Journal of Systems Engineering and Electronics ›› 2006, Vol. 17 ›› Issue (4): 896-900.doi: 10.1016/S1004-4132(07)60034-6

• SOFTWARE ALGORITHM AND SIMULATION • Previous Articles     Next Articles

Decision feedback equalizer based on non-singleton fuzzy regular neural networks

Song Heng1, Wang Chen1, He Yin2, Ma Shiping& Zuo Jizhang1
  

  1. 1.School of Engineering, Air Force Engineering Univ., Xi'an 710038, P. R. China; 2.School of Information Engineering, Fudan Univ., Shanghai 210433, P.R. China.
  • Online:2006-12-25 Published:2019-12-20

Abstract:

A new equalization method is proposed in this paper for severely nonlinear distorted channels. The structure of decision feedback is adopted for the non-singleton fuzzy regular neural network that is trained by gradient-descent algorithm. The model shows a much better performance on anti-jamming and nonlinear classification, and simulation is carried out to compare this method with other nonlinear channel equalization methods. The results show the method has the least bit error rate (BER).

Key words: