standard classification methods. In this paper we propose a novel method that uses a layered learning architecture to address these tasks. One key contribution of our work is the idea of pre-conditional events, which denote arbitrary but computable relaxed versions of the event of interest. We leverage this idea to break the original problem into two hierarchical layers, which we hypothesize are easier to solve. The results suggest that the proposed approach leads to a better performance relative to state of the art approaches for critical health episode prediction.

Human-related anomalous event detection via spatial-temporal graph convolutional autoencoder with embedded long short-term memory network

Automatic detection of human-related anomalous events in surveillance videos is challenging, owing to unclear definition of anomalies and insufficiency of training data. Generally, the irregular human motion patterns can be regarded as human-related abnormal events. Therefore, we propose a novel method to operate directly on sequences of human skeleton graphs for discovering the normal patterns of human motion. The sequence of skeleton graphs is decomposed into two sub-components: global movement and local posture sequences. The global component is utilized to compute local component. The local component sequences are then input to our network for capturing normal spatial-temporal motion patterns of human skeleton. Our network is established on a Spatial-temporal Graph Convolutional Autoencoder (ST-GCAE) and embedded with Long Short-Term Memory (LSTM) network in hidden layers for exploring the temporal cues, which is thus called Spatial-temporal Graph Convolutional Autoencoder with Embedded Long Short-Term Memory Network (STGCAE-LSTM). Different from traditional autoencoder, STGCAE-LSTM owns a single-encoder-dual-decoder architecture, which is capable of reconstructing the input and predicting the unseen future simultaneously. Then, samples that deviate from normal patterns are detected as anomalies with fusion of reconstruction and prediction errors. Experimental results on four challenging datasets demonstrate advantages of our method over other state-of-the-art algorithms.

Detection and Prediction of Adverse and Anomalous Events in Medical Robots

Adverse and anomalous (A&A) events are a serious concern in medical robots. We describe a system that can rapidly detect such events and predict their occurrence. As part of this system, we describe simulation, data collection and user interface tools we build for a robot for small animal biopsies. The data we collect consists of both the hardware state of the robot and variables in the software controller. We use this data to train dynamic Bayesian network models of the joint hardware-software state-space dynamics of the robot. Our empirical evaluation shows that (i) our models can accurately model normal behavior of the robot, (ii) they can rapidly detect anomalous behavior once it starts, (iii) they can accurately predict a future A&A event within a time window of it starting and (iv) the use of additional software variables beyond the hardware state of the robot is important in being able to detect and predict certain kinds of events.

Measure invariance of ergodic symbolic systems for low-delay detection of anomalous events

This paper investigates the spectral properties of ergodic measure of symbolic systems for applications to signal processing, pattern recognition, and anomaly detection in uncertain dynamical systems. The underlying algorithm is built upon the concept of ergodic sequences of measure-preserving transformations (MPT) on probability spaces, where non-stationary probabilistic finite state automata (PFSA) are constructed using short-length windows of time-series measurements. The resulting PFSA are non-homogeneous Markov, in which spectral properties of the MPT depend on several parameters that include the window length. The paper also develops an MPT-based metric to quantify the divergence of the evolving PFSA from that of the nominal PFSA; this information is then used for pattern classification and anomaly detection with low-delay tolerance. The MPT-based methodology has been validated with experimental data generated from a laboratory apparatus that deals with detection of thermoacoustic instabilities in combustion processes, which are known to have chaotic characteristics on a time scale of milliseconds. In this application, the concepts of MPT and ergodicity have been used to develop a novel symbolic time series analysis (STSA)-based detection method, whose performance is validated by comparison with two well-known techniques, namely, hidden Markov model (HMM) and cumulative sum (CUSUM), on the same experimental data. The results consistently show superior performance of the proposed MPT-based STSA.

