Joint Channel Tracking and Decoding for BICM–OFDM Systems Using Consistency Tests and Adaptive Detection Selection

This paper presents an iterative channel tracking and symbol decoding for combined bit-interleaved coded modulation (BICM) and orthogonal frequency-division multiplexing (OFDM) systems over time-varying channels. The proposed scheme alternates between a maximum
a posteriori
(MAP) decoder and a Kalman filter (KF) to track the channel taps. A key assumption in formulating the model is that the state-space model is perfectly specified. However, decision-directed detection methods make use of the decoded symbols to form the state-space model and may be inaccurate due to misdecoded symbols. The KF lacks robustness to outliers, and therefore, filter divergence may occur and give rise to error propagation, bringing about the well-known ldquoerror floorrdquo phenomenon. A novel approach to minimize the state-space mismatch is presented in this paper. In particular, we introduce a new component in the iterative receiver, which evaluates the filter's performance. By performing a statistical comparison of the innovation process against the theoretical predicted values, the new component adapts the number of decoded symbols used for channel tracking according to their reliability, thus reducing the mismodeling effect. We give an analysis of the effect of misdecoded symbols on the KF and how misdecoded symbols affect the channel frequency response estimation error. It is shown that this method can significantly reduce the error propagation effect, leading to a reduced bit error rate (BER).