Comparisons of Short-Prefix Based Channel Estimation in Single-Carrier Communication Systems

In this paper, we compare the performance between channel estimation based on compressed sensing (CS) and time-domain least square (LS) for single-carrier (SC) communication system. Unlike the conventional channel estimation techniques such as frequency domain LS which is used in the condition that the length of pilot sequence is equal to data sequence, the estimation scheme based on CS requires smaller length of pilot sequence. In this paper, the corresponding system structure is presented. Zadoff-Chu sequence is used to generate the pilot sequence, which is shown to perform better in forming measurement matrix of CS than pseudo random sequence. Simulation results demonstrate that channel estimation based on CS achieves a better bit error rate (BER) performance than time domain LS with a smaller pilot sequence and thus raising data rate of the SC communication system.