ВЫБОР МЕТРИЧЕСКИХ АТРИБУТОВ РЕДКИХ АНОМАЛЬНЫХ СОБЫТИЙ КОМПЬЮТЕРНОЙ СИСТЕМЫ МЕТОДАМИ ИНТЕЛЛЕКТУАЛЬНОГО АНАЛИЗА ДАННЫХ

The process of marking multi-attribute experimental data for subsequent use by means of data mining in problems of detection and classification of rare anomalous events of computer systems (CS) is considered. The labeling process is carried out using three methods: manual preprocessing, statistical analysis and cluster analysis. Among the attributes of the metric type, the authors identified two macrogroups: “integral attributes” and “impulse attributes”. It is shown that the combination of statistical and cluster analysis methods increases the accuracy of detecting anomalous events in the CS, and also allows the selection of attributes according to their information significance. The expediency of manual preprocessing of data before clustering is shown by the example of dividing attributes into macrogroups, analyzing the density distribution using violin plot and removing the trend component using the method difference stationary series. With the help of construction of violin diagrams (Violin plot) for the attribute of the “integral” macrogroup, the distribution of states of the CS is shown. It is shown that the removal of the trend component by the DS-series method, normalization and reduction to absolute values allows more accurate marking of anomalous outliers, but this is not always acceptable. The interpretation of the clustering results performed for each normalized attribute shows that the normal values for all attributes are concentrated around zero values. The result of labeling experimental data is attribute-labeled data, where each attribute at the current time is assigned one of two states: abnormal or normal.

Detecting Telecommunication Fraud with Visual Analytics: A Review

The detection of anomalous events in large multivariate data is sought in many domains. Analysis of data is an important fraud detection procedure in detecting suspicious events and prevent attempts to defraud. While now the data is becoming more complicated and difficult as data scales and complexities increase than ever before, the rich insights within the data may be difficult to identify by traditional means and often remain hidden. People require powerful tools to extract valid conclusions from the data while maintaining trustworthy and interpretable results. Hence, various fraud detection approaches have started to exploit Visual Analytics (VA) techniques to reveal the hidden knowledge in such fraudulent activities. Interactive data visualization tools have substantial potential for making the detection of fraudulent transactions more efficient and effective by allowing the investigator to change the representation of data from text and numeric into graphics and filter out subsets of transactions for further fraud investigation. However, little research to date has directly examined the efficacy of data visualization techniques for fraud detection especially telecommunication fraud. In this paper, we present an overview of several fraud detection solutions that use data visualization techniques to detect fraudulent transactions in the telecommunication domain. The paper concludes by discussing how academic research might proceed in investigating the efficacy of interactive data visualization tools for fraud detection.

Resettable skin interfaced microfluidic sweat collection devices with chemesthetic hydration feedback

Recently introduced classes of thin, soft, skin-mounted microfluidic systems offer powerful capabilities for continuous, real-time monitoring of total sweat loss, sweat rate and sweat biomarkers. Although these technologies operate without the cost, complexity, size, and weight associated with active components or power sources, rehydration events can render previous measurements irrelevant and detection of anomalous physiological events, such as high sweat loss, requires user engagement to observe colorimetric responses. Here we address these limitations through monolithic systems of pinch valves and suction pumps for purging of sweat as a reset mechanism to coincide with hydration events, microstructural optics for reversible readout of sweat loss, and effervescent pumps and chemesthetic agents for automated delivery of sensory warnings of excessive sweat loss. Human subject trials demonstrate the ability of these systems to alert users to the potential for dehydration via skin sensations initiated by sweat-triggered ejection of menthol and capsaicin.

Anomalous Behaviors Detection in Moving Crowds Based on a Weighted Convolutional Autoencoder-Long Short-Term Memory Network

We propose an anomaly detection approach by learning a generative model of moving pedestrians to guarantee public safety. To resolve the existing challenges of anomaly detection in complicated definitions, complex backgrounds, and local occurrence, a weighted convolutional autoencoder-long short-term memory network is proposed to reconstruct raw data and their corresponding optical flow and then perform anomaly detection based on reconstruction errors. Unlike equally treating raw data and optical flow, a novel two-stream framework is proposed to take the reconstructed optical flow as supplementary cues that encode pedestrian motions. A weighted Euclidean loss is proposed to enable the network to concentrate on moving foregrounds, thus restraining background influence. Global-local analysis is used to jointly detect and localize local anomaly in reconstructed raw data. Final detection of anomalous events is achieved by jointly considering the results of the global-local analysis and reconstructed optical flow. Qualitative evaluations verify the effectiveness of our two-stream framework, the weighted Euclidean loss, and the global-local analysis. Moreover, comparisons with state-of-the-art approaches indicate the superiority of our approach in anomaly detection.

