Smoothing techniques for decision-directed MIMO OFDM channel estimation

摘要

With the purpose of supplying the demand of faster and more reliablecommunication, multiple-input multiple-output (MIMO) systems in conjunction withOrthogonal Frequency Division Multiplexing (OFDM) are subject of extensiveresearch. Successful Decoding requires an accurate channel estimate at thereceiver, which is gained either by evaluation of reference symbols whichrequires designated resources in the transmit signal or decision-directedapproaches. The latter offers a convenient way to maximize bandwidth efficiency,but it suffers from error propagation due to the dependency between the decodingof the current data symbol and the calculation of the next channel estimate. Inour contribution we consider linear smoothing techniques to mitigate errorpropagation by the introduction of backward dependencies in the decision-basedchannel estimation. Designed as a post-processing step, frame repeat requestscan be lowered by applying this technique if the data is insensitive to latency.The problem of high memory requirements of FIR smoothing in the context ofMIMO-OFDM is addressed with an recursive approach that acquires minimalresources with virtual no performance loss. Channel estimate normalized meansquare error and bit error rate (BER) performance evaluations are presented. Forreference, a median filtering technique is presented that operates on the MIMOtime-frequency grids of channel coefficients to reduce the peak-like outliersproduced by wrong decisions due to unsuccessful decoding. Performance in termsof Bit Error Rate is compared to the proposed smoothing techniques.