|Article title||METHOD OF MONITORING WITH FORECASTING THE STATUS OF SENSOR NETWORKS BASED ON STATISTICS|
|Authors||D. A. Bespalov, A. S. Boldyrev|
|Section||SECTION IV. MONITORING AND CONTROL IN TECHNICAL SYSTEMS|
|Month, Year||03, 2018 @en|
|Abstract||In the modern world, there are more and more data sources that are used in distributed systems to solve problems of analysis, management and notification. They have a different physical nature and degree of complexity. Examples of such sources for intelligent systems are sensors and sensor networks. The method for monitoring sensor networks based on statistics to solve the problem of detection of anomalies and forecasting is developed. The proposed method is based on the exponential smoothing using different-scale statistics. Theoretical prerequisites of the work are based on the interpretation of the behavior of the object as a system with normal and anomalous areas of work, estimated by the sensor network. In this case, both input and output time series of parameters can be interpreted as approximations with a certain accuracy, as well as with zero variability or the presence of artifacts of behavior. The effectiveness of such an approach for perimeter security systems, prevention of penetration into the protected area or increase of the level of danger in a certain area is determined by the possibility of preliminary threat assessment on the spot and issuing an appropriate alert to the central node of the distributed system. In this regard, this article also describes the structure of the distributed and distributed information monitoring system with prediction, implemented in compliance with the modern concept of microservices, the use of modern software technologies and specialized databases. Such a system is scalable both horizontally and vertically, which makes it effective for practical use. The novelty of this approach is the use of a non-standard prediction algorithm for the analysis of signals of sensor systems and in combination with the method of preliminary data processing by means of wavelet analysis.|
|Keywords||Density networks; statistics; anomaly; prediction; trend; smoothing; series.|
|References||1. Boldyreff A.S., Bespalov D.A., Adzhiev A.K. Automated information-analytical system for thunderstorm monitoring and early warning alarms using modern physical sensors and information technologies with elements of artificial intelligence, Proceedings of SPIE. "Autonomous Air and Ground Sensing Systems for Agricultural Optimization and Phenotyping II", 2017, pp. 102180P.
2. Bespalov D.A. Predvaritel'naya obrabotka vremennykh ryadov parametrov v zadache obnaruzheniya anomaliy v rabote informatsionnykh sistem, ispol'zuyushchikh plastikovye karty [Pre-processing of time series of parameters in the problem of detection of anomalies in the information systems using plastic cards], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engineering Sciences], 2017, No. 5-6 (190-191), pp. 48-65.
3. Bespalov D.A., Anan'ev A.A. Sposob obnaruzheniya anomaliy v rabote informatsionnykh sistem, ispol'zuyushchikh plastikovye karty [Method of detection of anomalies in the information systems using plastic cards], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engineering Sciences], 2017, No. 5-6 (190-191), pp. 38-48.
4. Helge Brügner. Holt-Winters Traffic Prediction on Aggregated Flow Data. Network Architectures and Services, September 2017. Seminars FI / IITM SS 17, pp. 25-32.
5. Bermúdez J.D., Segura J.V. & Vercheri E. Bayesian forecasting with the Holt–Winters model, Journal of the Operational Research Society, 2010, Vol. 61, pp. 164-171.
6. Gelper S., Fried R. & Croux C. Robust forecasting with exponential and Holt-Winters smoothing, Journal of Forecasting, 2010, Vol. 29, pp. 285-300.
7. Adzhiev Anatoly Kh., Boldyrev Anton S., Bolgov Yuriy V., Manfred Wendisch, Bondareva Olga. V. Advanced remote sensing of thunderstorm events and atmospheric electric field, Proceedings of SPIE "Remote Sensing of Clouds and the Atmosphere XXII" 2017, pp. 104240M.
8. Boldyrev A.S., Boldyreva K.A. Modeling of the Electric Field near the Surface Layer under Strong Turbulent Mixing. Geophysical Research Abstracts. Vol. 17, EGU General Assembly, 2015.
9. Adzhiev A.Kh., Boldyreff A.S., Kazakova S. Method of Thunderstorm Activity Monitoring Using Lightning Sensors and Electric Field Mills. Geophysical Research Abstracts. Vol. 17, EGU General Assembly, 2015.
10. Adzhiev Anatoly, Boldyreff Anton, Kuliev Dalhat, Kondratyeva Natalia and Chochayev Khizir. Characteristics of Thunderstorm Activity and Parameters of Lightning in the South of Russia // European Conference on Severe Storms 2015 14;18 September 2015, Wiener Neustadt, Austria ECSS2015-91.
11. John Shahid. InfluxDB Documentation. Release 4.1.1. Sep. 13, 2017, 21 p.
12. Syeda Noor Zehra Naqvi. Sofia Yfantidou. Time Series Databases and InfluxDB. Universite libre de Bruxelles. Advanced Databases Winter Semester 2017-2018. December 17, 2017, 41 p.
13. Andrew Lahiff. Monitoring with InfluxDB and Grafana. STFC RAL. HEPiX 2015 Fall Workshop, BNL. UK Computing for Particle Physics. 2015, 30 p.
14. Josiah L. Carkson. Redis in Action. Manning Publications. Shelter Island. 2013, 322 p.
15. Wes McKinney. Python for Data Analysis. OReilly Media, Inc, 2013, 470 p.
16. Jake VanderPlas. Python Data Science Handbook. OReilly Media, Inc, 2017, 548 p.
17. Arnaud Beck. Data analysis with python. Laboratoire Leprince-Ringuet, École Polytechnique, CNRS/IN2P3.
18. Joel Grus. Data science from scratch. OReilly Media, Inc, 2015, 464 p.
19. Edouard Duchesnay, Tommy Lofstedt. Statistics and Machine Learning in Python. Release 2. Jun 22, 2018, 201 p.
20. Allen B. Downey. Think Stats. Green Tea Press, Needham, Massachusetts. 2014, 264 p.
21. Malla S. Veyvlety v obrabotke signalov [Wavelets in signal processing]. Moscow: Izd-vo Mir, 2005, 672 p.