A Low-Complexity Hardware-Friendly DFT-Based Channel Estimator for the LTE Uplink Channel

Abstract

In this work, a low-complexity hardware-friendly discrete Fourier transform (DFT)-based channel estimator with almost no inherent “edge effect” is designed for the long-term evolution (LTE) uplink channel. Specifically, we propose to perform the border symmetric extension (BSE) operation on the border subcarriers of the channel frequency response (CFR), such that the length of the extended CFR fulfills $2^{\lceil {{\text {log}}_2N}\rceil }$, where N denotes the non-radix-2 length of the original CFR, and $\lceil {x}\rceil$ stands for the integer ceiling function. Based on the proposed BSE operation, the discontinuities at the CFR’s border subcarriers are significantly lessened and a better power concentration of the transform-domain channel impulse response is realized. As a result, the inherent “edge effect” caused by the virtual subcarriers in LTE systems can be substantially reduced. A further advantage of the proposed method is that the cumbersome application specific integrated circuit-based implementation of 34 different non-radix-2 length DFT/IDFT operations can be accomplished in a single structure by their fast Fourier transform and inverse fast Fourier transform counterparts. Numerical results illustrate that the proposed DFT-based channel estimator with BSE achieves significant performance gains over the conventional counterpart, despite imposing a reduced computational complexity.