Abstract:

A fast reconstruction algorithm based on inner product optimization and sparsity updating constraint was proposed for OFDM system channel estimation when the number of channel paths was unknown.By constructing and updating the selection vector, the inner product operation was reduced by using the atoms corresponding to the non-zero index of the selection vector.The atoms were optimized based on compressed sampling and backtracking strategies, and the channel estimation was completed by matching pursuit.The sparsity update and the stop condition for the algorithm was achieved by the energy difference between the two adjacent channel estimation so as to ensure fast convergence of the algorithm.The simulation results show that the proposed algorithm has better channel estimation performance than the least square algorithm, minimum mean square error algorithm, sparsity adaptive matching pursuit algorithm and adaptive regularized compressed sampling matching pursuit algorithm, and consumes less channel estimation time than the two adaptive methods.

Key words: compressed sampling, inner product, backtracking strategy, sparsity adaptive, channel estimation