On the Selection of Significant Channel Parameters for OFDM Systems: A Hypotheses Testing Based Approach

This paper considers multiple hypothesis techniques for the identification of significant channel parameters in OFDM systems. Unlike several OFDM channel estimation methods, where the channel response is estimated based on a pre-defined or an estimated length. In this work, the channel length is assumed unknown and only significant channel parameters will be identified. First, using a training based scenario, a model is proposed for the channel response. Second the model parameters are estimated using least squares (LS). Based on those estimates, multiple hypothesis tests based on F-statistics are constructed to classify each significant parameter. Simulation results show that the method is capable of identifying significant channel parameters with high probability under various SNR. In addition a receiver based on those parameters have a better performance than one which is based on the full model