Detecting Anomalies in the Fm Frequency Band Using Statistical Methods
| DOWNLOAD | DOI: 10.62897/COS2023.1-1.54 |
|
Author: Szilárd L. Takács Széchenyi István University, Egyetem tér 1, 9026 Győr, Hungary |
Abstract:
Based on Act C of 2003 on electronic communications, in Hungary, the National Media and Infocommunications Authority is responsible for ensuring harmful interference-free frequency usage and electromagnetic compatibility. Continuous measurements are conducted nationwide in order to reach this goal, but the evaluation and analysis of anomalies are time-consuming.
This research focuses on the detection of anomalies in the FM radio frequency spectrum. Within that, the study was concerned with the outages of radio transmission and the outages of modulation. The goal of this study is to automate the detection process, providing real-time alerts for potential anomalies and saving valuable time for spectrum monitoring engineers.
In order to solve the problem, statistical learning was used, including classification algorithms. Comparing the following algorithms: k nearest neighbor classification method, logistic regression, linear discriminant analysis, quadratic discriminant analysis, naive Bayes classification, support vector machines, and random forests. The most efficient method for this is Support Vector Machines, which can identify the phenomena with 93.28 % accuracy.
Statistical machine learning is highly efficient at identifying known phenomena in spectrum monitoring and generating real-time alerts. Alerts can be generated within a minute, effectively providing real-time information.
REFERENCES
Alpaydin E., 2020, Introduction to machine learning, MIT press, Cambridge, UK, 140-145.
Ghojogh B., Crowley M., 2019, Linear and Quadratic Discriminant Analysis: Tutorial, arXiv preprint arXiv:1906.02590.
Hart E. P., Stork, G. D., Duda O. R., 2000, Pattern classification. Wiley, Hoboken, NJ, United States.
Hastie T., Tibshirani R., Friedman J.H., 2017. The elements of statistical learning: data mining, inference, and prediction, Second edition, corrected at 12th printing 2017. 2nd Ed, Springer series in statistics. Springer, New York, NY, United States, DOI: 10.1007/b94608.
Kotsiantis S. B., 2013, Decision trees: a recent overview. Artificial Intelligence Review, 39, 261-283.
National Media and Infocommunications Authority decree - About the national frequency allocation and the rules for the use of frequency bands. In: (7/2015. (XI. 13.)), Hungarian National Legislation-repository.
Peng C., Hu W., Wang L., 2022, Spectrum anomaly detection based on spatio-temporal network prediction. Electronics, 11(11), 1770.
Rajendran S., Meert W., Lenders V., Sofie P., 2019, Unsupervised wireless spectrum anomaly detection with interpretable features. IEEE Transactions on Cognitive Communications and Networking, 5(3), 637-647.
Russell S., Norvig P., 2005, Artificial Intelligence A Modern Approach. Hungarian Translation Panem, HU, 635- 638, 726- 727
Takács S. L., 2023, Detection of FM transmission outages and modulation outages with classification methods, MSc Dissertation, Széchenyi István University, Győr, Hungary.
Yokkampon U., Chumkamon S., Mowshowitz A., 2020, Anomaly detection using variational autoencoder with spectrum analysis for time series data. In 2020 Joint 9th International Conference on Informatics, Electronics & Vision (ICIEV) and 2020 4th International Conference on Imaging, Vision & Pattern Recognition (icIVPR) (pp. 1-6). IEEE.
Wang X. Y., Qian R. R., Huang C. X., Huang M., Yang J. J., 2022, An Anomaly Detector Using Filtering Stockwell Transform and Siamese Convolutional Neural Network in Radio Monitoring (AnoFSTSCNN). URSI Radio Science Letters, 4, 35.