Channel Estimation Algorithms for Hybrid Antenna Arrays
At the millimeter wave and higher frequency bands the radio channel can often be expressed as a linear combination of a small number of scattering clusters. Hence, the number of angles of arrivals with significant components is limited. Due to severe path losses, the receiver must be equipped with an antenna array capable of forming narrow beams. The channel estimation with narrow beams is challenging. Algorithms developed for sparse estimation problems can be utilized to overcome the problem. In this paper, the performance and computational complexity of channel estimation methods for millimeter and terahertz frequency bands are compared. The methods considered are based on Bayesian learning with the relevance vector machine, orthogonal matching pursuit and the least absolute shrinkage and selection operator optimization. The conventional least squares channel estimator is used as a reference method. The complexity of the least squares estimator is found to be the smallest. The estimation accuracy of the Bayesian learning based estimator is the best but with increased computational complexity.