The paper concerns the detection of fall events based on human silhouette shape variations. The detection of fall events is addressed from the statistical point of view as an anomaly detection problem. Specifically, the paper investigates the multivariate exponentially weighted moving average (MEWMA) control chart to detect fall events. Towards this end, a set of ratios for five partial occupancy areas of the human body for each frame are collected and used as the input data to MEWMA chart. The MEWMA fall detection scheme has been successfully applied to two publicly available fall detection databases, the UR fall detection dataset (URFD) and the fall detection dataset (FDD). The monitoring strategy developed was able to provide early alert mechanisms in the event of fall situations.

Chemical and petrochemical processes require continuous monitoring to detect abnormal events and to sustain normal operations. Furthermore, process monitoring enhances productivity, efficiency, and safety in process industries. Here, we propose an innovative statistical approach that exploits the advantages of multiscale partial least squares (MSPLS) models and generalized likelihood ratio (GLR) tests for fault detection in processes. Specifically, we combine an MSPLS algorithm with wavelet analysis to create our modeling framework. Then, we use GLR hypothesis testing based on the uncorrelated residuals obtained from the MSPLS model to improve fault detection. We use simulated distillation column data to evaluate the MSPLS-based GLR chart. Results show that our MSPLS-based GLR method is more powerful than the PLS-based Q and GLR method and MSPLS-based Q method, especially in early detection of small faults with abrupt or incipient behavior.

Accurate and effective anomaly detection and diagnosis of modern engineering systems by monitoring processes ensure reliability and safety of a product while maintaining desired quality. In this paper, an innovative method based on Kullback-Leibler divergence for detecting incipient anomalies in highly correlated multivariate data is presented. We use a partial least square (PLS) method as a modeling framework and a symmetrized Kullback-Leibler distance (KLD) as an anomaly indicator, where it is used to quantify the dissimilarity between current PLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, this paper reports the development of two monitoring charts based on the KLD. The first approach is a KLD-Shewhart chart, where the Shewhart monitoring chart with a three sigma rule is used to monitor the KLD of the response variables residuals from the PLS model. The second approach integrates the KLD statistic into the exponentially weighted moving average monitoring chart. The performance of the PLS-based KLD anomaly-detection methods is illustrated and compared to that of conventional PLS-based anomaly detection methods. Using synthetic data and simulated distillation column data, we demonstrate the greater sensitivity and effectiveness of the developed method over the conventional PLS-based methods, especially when data are highly correlated and small anomalies are of interest. Results indicate that the proposed chart is a very promising KLD-based method because KLD-based charts are, in practice, designed to detect small shifts in process parameters. © 2016 Elsevier Ltd

Fault detection has a vital role in the process industry to enhance productivity, efficiency, and safety, and to avoid expensive maintenance. This paper proposes an innovative multivariate fault detection method that can be used for monitoring nonlinear processes. The proposed method merges advantages of nonlinear projection to latent structures (NLPLS) modeling and those of Hellinger distance (HD) metric to identify abnormal changes in highly correlated multivariate data. Specifically, the HD is used to quantify the dissimilarity between current NLPLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, to enhance further the robustness of these methods to measurement noise, and reduce the false alarms due to modeling errors, wavelet-based multiscale filtering of residuals is used before the application of the HD-based monitoring scheme. The performances of the developed NLPLS-HD fault detection technique is illustrated using simulated plug flow reactor data. The results show that the proposed method provides favorable performance for detection of faults compared to the conventional NLPLS method.

This paper proposes an efficient data-based anomaly detection method that can be used for monitoring nonlinear processes. The proposed method merges advantages of nonlinear projection to latent structures (NLPLS) modeling and those of Hellinger distance (HD) metric to identify abnormal changes in highly correlated multivariate data. Specifically, the HD is used to quantify the dissimilarity between current NLPLS-based residual and reference probability distributions. The performances of the developed anomaly detection using NLPLS-based HD technique is illustrated using simulated plug flow reactor data.

The global historical radiosonde archives date back to the 1920s and contain the only directly observed measurements of temperature, wind, and moisture in the upper atmosphere, but they contain many random errors. Most of the focus on cleaning these large datasets has been on temperatures, but winds are important inputs to climate models and in studies of wind climatology. The bivariate distribution of the wind vector does not have elliptical contours but is skewed and heavy-tailed, so we develop two methods for outlier detection based on the bivariate skew-t (BST) distribution, using either distance-based or contour-based approaches to flag observations as potential outliers. We develop a framework to robustly estimate the parameters of the BST and then show how the tuning parameter to get these estimates is chosen. In simulation, we compare our methods with one based on a bivariate normal distribution and a nonparametric approach based on the bagplot. We then apply all four methods to the winds observed for over 35,000 radiosonde launches at a single station and demonstrate differences in the number of observations flagged across eight pressure levels and through time. In this pilot study, the method based on the BST contours performs very well.

