Другие журналы
scientific edition of Bauman MSTU
SCIENCE & EDUCATION
Bauman Moscow State Technical University. El № FS 77 - 48211. ISSN 1994-0408
Adaptation of the Neural Network Recognition System of the Helicopter on Its Acoustic Radiation to the Flight Speed
# 05, May 2015
DOI: 10.7463/0515.0776347
Article file:
SE-BMSTU...o153.pdf
(1263.22Kb)
SE-BMSTU...o153.pdf
(1263.22Kb)
The article concerns the adaptation of a neural tract that recognizes a helicopter from the aerodynamic and ground objects by its acoustic radiation to the helicopter flight speed. It uses non-centered informative signs-indications of estimating signal spectra, which correspond to the local extremes (maximums and minimums) of the power spectrum of input signal and have the greatest information when differentiating the helicopter signals from those of tracked vehicles. The article gives justification to the principle of the neural network (NN) adaptation and adaptation block structure, which solves problems of blade passage frequency estimation when capturing the object and track it when tracking a target, as well as forming a signal to control the resonant filter parameters of the selection block of informative signs. To create the discriminatory characteristics of the discriminator are used autoregressive statistical characteristics of the quadrature components of signal, obtained through the discrete Hilbert Converter (DGC) that perfor
Mathematical modeling of the tracking meter using the helicopter signals obtained in real conditions is performed. The article gives estimates of the tracking parameter when using a tracking meter with DGC by sequential records of realized acoustic noise of the helicopter. It also shows a block-diagram of the adaptive NN. The scientific novelty of the work is that providing the invariance of used informative sign, the counts of local extremes of power spectral density (PSD) to changes in the helicopter flight speed is reached due to adding the NN structure and adaptation block, which is implemented as a meter to track the apparent passage frequency of the helicopter rotor blades using its relationship with a function of the autoregressive acoustic signal of the helicopter.
Specialized literature proposes solutions based on the use of training classifiers with different parametric methods of spectral representations, in particular, linear prediction and cepstrum, as well as methods based on wavelet transformations and robust learning. Adaptive approach allows solving tasks in a wide range of changing helicopter speeds.
References
- Astapov Yu.M., Kozlov V.I., Soboleva N.S., Khokhlov V.K.; Borzov A.B., ed. Avtonomnye informatsionnye i upravliaiushchie sistemy: Trudy kafedry “Avtonomnye infor-matsionnye i upravlyayushchie sistemy” MGTU im. N.E. Baumana. V 4 t. T. 1 [Autonomous information and control systems: Proc. of department “Autonomous information and control systems” of Bauman MSTU. In 4 vols. Vol. 1]. Moscow, JSC “Inzhener” Publ., JSC “Oniko-M” Publ., 2011. 468 p. (in Russian).
- Khokhlov V.K. The Initial Regression Statistical Characteristics of Intervals Between Zeros of Random Processes. Nauka i obrazovanie MGTU im. N.E. Baumana = Science and Education of the Bauman MSTU , 2014, no. 9, pp. 132-147. DOI: 10.7463/0914.0726720 (in Russian).
- Pavlov G.L., Khokhlov V.K. Neural network algorithms for classification problem of objects according to their acoustic radiation. Nauka i obrazovanie MGTU im. N.E. Baumana = Science and Education of the Bauman MSTU , 2012, no. 5, pp. 247-258. DOI: 10.7463/0512.0367620 (in Russian).
- Pavlov G.L., Khokhlov V.K. Adaptation of neural network algorithm to velocities classified by acoustic radiation facilities. Nauka i obrazovanie MGTU im. N.E. Baumana = Science and Education of the Bauman MSTU , 2012, no. 10, pp. 241-250. DOI: 10.7463/1012.0462849 (in Russian).
- Khokhlov V.K., Gulin Iu.Iu. Informative Signs Selection in Autonomous Information Systems with Neuron Net Paths of Signal Processing. Vestnik MGTU im. N.E. Baumana. Ser. Priborostroenie = Herald of the Bauman Moscow State Technical University. Ser. Instrument Engineering , 2003, no. 3, pp. 70-83. (in Russian).
- Tang Haifeng, Sun Degang. Real Time Multisensor Target Recognition Based on DSP. 8th International Conference on Electronic Measurement and Instruments (ICEMI '07). IEEE Publ., 2007, pp. 4-24 – 4-28. DOI: 10.1109/ICEMI.2007.4351127
- Elshafei M., Akhtar S., Ahmed M.S. Parametric models for helicopter identification using ANN. IEEE Transactions on Aerospace and Electronic Systems, 2000, vol. 36, iss. 4, pp. 1242-1252. DOI:10.1109/7.892672
- Du Yinggang, Lu Jinhui, Shi xiangquan, GuYalin. Target identification based on the optimal base number. Proceedings 1998 Fourth International Conference on Signal Processing (ICSP '98). Vol. 1 . IEEE Publ., 1998, pp. 271-274. DOI:10.1109/ICOSP.1998.770204
- Moukas P., Simson J., Norton-Wayne L. Automatic Identification of Noise Pollution Sources. IEEE Transactions on Systems, Man and Cybernetics , 1982, vol. 12, iss. 5, pp. 622-634. DOI:10.1109/TSMC.1982.4308881
- Khokhlov V.K., Pylaev V.A., Volchikhin I.V., Stepanenko N.V. Pelengator istochnikov akusticheskikh izluchenii [D irection finder of sources of acoustic radiation ]. Patent RF, no. 2048678, 1995. (in Russian).
Publications with keywords: neural network, recognition, a helicopter, informative features of signals, neural adaptation network, passive acoustic information system, power spectral density
Publications with words: neural network, recognition, a helicopter, informative features of signals, neural adaptation network, passive acoustic information system, power spectral density
See also:
Поделиться:
| Authors |
| Press-releases |
| Library |
| Conferences |
| About Project |
| Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00) |
|
 |
|||
| © 2003-2024 «Наука и образование» Перепечатка материалов журнала без согласования с редакцией запрещена Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00) | |||||
