Improved noise power and SNR estimation in severe fading channels for OFDM systems

Abstract

The purpose of this study is to estimate signal to noise power ratio (SNR) in a se-vere fading channels. Signal power and SNR estimation is hard to estimate in a severe fading channels because channel variation occurs more extremely. In this situation, noise power estimates are generally higher than the actual values because of channel estimation errors, channel variations, and multipath delays. Moreover, current communi-cation systems reduce pilot density to increase data rate. As a result, accurate SNR estimation is becoming more difficult. This paper analyses SNR estimation techniques and proposes improved signal and SNR estimation schemes in severe fading channels to reduce estimation error. We analyze representative SNR estimators. The proposed signal power and SNR estimators use a chan-nel estimation filter (CEF) and differential scheme. The simulation results show that the differential scheme SNR estimator provides better performance in severe fading channel. The proposed techniques can reduce error because they select a better condi-tion channel domain. The proposed SNR estimation technique uses OFDM characteristics in multi-dimensional domain. Channel status is different along the time and frequency domain. Choosing a better domain can reduce error in SNR estimation. The proposed SNR estimator uses SNR information of each domain to choose the proper domain. The simulation result shows that the proposed estimation technique can reduce estimation error up to 5dB in a frequency selective or fast fading channel.The study result can be applied to contemporary communication system channel esti-mation such as LTE and WiMax, which provide high data rates with low pilot density with high interference channel or fast fading channel. ⓒ 2013 DGIST