Compressed Sensing OFDM Time Delay Estimation Based on Atomic Evolution and Elimination

Abstract

Aiming at the problem of time delay estimation for orthogonal frequency division multiplexing (OFDM) signals with small samples, a new compressed sensing time delay estimation method based on atomic evolution and elimination is proposed. This method introduces the idea of evolution and elimination from Darwin's evolutionism, abandons the fixed and invariable way of grid points in the traditional algorithm, takes grid points as adjustable parameters, and updates grid points in each iteration process to promote the evolution of atoms. In order to deal with the off-grid effect, the exponential function is directly modified instead of linear approximation, which effectively eliminates the modeling error caused by the original off-grid algorithm. In the process of calculation, the atomic elimination mechanism is adopted, which reduces the amount of calculation and improves the speed of it. The simulation results show that the proposed algorithm can significantly improve the performance of time delay estimation, especially when using coarse grids. Based on the universal software radio peripheral (USRP), the validity of the proposed algorithm is verified by the actual signal.