Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Anomaly detection is an important problem that has been researched within diverse research areas and application domains. Many anomaly detection techniques have been specifically developed for certain application domains, while others are more generic. This survey tries to provide a structured and comprehensive overview of the research on anomaly detection. We have grouped existing techniques into different categories based on the underlying approach adopted by each technique. For each category we have identified key assumptions, which are used by the techniques to differentiate between normal and anomalous behavior. When applying a given technique to a particular domain, these assumptions can be used as guidelines to assess the effectiveness of the technique in that domain. For each category, we provide a basic anomaly detection technique, and then show how the different existing techniques in that category are variants of the basic technique. This template provides an easier and more succinct understanding of the techniques belonging to each category. Further, for each category, we identify the advantages and disadvantages of the techniques in that category. We also provide a discussion on the computational complexity of the techniques since it is an important issue in real application domains. We hope that this survey will provide a better understanding of the different directions in which research has been done on this topic, and how techniques developed in one area can be applied in domains for which they were not intended to begin with.

/pdf/anomaly-detection-a-survey-1x33g9m0a3.pdf

Anomaly detection: A survey

物件導向軟體之架構(Object-Oriented Software Construction)探討

A review on machinery diagnostics and prognostics implementing condition-based maintenance

This report contains a review of the technical literature concerning the detection, location, and characterization of structural damage via techniques that examine changes in measured structural vibration response. The report first categorizes the methods according to required measured data and analysis technique. The analysis categories include changes in modal frequencies, changes in measured mode shapes (and their derivatives), and changes in measured flexibility coefficients. Methods that use property (stiffness, mass, damping) matrix updating, detection of nonlinear response, and damage detection via neural networks are also summarized. The applications of the various methods to different types of engineering problems are categorized by type of structure and are summarized. The types of structures include beams, trusses, plates, shells, bridges, offshore platforms, other large civil structures, aerospace structures, and composite structures. The report describes the development of the damage-identification methods and applications and summarizes the current state-of-the-art of the technology. The critical issues for future research in the area of damage identification are also discussed.

/pdf/damage-identification-and-health-monitoring-of-structural-3s8adkph2w.pdf

Damage identification and health monitoring of structural and mechanical systems from changes in their vibration characteristics: A literature review

Wireless sensor nodes with impedance measurement capabilities, often based on the Analog Devices AD5933
impedance chip and Atmel's 8-bit ATMega 1281 microcontroller, have been demonstrated to be effective in collecting
data for localized damage detection (such as for loose bolt detection) and for sensor self-diagnostics. Previouslydeveloped
nodes rely on radio telemetry and off-board processing (usually via a PC) to ascertain damage presence or
sensor condition. Recent firmware improvements for the wireless impedance device (WID) now allow seamless
integration of the WID with SHMTools and mFUSE, an open-source function sequencer and SHM process platform for
Matlab. Furthermore, SHM processes developed using mFUSE can be implemented in hardware on the WID, allowing
greater autonomy among the sensor nodes to identify and report damage in real time. This paper presents the capabilities
of the newly integrated hardware and software, as well as experimental validation.

/pdf/embedded-processing-for-shm-with-integrated-software-control-1w1kior9q3.pdf

Embedded processing for SHM with integrated software control of a wireless impedance device

Video cameras offer a versatile high resolution alternative to traditional sensor apparatuses for full field structural dynamics analysis. Previous work has extended phase based motion magnification techniques to achieve state of the art results in vibrational and structural analysis. One limitation on the current approach is that it does not work very well in situations where substantial rigid body motion is present in the scene. The large rigid motion drowns out the tiny vibrations that are relevant for a full-field structural analysis. This work extends upon previous phase based motion algorithms in order to allow them to work even in cases with large rigid body motion present. Through a keypoint tracking and video cropping scheme rigid body motion is subtracted from video streams, while preserving the smaller motions of interest. After this preprocessing step, the new video stream can be passed into the full-field analysis scheme as usual, in order to get a full-field modal decomposition.

Extraction of Full-Field Structural Dynamics from Digital Video Measurements in Presence of Large Rigid Body Motion

Health Monitoring of Pipeline Systems using Macro-fiber Composite Active-Sensors

Structural health monitoring is a process of evaluating and ensuring the safety and integrity of structural systems by converting sensor data into structural health monitoring information. The first and most important objective of structural health monitoring is to ascertain with confidence if damage is present or not. The idea is to characterize only the normal condition of a structure and the baseline data are used as a reference. When data are measured during continuous monitoring, the new data are compared with the reference. A significant deviation from the reference will signal damage. This decision-making procedure necessitates the establishment of a decision boundary. Choosing the decision boundary is often based on the assumption that the distribution of the data is Gaussian in nature. While the problem of damage detection focuses attention on the outliers or extreme values of the data, i.e. those points in the tails of the distribution, the establishment of the decision boundary based on the normality assumption weighs the central population of the data. This unwarranted assumption about the nature of the data distribution can significantly impair the performance of a structural health monitoring system by increasing false positive and negative indications of damage. In this study, a robust and automated damage classifier is developed by properly modeling the tails of the distribution using extreme value statistics. 1 Assistant Professor 2 Research Associate 3 Professor 4 Technical Staff Member

Minimizing misclassification of damage using extreme values statistics

The monitoring of adhesively-bonded joints through the use of ultrasonic guided waves is the general topic of this paper. Specifically, composite-to-composite joints representative of the wing skin-to-spar bonds of Unmanned Aerial Vehicles (UAVs) are examined. This research is the first step towards the development of an on-board structural health monitoring system for UAV wings based on integrated ultrasonic sensors. The study investigates two different lay-ups for the wing skin and two different types of bond defects, namely poorly-cured adhesive and disbonded interfaces. The guided wave propagation problem is studied numerically by a semi-analytical finite element method that accounts for viscoelastic damping, and experimentally by utilizing macro fiber composite (MFC) transducers which are inexpensive, flexible, highly robust, and viable candidates for application in on-board monitoring systems. Based upon change in energy transmission, the presence of damage is successfully identified through features extracted in both the time domain and discrete wavelet transform domain. A unique "passive" version of the diagnostic system is also demonstrated experimentally, whereby MFC sensors are utilized for detecting and locating simulated active damage in an aluminum plate. By exploiting the directivity behavior of MFC sensors, a damage location algorithm which is independent of wave speed is developed. Application of this approach in CFRP components may alleviate difficulties associated with damage location in highly anisotropic systems.

Charles R. Farrar

Papers

Embedded processing for SHM with integrated software control of a wireless impedance device

Extraction of Full-Field Structural Dynamics from Digital Video Measurements in Presence of Large Rigid Body Motion

Health Monitoring of Pipeline Systems using Macro-fiber Composite Active-Sensors

Minimizing misclassification of damage using extreme values statistics

Ultrasonic Guided Wave Monitoring of Composite Bonded Joints Using Macro Fiber Composite Transducers