Trajectory-Based Surveillance Analysis: A Survey

Due to the advancement of camera hardware and machine learning techniques, video object tracking for surveillance has received noticeable attention from the computer vision research community. Object tracking and trajectory modeling have important applications in surveillance video analysis. For example, trajectory clustering, summarization or synopsis generation, and detection of anomalous or abnormal events in videos are mainly being exploited by the research community. However, barring one research work (which is almost a decade old), there is no recent review that emphasizes the use of video object trajectories, particularly in the perspective of visual surveillance. This paper presents a survey of trajectory-based surveillance applications with a focus on clustering, anomaly detection, summarization, and synopsis generation. The methods reviewed in this paper broadly summarize the abovementioned applications. The main purpose of this survey is to summarize the state-of-the-art video object trajectory analysis techniques used in the indoor and outdoor surveillance.

Similarity-aware data aggregation using fuzzy c-means approach for wireless sensor networks

For resource-constrained IoT systems, data collection is one of the fundamental operations to reduce the energy dissipation of sensor nodes and improve the network lifetime. However, an anomaly or deviation will exert a great influence on the quality of data collected, especially for a data aggregation scheme. By taking into account data-aware clustering and detection of anomalous events, a similarity-aware data aggregation using a fuzzy c-means approach for wireless sensor networks is proposed. Firstly, by using a fuzzy c-means approach, the clustering process can be performed to organize sensors into clusters based on data similarity. Next, an effective support degree function is defined for further outlier diagnosis. Afterwards, the appropriate weight of valid data can be obtained by taking advantage of the probability distribution characteristics of normal samples within a certain period. Finally, the aggregation result in the cluster can be estimated. Practical database-based simulations have confirmed that the proposed data aggregation method can achieve better performance than traditional methods in terms of data outlier detection accuracy and relative recovery error.

Visual analytics for event detection: Focusing on fraud

The detection of anomalous events in huge amounts of data is sought in many domains. For instance, in the context of financial data, the detection of suspicious events is a prerequisite to identify and prevent attempts to defraud. Hence, various financial fraud detection approaches have started to exploit Visual Analytics techniques. However, there is no study available giving a systematic outline of the different approaches in this field to understand common strategies but also differences. Thus, we present a survey of existing approaches of visual fraud detection in order to classify different tasks and solutions, to identify and to propose further research opportunities. In this work, fraud detection solutions are explored through five main domains: banks, the stock market, telecommunication companies, insurance companies, and internal frauds. The selected domains explored in this survey were chosen for sharing similar time-oriented and multivariate data characteristics. In this survey, we (1) analyze the current state of the art in this field; (2) define a categorization scheme covering different application domains, visualization methods, interaction techniques, and analytical methods which are used in the context of fraud detection; (3) describe and discuss each approach according to the proposed scheme; and (4) identify challenges and future research topics.

Performance evaluation of social network anomaly detection using a moving window–based scan method

AbstractTimely detection of anomalous events in networks, particularly social networks, is a problem of increasing interest and relevance. A variety of methods have been proposed for monitoring such networks, including the window‐based scan method proposed by a previous study. However, research assessing the performance of this and other methods has been sparse. In this article, we use simulated social network structures to study the performance of the Priebe et al method. The detection power is high only when more than half of the social network experiences anomalous behavior or if the anomalous behavior is extreme. Both can be represented by high signal‐to‐noise ratios in the network. More precisely, Priebe's scan method performs well when the signal‐to‐noise ratio is above 20. Simulation studies are used to show that an improved detection rate and shortened monitoring delays can be achieved by lagging the moving window used for standardization, lowering the signaling threshold, and using shorter moving windows at the initial stage of monitoring. We suggest a community detection method to be used after an anomalous event has been identified to help determine the subnetwork associated with this anomalous behavior.

