Fast-Decoding Channel Estimation Technique for Downlink Control Channel in LTE-MTC Systems

The future Machine Type Communication (MTC) forces new requirements on cellular networks such as low-power, low-cost, narrow band, and extended coverage. To support these features, the 3GPP LTE-A introduced a new user equipment (UE) category, namely CAT-M1 raising new challenges for the system designers. For instance, the device should be able to setup a call at a signal-to-noise ratio (SNR) of -15 dB in the extended coverage mode with only one receive antenna and almost no frequency diversity. A new Physical Downlink Control Channel (PDCCH), namely MPDCCH, is introduced to satisfy the new system requirements. The new design relies on signal repetition to enhance the detection performance under very low SNRs. While the performance of this control channel will mainly depend on the channel estimation techniques used for coherent equalization during transmission, the choice of such technique will affect both the complexity and the performance of the UE. In this paper, we propose a channel estimation for MPDCCH that depends on linear processing of the pilot symbols and first-order polynomial interpolations. In order to enhance the performance of this technique we use a de-noising step using decision-feedbacks to help the UE reduce the power consumption of the reception and detection of the MPDCCH. The proposed scheme is able to use less than 30% of the supported repetitions in enhanced coverage modes to achieve the required detection error rate at moderate and low SNRs.