Hybrid Iterative Updates Detection Scheme for Uplink Dynamic Multiuser Massive MIMO Systems

A new signal detection scheme based on joint Chebyshev polynomials acceleration and semi-iterative plus symmetric successive over relaxation (SSOR+CP) method is proposed to fast update the linear minimum mean square error (MMSE) detection when a user is added to or removed from the massive multipleinput multiple- output (MIMO) systems. Chebyshev polynomials (CP) acceleration and semi-iterative method not only is employed to construct a secondary iteration for the SSOR method to significantly accelerate the convergence rate, but also has inexpensive cost. Furthermore, by utilizing Schur complements and the block matrix inverse lemma, we can fast update a matrix inverse that further reduce the complexity by an order of magnitude in real massive MIMO systems with a dynamic set of users. Through numerical simulations, it is observed that the proposed hybrid iterative updates method, with only a few operations, is almost able to perform the optimal detection performance in contrast with recently proposed detection algorithms, even when the number of user equipments (UEs) significantly increased. Meanwhile, our hybrid iterative updates method can provide more robust signal detection in high-spatial correlation of massive MIMO channels.