A Low-complexity Successive Detection Method for OFDM Systems over Doubly Selective Channels

Abstract

For OFDM systems, the time variation of a multipath channel results in inter-carrier interference (ICI). It leads to performance degradations. In reducing the problem, current successive detection methods cost very high computationalcomplexities. Among them, the minimum-mean squareerror successive detection (MMSE-SD) method has the best performance. This work proposes an improved data detection method with low complexity by integrating the techniques of Newton’s iterative matrix inversion method and the MMSE-SD method which considers the effects of signal-to-interference-plusnoise ratio (SINR). In order to efficiently integrate Newton’s and MMSE-SD method, we develop an effective scheme with low complexity for generating the initial values required by Newton’s method. Based on the new initial value scheme, we are able tosimplify the criterion of maximum SINR determination to an equivalent one with lower complexity. As a result, the proposed algorithm has a much lower complexity of O(N2) than O(N3) of the MMSE-SD algorithm, where N is the number of subcarriers. Moreover, simulation results in different channel conditions show that performances of the proposed MMSE-SD method are very close to the MMSE-SD method.