Wyniki wyszukiwania - BazTech

1

An approach to identify the interactions between the control factors in a Mahalanobis-Taguchi System

Labidi Amal, Tanabe Ikuo, Tanabe Yuko, Isobe Hiromi

Journal of Machine Engineering

|

2022

|

Vol. 22, No. 1

96--110

EN

The Mahalanobis-Taguchi System (MTS) is, today, widely used to define the optimal conditions for the design stage of product development especially, in the field of Artificial Intelligence (AI) considering the non-linear properties and non-digital data. In this paper, an approach to identify the several interactions in a MTS is proposed. The MTS contains four methods; Mahalanobis-Taguchi (MT) method, Mahalanobis Taguchi Adjoint (MTA) method, Recognition Taguchi (RT) method and Taguchi (T) method. The method to use for the analysis is selected based on the system’s properties. For the case of study used in this research, the unit space is created through the RT method and used to calculate the Mahalanobis-Taguchi distances (MTD). For the method proposed in this paper, the relationships between control factors and MTDs were firstly clarified by MTS (RT), then the same relationships were clarified using a modified design of experiments method, and the several interactions between control factors in MTS (RT) were finally identified by comparing the two relationships. Then effectiveness of the proposed method was evaluated by using a mathematical model.

2

Diagnosis of air compressor condition using minimum redunduncy maximum relevance (MRMR) algorithim and distance metric based classification

Al-Bugharbee Hussein, Samaka Hatem, Zubaidi Salah L.

Diagnostyka

|

2021

|

Vol. 22, No. 4

25--32

EN

Finding a reliable machines condition monitoring technique has been attracted many researchers to avoid the sudden failure in machines and the unexpected consequences. This work proposes a fault diagnosis of air compressors using frequency-based features and distance metric-based classification. The analyzed experimental datasets contain one healthy condition and seven different fault conditions. Features are extracted from the frequency spectrum, then the best feature sets are selected using MRMR algorithm and eventually the classification is conducted using a distance metric classifier. The results demonstrated the automatic classification with more than 97% correct classification rate. The effect of selected feature set size, training sample size on the classification accuracy is also investigated. From the results, this method of analysis can be used for early detection of faults with very great accuracy.

3

Wykrywanie anomalii bazujące na wskazanych przykładach

Kwiatkowski W.

Przegląd Teleinformatyczny

|

2018

|

T. 6, Nr 1-2 (46)

3--21

PL

Rozpatrywany jest problem wykrywania anomalii na podstawie zarejestrowanych obserwacji zachowania systemu. Problem jest sformułowany jako zadanie rozpoznawania wzorców zachowania normalnego i zachowania nietypowego. Obydwa wzorce są określane przez wskazanie odpowiednich przykładów. Osobliwość rozwiązywanego zadania wynika z faktu, że zwykle liczebność przykładów jest dużo mniejsza od wymiaru wektora obserwacji. W artykule zostały przedstawione dwie metody detekcji anomalii bazujące na wyznaczaniu rzutów obserwacji na podprzestrzenie wzorców. Wyróżnikiem pierwszej metody jest wykorzystywanie odległości wektora obserwacji od podprzestrzeni wzorców. Druga metoda polega na przeniesieniu zadania rozpoznawania wzorców do podprzestrzeni wzorców.

EN

The paper considers the issue of anomalies detection based on registered observations of a system behavior. The problem is formulated as recognition of normal and anomalous behavior patterns. Both types of patterns are identified by indication of appropriate examples. A peculiarity of this task is that usually the number of examples is far lower than the dimension of vectors describing the observations. Two methods to solve this task have been presented in the paper, based on projecting the observations on the subspace of examples. The first method is based on a distance of the observation vector from the subspace of examples. The second method is based on transferring the pattern recognition problem to the subspace of examples.

4

The structure of classifiers of hand gestures with the use of the active contour model and Fourier descriptors

Golański P., Szczekala M.

Aviation Advances & Maintenance

|

2018

|

Vol. 41, Issue 1

63--77; 78--92

EN

In this article, the research results of the usage of selected methods of the analysis of images for the recognition of hand gestures in human-computer interaction was depicted. The usage of this type of interaction is important in case of the so-called wearable computers (computer is integrated with the work clothing of an operator. For the recognition of gestures, the combination of two methods associated with the image processing was suggested and that is the Chan-Vese active contour model enabling to recognize objects on a given image, based on the curve evolution technique, Mumford-Shah functional, level-sets and the methods to create shapes with the use of Fourier descriptors. For the classification criteria as a compatibility measure a scalable Mahalanobis distance was used.

