Production of time domain fuzzy control rules for channel estimation and equalization in OFDM based communication systems


Ozen A., Soysal B., Kaya I. , Altas I. H.

IEEE 14th Signal Processing and Communications Applications, Antalya, Turkey, 16 - 19 April 2006, pp.245-246 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/siu.2006.1659685
  • City: Antalya
  • Country: Turkey
  • Page Numbers: pp.245-246

Abstract

While data speed provided by communication technologies and the data speed required for multimedia applications are competing with each other, the importance of OFDM technology, which has less intersymbol interference influences that occur in the channel, has been increasing. Deep fadings occurred as a result of multipath propagation cause some sub carriers that loses its information as well as fading in a non-recoverable degree, and consequently results in a poorer system performance. Since the training speed of LMS algorithm used in estimation of channel coefficients in time domain for OFDM based systems is not generally sufficient to achieve the desired performance levels in training time defined by standards, a fuzzy logic based algorithm is adapted to get a faster performance. An application of fuzzy logic based high performance LMS algorithm for OFDM systems is investigated in this paper. The aim of the proposed method is to produce the required fuzzy rule base and fuzzy subsets for fuzzy logic controller in order to speed up the LMS algorithm. The generations of fuzzy rules are based on the comparison of estimated signal by equalizer or channel output. The obtained results of simulations show that the LMS algorithm based on fuzzy logic (F-LMS) increases the performance in OFDM systems significantly and converges to the performance of the RLS algorithm.