Now showing 1 - 3 of 3
  • Publication
    ZeChipC: Time Series Interpolation Method Based on Lebesgue Sampling
    In this paper, we present an interpolation method based on Lebesgue sampling that could help to develop systems based time series more efficiently. Our methods can transmit times series, frequently used in health monitoring, with the same level of accuracy but using much fewer data. Our method is based in Lebesgue sampling, which collects information depending on the values of the signal (e.g. the signal output is sampled when it crosses specific limits). Lebesgue sampling contains additional information about the shape of the signal in-between two sampled points. Using this information would allow generating an interpolated signal closer to the original one. In our contribution, we propose a novel time-series interpolation method designed explicitly for Lebesgue sampling called ZeChipC. ZeChipC is a combination of Zero-order hold and Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) interpolation. ZeChipC includes new functionality to adapt the reconstructed signal to concave/convex regions. The proposed methods have been compared with state-of-the-art interpolation methods using Lebesgue sampling and have offered higher average performance.
      12
  • Publication
    Cutting Through the Emissions: Feature Selection from Electromagnetic Side-Channel Data for Activity Detection
    Electromagnetic side-channel analysis (EM-SCA) has been used as a window to eavesdrop on computing devices for information security purposes. It has recently been proposed to use as a digital evidence acquisition method in forensic investigation scenarios as well. The massive amount of data produced by EM signal acquisition devices makes it difficult to process in real-time making on-site EM-SCA infeasible. Uncertainty surrounds the precise information leaking frequency channel demanding the acquisition of signals over a wide bandwidth. As a consequence, investigators are left with a large number of potential frequency channels to be inspected; with many not containing any useful information leakages. The identification of a small subset of frequency channels that leak a sufficient amount of information can significantly boost the performance enabling real-time analysis. This work presents a systematic methodology to identify information leaking frequency channels from high dimensional EM data with the help of multiple filtering techniques and machine learning algorithms. The evaluations show that it is possible to narrow down the number of frequency channels from over 20,000 to less than a hundred (81 channels). The experiments presented show an accuracy of 0.9315 when all the 20,000 channels are used, an accuracy of 0.9395 with the highest 500 channels after calculating the variance between the average value of each class, and an accuracy of 0.9047 when the best 81 channels according to Recursive Feature Elimination are considered.
    Scopus© Citations 7  25
  • Publication
    Valve Health Identification Using Sensors and Machine Learning Methods
    Predictive maintenance models attempt to identify developing issues with industrial equipment before they become critical. In this paper, we describe both supervised and unsupervised approaches to predictive maintenance for subsea valves in the oil and gas industry. The supervised approach is appropriate for valves for which a long history of operation along with manual assessments of the state of the valves exists, while the unsupervised approach is suitable to address the cold start problem when new valves, for which we do not have an operational history, come online. For the supervised prediction problem, we attempt to distinguish between healthy and unhealthy valve actuators using sensor data measuring hydraulic pressures and flows during valve opening and closing events. Unlike previous approaches that solely rely on raw sensor data, we derive frequency and time domain features, and experiment with a range of classification algorithms and different feature subsets. The performing models for the supervised approach were discovered to be Adaboost and Random Forest ensembles. In the unsupervised approach, the goal is to detect sudden abrupt changes in valve behaviour by comparing the sensor readings from consecutive opening or closing events. Our novel methodology doing this essentially works by comparing the sequences of sensor readings captured during these events using both raw sensor readings, as well as normalised and first derivative versions of the sequences. We evaluate the effectiveness of a number of well-known time series similarity measures and find that using discrete Frechet distance or dynamic time warping leads to the best results, with the Bray-Curtis similarity measure leading to only marginally poorer change detection but requiring considerably less computational effort.
    Scopus© Citations 2  277