PL

W artykule przedstawiono wyniki badań nad wykorzystaniem wybranych metod analizy obrazów do detekcji gestów dłoni w komunikacji człowiek–komputer. Wykorzystanie tego typu komunikacji ma duże znaczenie w przypadku obsługi komputerów zintegrowanych z odzieżą roboczą operatora, tzw. komputerów do noszenia (wearables computers). Do detekcji gestów zaproponowano połączenie dwóch metod związanych z obróbką obrazu: metodę aktywnych konturów Chana-Vese, umożliwiającą wykrywanie obiektów na danym obrazie, opartą na technikach ewolucji krzywych, funkcjonale Mumforda-Shaha oraz zbiorach poziomicowych, oraz metodę tworzenia klasyfikatorów kształtów z wykorzystaniem deskryptorów Fouriera. Do kryterium klasyfikacji jako miarę zgodności z wzorcem wykorzystano skalowaną odległość Mahalanobisa.

5

A study on welding quality for the automatic vertical-position welding process based on Mahalanobis Distance method

Jin B.-J., Park M.-H., Yun T.-J., Shim J.-Y., Kang B.-Y., Kim I.-S.

Archives of Materials Science and Engineering

|

2018

|

Vol. 91, nr 1

31--40

EN

Purpose: The welding quality and reducing production cost could be achieved by developing the automatic on-line welding quality monitoring system. However, investigation of welding fault to quantify the welding quality on the horizontal-position welding has been concentrated. Therefore, MD (Mahalanobis Distance) method on the vertical-position welding process by analysing the transform arc voltage and welding current gained from the on-line monitoring system has been applied. Design/methodology/approach: The transformed welding current and arc voltage data were taken from the experiment whereby the data number was 2500 data/s. The prediction of Contact Tip to Work Distance (CTWD) to gain best welding quality using the waveform variations were then taken from the experimental results. MD was employed to quantify the welding quality by analysing the transformed arc voltage and welding current. Finally, the optimal CTWD setting has verified the developed algorithms through additional experiments. Two kinds of experiments has been carried out by changing welding parameters artificially to verify the sensitivity and feasibility of WQ (Welding Quality) based on the concepts of MD and normal distribution. Findings: The results represented that WQ was fully capable of quantifying and qualifying the welding faults for automatic vertical-position welding process. Research limitations/implications: The arc welding process on the vertical-position compared to a horizontal-position welding is much more difficult because the metal transfer is influenced by the gravity force. To solve the problem, a new algorithm to monitor and control the welding fault during the arc welding process has been developed. Furthermore, optimization of welding parameters for the vertical-position welding process was really difficult to use the developed algorithms because they are only useful in selecting stored data and not for evaluating the effect of the variation of welding parameters on the weld ability. Practical implications: The developed algorithm could be achieved the highest welding quality at 15mm CTWD setting which the welding quality is 99.50% for the start section and 99.68% at the middle section. Originality/value: This paper proposed a new algorithm which employed the concepts of MD (Mahalanobis Distance) and normal distribution to describe a good quality welding.

6

Application of Improved Robust Kalman Filter in Data Fusion for PPP/INS Tightly Coupled Positioning System

Li Z., Yao Y., Wang J., Gao J.

Metrology and Measurement Systems

|

2017

|

Vol. 24, nr 2

289--301

EN

A robust Kalman filter improved with IGG (Institute of Geodesy and Geophysics) scheme is proposed and used to resist the harmful effect of gross error from GPS observation in PPP/INS (precise point positioning/inertial navigation system) tightly coupled positioning. A new robust filter factor is constructed as a three-section function to increase the computational efficiency based on the IGG principle. The results of simulation analysis show that the robust Kalman filter with IGG scheme is able to reduce the filter iteration number and increase efficiency. The effectiveness of new robust filter is demonstrated by a real experiment. The results support our conclusion that the improved robust Kalman filter with IGG scheme used in PPP/INS tightly coupled positioning is able to remove the ill effect of gross error in GPS pseudorange observation. It clearly illustrates that the improved robust Kalman filter is very effective, and all simulated gross errors added to GPS pseudorange observation are successfully detected and modified.

7

Reconstruction of hit time and hit position of annihilation quanta in the J-PET detector using the Mahalanobis distance

Sharma N. G., Silarski M., Bednarski T., Białas P., Czerwiński E., Gajos A., Gorgol M., Jasińska B., Kamińska D., Kapłon Ł., Korcyl G., Kowalski P., Kozik T., Krzemień W., Kubicz E., Niedźwiecki S., Pałka M., Raczyński L., Rudy Z., Rundel O., Słomski A., Strzelecki A., Wieczorek A., Wiślicki W., Zieliński M., Zgardzińska B., Moskal P.

