Anotace:
In this paper, we have proposed a new paradigm for user scheduling in large-scale multiple-input multiple-output (MIMO) time division duplexing (TDD) system. In this paper, we have selected the users from different groups with semi-orthogonal (SO) and random criterion. We separate the users in different groups with the K-means clustering algorithm which assigns the users into different groups. After user groups are so determined, we use two new user selection paradigms where users are selected in two methods- firstly we select the users in intra-group those are SO with each other along with it also SO with other groups’ users. Secondly, users are selected from inter-group those are SO with each other also SO with other group users. In both the selection schemes antennas are scheduled based on the maximum gain of the channel. In the results, it is noticed that in intra-group with semi-orthogonal user selection (SUS) and antenna selection (AS) using the zero-forcing (ZF) precoding shown the highest systems rate. We also evaluated the computation cost of our modified proposed algorithm which is exposed in table-1. We explored the efficiency of the proposed schemes through MATLAB simulations.