Title :
A neural network for spectrum estimation of quasi-stationary signal
Author :
Chen, Qiming ; Zhang, Shujing
Author_Institution :
Inst. of Inf. Sci., Northern Jiaotong Univ., Beijing, China
Abstract :
The authors propose a system for adaptive processing of quasi-stationary signals, which accomplishes pattern clustering and recognition with spectral characteristics of stationary segments. With input being AR coefficients, Kohonen´s self-organizing neural network plays a significant role in determining model transition time, pattern clustering and recognition, adaptive spectral estimation, and discovery of tendency or regularity in quasi-stationary signals. Finally, simulation results are given to prove its performance
Keywords :
pattern recognition; self-organising feature maps; signal processing; spectral analysis; AR coefficients; Kohonen´s self-organizing neural network; model transition time; neural network; pattern clustering; pattern recognition; quasi-stationary signal; simulation; spectral characteristics; spectrum estimation; Adaptive systems; Brain modeling; Filters; Gaussian noise; Humans; Neural networks; Parameter estimation; Signal resolution; Spectral analysis; Time series analysis;
Conference_Titel :
Industrial Electronics, 1992., Proceedings of the IEEE International Symposium on
Conference_Location :
Xian
Print_ISBN :
0-7803-0042-4
DOI :
10.1109/ISIE.1992.279540
