An expectation-maximization solution to interpolated OFDM systems

We propose a hybrid channel estimation process for orthogonal frequency division multiplexing (OFDM) systems. We show that the OFDM channel estimation process can be substantially improved by using,a blind estimation module, namely the expectation-maximization (EM) algorithm, in addition to a conventional non-blind estimation module such as pilot symbol-aided channel estimation (PSAC). We provide a mathematical framework for the EM module and present computer simulations on its performance for an OFDM system operating in a frequency dispersive mobile channel. An additional benefit is revealed by our simulations which show that the proposed method is particularly powerful when the pilot insertion rate, i.e. overhead, is low.