Nukleonika

|

2015

|

Vol. 60, No. 4, part 1

765--769

EN

The J-PET detector being developed at the Jagiellonian University is a positron emission tomograph composed of the long strips of polymer scintillators. At the same time, it is a detector system that will be used for studies of the decays of positronium atoms. The shape of photomultiplier signals depends on the hit time and hit position of the gamma quantum. In order to take advantage of this fact, a dedicated sampling front-end electronics that enables to sample signals in voltage domain with the time precision of about 20 ps and novel reconstruction method based on the comparison of examined signal with the model signals stored in the library has been developed. As a measure of the similarity, we use the Mahalanobis distance. The achievable position and time resolution depend on the number and values of the threshold levels at which the signal is sampled. A reconstruction method as well as preliminary results are presented and discussed.

8

Binary Classification of Heart Failures Using k-NN with Various Distance Metrics

Udovychenko Y., Popov A., Chaikovsky I.

International Journal of Electronics and Telecommunications

|

2015

|

Vol. 61, No. 4

339-344

EN

Magnetocardiography is a sensitive technique of measuring low magnetic fields generated by heart functioning, which is used for diagnostics of large number of cardiovascular diseases. In this paper, k-nearest neighbor (k-NN) technique is used for binary classification of myocardium current density distribution maps (CDDM) from patients with negative T-peak, male and female patients with microvessels (diffuse) abnormalities and sportsmen, which are compared with normal control subjects. Number of neighbors for k-NN classifier was selected to obtain highest classification characteristics. Specificity, accuracy, precision and sensitivity of classification as functions of number of neighbors in k-NN are obtained for classification with several distance measures: Mahalanobis, Cityblock, Eucleadian and Chebyshev. Increase of the accuracy of classification for all groups up to 10% was obtained using Cityblock distance metric in binary k-NN classifier with 19 - 27 neighbors, comparing to other metrics. Obtained results are acceptable for further patient’s state evaluation.

9

Geodesic survey and modernization of a route as the task of optimization

Nowak E.

Geodesy and Cartography

|

2014

|

Vol. 63, no. 1

75--87

EN

A geodesic survey of an existing route requires one to determine the approximation curve by means of optimization using the total least squares method (TLSM). The objective function of the LSM was found to be a square of the Mahalanobis distance in the adjustment field ν. In approximation tasks, the Mahalanobis distance is the distance from a survey point to the desired curve. In the case of linear regression, this distance is codirectional with a coordinate axis; in orthogonal regression, it is codirectional with the normal line to the curve. Accepting the Mahalanobis distance from the survey point as a quasi-observation allows us to conduct adjustment using a numerically exact parametric procedure. Analysis of the potential application of splines under the NURBS (non-uniform rational B-spline) industrial standard with respect to route approximation has identified two issues: a lack of the value of the localizing parameter for a given survey point and the use of vector parameters that define the shape of the curve. The value of the localizing parameter was determined by projecting the survey point onto the curve. This projection, together with the aforementioned Mahalanobis distance, splits the position vector of the curve into two orthogonal constituents within the local coordinate system of the curve. A similar system corresponds to points that form the control polygonal chain and allows us to find their position with the help of a scalar variable that determines the shape of the curve by moving a knot toward the normal line.

PL

Inwentaryzacja istniejącej tras y wymaga wyznaczenia krzywej aproksymującej w wyniku optymalizacji realizowanej metodą najmniejszych kwadratów (TLSM). Analiza funkcji celu LSM wykazała, że jest ona kwadratem odległości Mahalanobisa w przestrzeni poprawek. W zadaniach aproksymacyjnych odległość Mahalanobisa jest miarą odstępu pikiety od wyznaczanej krzywej, w przypadku regresji zwykłej odstęp ten ma kierunek osi układu współrzędnych a w przypadku regresji ortogonalnej odstęp ma kierunek normalnej do krzywej. Uznanie odległości Mahalanobisa pikiety od wyznaczanej krzywej za quasi-obserwację pozwala na wykonanie wyrównania dopracowaną numerycznie procedurą parametryczną. Badanie możliwości zastosowania funkcji sklejanych w przemysłowym standardzie NURBS do aprosymacji przebiegu trasy wykazało dwa problemy: brak wartości parametru lokalizującego dla pikiety oraz operowanie parametrami wektorowymi defi lującymi kształt krzywej. Wartość parametru lokalizującego wyznaczono przez rzut pikiety na krzywą - łącznie z opisaną wyżej odległością Mahalanobisa stanowi on rozkład wektora wodzącego pikiety na dwie składowe podłużną i poprzeczną w lokalnym układzie krzywej. Analogiczny układ w punktach tworzących łamaną kontrolną funkcji Beziera pozwala na wyznaczenie ich położenia za pośrednictwem niewiadomej skalarnej modelującej kształt krzywej poprzez przesunięcia węzła w kierunku normalnej.

