LS and MMSE Channel Estimation in Dual Pulse Shaping Systems

The dual pulse shaping (DPS) technology is an innovative broadband wireless communication solution that enables full-rate transmission with low-performance and low-cost data converters. This is achieved by splitting the data stream into two half-symbol rate streams, and shaping them by Nyquist pulses and complementary Nyquist pulses. In this paper, we investigate linear channel estimation methods, e.g., least squares (LS) and minimum mean square error (MMSE) methods for DPS systems. Simulation results show that the MMSE channel estimation outperforms the LS one in terms of mean square error of channel estimation and bit error rate (BER) performance. The MMSE one exhibits about 2 dB and 5 dB difference at the BER of 20 dB compared to the case that the channel state information is perfectly known at the receiver, when the length of training sequences is set to 32 and 64, respectively.