Sparse channel estimation using overcomplete dictionaries in OFDM systems

Abstract

With the in-depth study of the wireless channel, more and more experimental evidence show that many wireless channels are sparse in the conditions of large bandwidth and long signaling durations. Thus, Compressed Sensing theory applied for sparse channel estimation can reduce the number of pilots, so as to increase spectral efficiency. However, the non-integer times of sampling period about the time-delay or Doppler frequency shift will lead to the energy leakage, and reduce the time delay-Doppler sparsity of the equivalent channel, thus affect the accuracy of channel estimation. In this paper, we utilize over-complete dictionaries based on super resolution to enhance the sparsity of the equivalent channel. Simulation results demonstrate that the overcomplete dictionary representation of the double-selective channel is much sparser than the classical delay-Doppler representation. The method proposed in this paper can effectively improve the performance of sparse reconstruction algorithms, and then obtain the better precision of channel estimation. © 2014 Springer International Publishing Switzerland.