A Novel SNR per Subcarrier Estimation Scheme for OFDM Systems in Frequency Selective Channels

Signal-to-noise ratio (SNR) is an important parameter for adaptive orthogonal frequency division multiplexing (OFDM) systems as it serves as a standard measure of signal quality. In this work, we present a novel SNR per subcarrier estimation scheme for wireless OFDM systems in frequency selective channels. The novel scheme can operate both in time and frequency, requiring only one previously known OFDM symbol. Firstly the noise variance per subcarrier estimate is obtained using the estimated channel coefficients and a novel analysis of its estimation error is provided. The noise variance per subcarrier estimate is used to obtain the SNR per subcarrier estimate. The estimation scheme takes under consideration the frequency selectivity of the channel and the variation of the noise statistics across the OFDM subcarrier index. The estimator's performance, in means of normalized mean square error (NMSE), is evaluated through simulations on a variety of channel models and is compared with another estimation scheme. The NMSE of the proposed scheme demonstrates its ability to identify the local statistics of the noise efficiently resulting into a reliable estimator in contrast with other similar approaches.