Optimized Iterative (Turbo) Reception for QAM OFDM with CFO over Unknown Double-Selective Channels
Paper in proceedings, 2011
A novel iterative (turbo) receiver is introduced, suitable for orthogonal frequency division multiplexing (OFDM) employing quadrature amplitude modulation (QAM) and receiver diversity. The system operates over a double-selective channel and includes a carrier frequency offset (CFO). We propose a maximum a posteriori probability expectation-maximization (MAP-EM) receiver with a different EM parameter division than standard methods. In such standard MAP-EM receivers, the E-step parameters correspond to the channel, while the M-step parameters correspond to the CFO and data symbols. This standard receiver parameter division results into a highly complex receiver for QAM, due to the large modulated symbol alphabet size, and the non-constant constellation symbol amplitude. In this paper, a new receiver framework introduces a different parameter division that leads to reduced complexity turbo receivers for QAM signaling, while still achieving close to optimal system performance. The new approach adapts the sum-product algorithm (SPA) parameter framework to the MAP-EM receiver. Thus, in the new receiver framework, the E-step parameters are data symbols, while the M-step parameters are the channel and the CFO. We evaluate the performance of the proposed receiver with and without automatic repeat request (ARQ), where in the former case packet combining applies to further improve performance.