Detection of Anomalous Noise Events for Real-Time Road-Traffic Noise Mapping: The Dynamap’s project case study

Abstract Environmental noise is increasing year after year, especially in urban and suburban areas. Besides annoyance, environmental noise also causes harmful health effects on people. The Environmental Noise Directive 2002/49/EC (END) is the main instrument of the European Union to identify and combat noise pollution, followed by the CNOSSOS-EU methodological framework. In compliance with the END legislation, the European Member States are required to publish noise maps and action plans every five years. The emergence of Wireless Acoustic Sensor Networks (WASNs) have changed the paradigm to address the END regulatory requirements, allowing the dynamic ubiquitous measurement of environmental noise pollution. Following the END, the LIFE DYNAMAP project aims to develop a WASN-based low-cost noise mapping system to monitor the acoustic impact of road infrastructures in real time. Those acoustic events unrelated to regular traffic noise should be removed from the equivalent noise level calculations to avoid biasing the noise map generation. This work describes the different approaches developed within the DYNAMAP project to implement an Anomalous Noise Event Detector on the low-cost sensors of the network, considering both synthetic and real-life acoustic data.Moreover, the paper reflects on several open challenges, discussing how to tackle them for the future deployment of WASN-based noise monitoring systems in real-life operating conditions.

Deep Learning with a Spatiotemporal Descriptor of Appearance and Motion Estimation for Video Anomaly Detection

The automatic detection and recognition of anomalous events in crowded and complex scenes on video are the research objectives of this paper. The main challenge in this system is to create models for detecting such events due to their changeability and the territory of the context of the scenes. Due to these challenges, this paper proposed a novel HOME FAST (Histogram of Orientation, Magnitude, and Entropy with Fast Accelerated Segment Test) spatiotemporal feature extraction approach based on optical flow information to capture anomalies. This descriptor performs the video analysis within the smart surveillance domain and detects anomalies. In deep learning, the training step learns all the normal patterns from the high-level and low-level information. The events are described in testing and, if they differ from the normal pattern, are considered as anomalous. The overall proposed system robustly identifies both local and global abnormal events from complex scenes and solves the problem of detection under various transformations with respect to the state-of-the-art approaches. The performance assessment of the simulation outcome validated that the projected model could handle different anomalous events in a crowded scene and automatically recognize anomalous events with success.

Detection of Anomalous Noise Events on Low-Capacity Acoustic Nodes for Dynamic Road Traffic Noise Mapping within an Hybrid WASN.

One of the main aspects affecting the quality of life of people living in urban and suburban areas is the continuous exposure to high road traffic noise (RTN) levels. Nowadays, thanks to Wireless Acoustic Sensor Networks (WASN) noise in Smart Cities has started to be automatically mapped. To obtain a reliable picture of the RTN, those anomalous noise events (ANE) unrelated to road traffic (sirens, horns, people, etc.) should be removed from the noise map computation by means of an Anomalous Noise Event Detector (ANED). In Hybrid WASNs, with master-slave architecture, ANED should be implemented in both high-capacity (Hi-Cap) and low-capacity (Lo-Cap) sensors, following the same principle to obtain consistent results. This work presents an ANED version to run in real-time on Controller-based Lo-Cap sensors of a hybrid WASN, discriminating RTN from ANE through their Mel-based spectral energy differences. The experiments, considering 9 h and 8 min of real-life acoustic data from both urban and suburban environments, show the feasibility of the proposal both in terms of computational load and in classification accuracy. Specifically, the ANED Lo-Cap requires around of the computational load of the ANED Hi-Cap, while classification accuracies are slightly lower (around 10%). However, preliminary analyses show that these results could be improved in around 4% in the future by means of considering optimal frequency selection.

Designing Size Consistent Statistics for Accurate Anomaly Detection in Dynamic Networks

An important task in network analysis is the detection of anomalous events in a network time series. These events could merely be times of interest in the network timeline or they could be examples of malicious activity or network malfunction. Hypothesis testing using network statistics to summarize the behavior of the network provides a robust framework for the anomaly detection decision process. Unfortunately, choosing network statistics that are dependent on confounding factors like the total number of nodes or edges can lead to incorrect conclusions (e.g., false positives and false negatives). In this article, we describe the challenges that face anomaly detection in dynamic network streams regarding confounding factors. We also provide two solutions to avoiding error due to confounding factors: the first is a randomization testing method that controls for confounding factors, and the second is a set of size-consistent network statistics that avoid confounding due to the most common factors, edge count and node count.