Reliable detection of traffic congestion provides pertinent information for improving safety and comfort by alerting the driver to crowded roads or providing useful information for rapid decision-making. This paper addresses the problem of road traffic congestion estimation and detection from a statistical approach. First, a piecewise switched linear traffic model (PWSL)-based observer is introduced. The proposed hybrid observer (HO) estimates the unmeasured traffic density, thus reducing the cost of implementing and maintenance sensors and measurements devices. Here, the observer gains of each mode are obtained by solving a set of linear matrix inequalities. Second, a novel method for efficiently monitoring traffic congestion is proposed by combining the proposed HO with a generalized likelihood ratio (GLR) test. Also, an exponentially-weighted moving average (EWMA) filter is applied to the residual data to reduce high-frequency noise. Thus, as the EWMA filter, aggregates all of the information from past and actual samples in the decision rule, it extends the congestion detection abilities of the GLR test to the detection of incipient changes. This study shows that a better performance is achieved when GLR is applied to filtered data than to unfiltered data. The effectiveness of the proposed approach is verified on traffic data from the four-lane State Route 60 (SR-60) and the three lanes Interstate 210 (I-210) in California freeways. Results show the efficacy of the proposed HO-based EWMA-GLR method to monitor traffic congestions. Also, the proposed approach is compared to that of the conventional Shewhart and EWMA approaches and found better performance.

Denial of service (DOS) and distributed DOS (DDOS) continue to be a significant concern in internet and networking systems. This paper targets to develop an anomaly detection mechanism based on the generalized likelihood ratio (GLR) scheme to detect TCP and ICMPv6 based DOS/DDOS attacks. The anomaly detection problem is addressed as a hypothesis testing problem. The proposed approach uses GLR test to monitor internet traffic for better detecting potential cyber- attacks. The decision threshold of GLR approach has been computed non parametrically based on kernel density estimation. To evaluate the performance of this approach, two network traffic datasets have been used namely the DARPA99 and ICMPv6 datasets. Results highlight the efficiency of the proposed method.

This paper proposes an intrusion detection scheme for Denial Of Service (DOS) and Distributed DOS (DDOS) attacks detection. We used Hellinger distance (HD), which is an effective measure to quantify the similarity between two distributions, to detect the presence of potential malicious attackers. Specifically, we applied HD-based anomaly detection mechanism to detect SYN and ICMPv6-based DOS/DDOS attacks. Here, Shewhart chart is applied to HD to set up a detection threshold. The proposed mechanism is evaluated using DARPA99 and ICMPv6 traffic datasets. Results indicate that our mechanism accomplished reliable detection of DOS/DDOS flooding attacks.

Monitoring inflow measurements of water resource recovery facilities (WRRFs) is essential to promptly detect abnormalities and helpful in the decision making of the operators to better optimize, take corrective actions, and maintain downstream processes. In this paper, we introduced a flexible and reliable monitoring soft sensor approach to detect and identify abnormal influent measurements of WRRFs to enhance their efficiency and safety. The proposed data-driven soft sensor approach merges the desirable characteristics of principal component analysis (PCA) with k-nearest neighbor (KNN) scheme. PCA performed effective dimension reduction and revealed interrelationships between inflow measurements, while KNN distances demonstrated superior detection capacity, robustness to underlying data distribution, and efficiency in handling high-dimensional dataset. Furthermore, nonparametric thresholds derived from kernel density estimation further enhanced detection results of PCA-KNN approach when compared with parametric counterparts. Moreover, the radial visualization plot is innovatively employed for fault analysis and diagnosis in combination with PCA and delineated interpretable visualization of anomalies and detector performances. The effectiveness of these soft sensor schemes is evaluated by using real data from a coastal municipal WRRF located in Saudi Arabia. Also, we compared the proposed soft sensor scheme with the conventional PCA-based approaches, including standard prediction error, Hotelling’s T2, and joint univariate methods. Results demonstrate that this soft sensor-based monitoring approach outperforms conventional PCA-based methods.