International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 06 | June 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1584 Interference Aware User Grouping Strategy for Downlink Massive MIMO Systems J.Roscia Jeya Shiney 1 , Dr.G.Indumathi 2 , Dr. A. Allwyn Clarence Asis 3 1 Research Scholar, Department of Electronics and Communication Engineering, Mepco Schlenk Engineering College, India 2 Professor, Department of Electronics and Communication Engineering, Mepco Schlenk Engineering College, India 3 Associate Professor, Department of Electrical and Electronics Engineering, Mepco Schlenk Engineering College, India ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract - In a Multiuser massive MIMO system with time division duplex scheme serving multiple users inter-user interference will degrade the system performance and so the users with correlated channel condition can be served in different time slots. In this paper a user grouping algorithm which separates the users based on the interference power is proposed. The grouping process first finds the difference in interference power of the users based on which the users are separated in to groups. Here the number of groups formed is not fixed but it depends on the user interference. The isolated users in the corresponding groups are then served in different time slots. The proposed scheme tends to serve all the users thereby ensuring fairness among the users. Simulation results show the effectiveness of the proposed scheme even in a highly correlated channel condition. Key Words: Massive MIMO, Channel state information, Time division duplex, signal to noise plus interference ratio, User grouping 1. INTRODUCTION Massive MIMO which is also known as very large MIMO or large scale antenna systems is one of the most promising technologies for the next generation cellular systems. It is envisioned that in massive MIMO, the BS is equipped with hundreds of antennas and serves tens of users at the same time. With excess spatial resources, massive MIMO can achieve all the merits of MIMO systems with a much greater scale. It is also assumed that the number of antennas M and the number of single-antenna users K go to infinity, while the ratio of K/M is fixed. In massive MIMO, time division duplex (TDD) scheme can be used where the uplink transmission and the downlink transmission use the same frequency band while they are separated by different time slots. In TDD scheme, there is a possibility to obtain channel reciprocity where the downlink channel matrix equals to the transpose of the uplink channel matrix. Then the downlink CSI can be achieved by uplink training in which the complexity of uplink training scales as the number of users. [1-5] Most of the existing MIMO schedulers estimate the orthogonality of user’s spatial channel and serve the users with near-orthogonal channels simultaneously [6]. Serving multiple users with correlated channels would bring severe degradation in system capacity and so there is a need for user grouping in a multiuser MIMO system where users with correlated channels should be served by different time slots. In [7] angle between users (ABU) is used to measure the correlation between different users is proposed. The two users are served within the same time slot only when the ABU is greater than a certain threshold value otherwise TDMA will be used instead. A hierarchical user grouping algorithm is suggested in [8] where merging of individual users based on certain criteria for user grouping is done. Eventually, all users can form one single group or it can terminate when the desired number of groups is reached. In [9-11] the grouping process relies on finding the correlation coefficients that are larger than a certain threshold value and isolating the corresponding users in separate groups who are then served in different scheduled time. The authors in [12-19] have studied the user grouping and scheduling problems based on a two-stage precoding framework for FDD massive MIMO systems where they have proposed weighted likelihood similarity measure, subspace projection based similarity measure, hierarchical clustering and K-medoids clustering for user grouping in order to achieve load balancing and user fairness for FDD massive MIMO systems. The importance of the proposed scheme is to take advantage of TDMA when the users are highly spatially correlated while maintaining the good performance provided by MU-MIMO when the inter-user correlation is low. Instead of randomly selecting users from the large pool of users the grouping process select users with less interference and thereby exploiting multi user diversity. As a result users in the same time slot will have low spatial correlation and therefore higher capacity can be achieved. Unlike the other existing works the proposed user grouping algorithm aim on finding the interference between users which is used as an criterion for user grouping process and is effective for improving the system performance even under highly correlated channel conditions.