OMP-based Channel Estimation with Dynamic Grid for mmWave MIMO positioning systems

Location information would become more critical in next-generation networks. In this letter, we introduce a novel estimation method for location-related parameters, i.e., Time-of-Arrival (TOA), Angle-of-Departure (AOD), Angle-of-Arrival (AOA), and channel amplitude in mmWave MIMO systems. Because of the sparse nature of the mmWave channel, we first coarsely estimate all parameters using compressive sensing (CS). To mitigate the error caused by basis mismatch, we coordinately apply Golden-Section search for AOD and AOA, followed by instant estimates of channel amplitude and TOA using Least Square (LS) with only two up-to-date atoms of AOA and AOD, in descent manner. Finally, channel amplitude and TOA is re-estimated by CS with mismatch-free basis, which is obtained from refined atoms. Simulations show that the proposed method can approach the exact value of all parameters with faster convergence and lower algorithmic complexity than full-dimensional search based on Newton-Type algorithms.