10

Data mining system for air quality monitoring networks

Czechowski P., Badyda A., Majewski G.

Archives of Environmental Protection

|

2013

|

Vol. 39, no. 4

123--147

EN

The use of quantitative methods, including stochastic and exploratory techniques in environmental studies does not seem to be sufficient in practical aspects. There is no comprehensive analytical system dedicated to this issue, as well as research regarding this subject. The aim of this study is to present the Eco Data Miner system, its idea, construction and implementation possibility to the existing environmental information systems. The methodological emphasis was placed on the one-dimensional data quality assessment issue in terms of using the proposed QAAH1 method - using harmonic model and robust estimators beside the classical tests of outlier values with their iterative expansions. The results received demonstrate both the complementarity of proposed classical methods solution as well as the fact that they allow for extending the range of applications significantly. The practical usefulness is also highly significant due to the high effectiveness and numerical efficiency as well as simplicity of using this new tool.

11

Wavelet energy-based mahalanobis distance metric for testing analog and mixed-signal circuits

Spyronasios A. D., Dimopoulos M. G., Hatzopoulos A. A.

International Journal of Microelectronics and Computer Science

|

2010

|

Vol. 1, nr 2

218-224

EN

In this paper a test method based on the wavelet transformation of the measured signal, be it supply current (Ips) or output voltage (Vout) waveform, is presented. In the wavelet analysis, a Mahalanobis distance test metric is introduced utilizing information from the wavelet energies of the first decomposition level of the measured signal. The tolerance limit for the good circuit is set by statistical processing data obtained from the fault-free circuit. Simulation comparative results on benchmark circuits for testing both hard faults and parametric faults are presented showing the effectiveness of the proposed testing scheme.

12

Wstępna selekcja kanałów spektralnych w procesie rozpoznania obrazowego

Walczykowski P., Orych A., Paluchowski Ł.

Biuletyn Wojskowej Akademii Technicznej

|

2010

|

Vol. 59, nr 2

325-337

PL

Ciągłe ulepszanie metod kamuflażu przyczynia się do poszukiwania malejących różnic w odbiciu spektralnym pomiędzy obiektami a tłem naturalnym. Główne problemy wynikające z natury tła i materiałów wykorzystywanych do kamuflażu to sposób prowadzenia rozpoznania, wybór kanałów spektralnych, dobór algorytmów umożliwiających przetworzenie zdjęć i poprawę kontrastu oraz metody wizualizacji wyników. W przeprowadzonych badaniach zastosowano algorytm do sprawdzania kontrastu na zobrazowaniach hiperspektralnych. Poddano analizie porównawczej metody wykrywania obiektów oparte na pojedynczych zobrazowaniach, dwóch kanałach spektralnych oraz metodę automatycznego tworzenia kompozycji hiperspektralnej. Dodatkowo sklasyfikowano metody pod kątem wyróżnienia obiektów o znanej i nieznanej charakterystyce odbiciowej. Zastosowana metodyka badań jest oparta na "odległości Mahalanobisa" i wskazuje na potrzebę prowadzenia rozpoznania wielokanałowego w celu sprawnego wykrycia obiektów.

EN

Constant advances in methods of camouflage are responsible for the progress in image reconnaissance and the distinguishing between objects and their natural background. The main problems attributable to the nature of the background and materials used to camouflage the object are: the way in which image reconnaissance should be conducted, the choice of spectral bands used, the choice of algorithms used to process the images and methods of visualizing the results. In our studies we have applied an algorithm to evaluate the contrast of the acquired hyperspectral images. We carried out a comparative analysis of methods used to recognize objects based on single images, on two spectral bands and using an automated method of creating hyperspectral compositions. Additionally, the methods had been classified in terms of their ability to recognize objects with a known and unknown spectral curve. This methodology is based on the "Mahalanobis distance". It proves that there is a need to acquire multiband imagery information in order to make the process of object recognition more efficient.