Zero Forcing Precoder Based Multiuser MIMO System Using K-Mean Grouping

Abstract

In today’s world, the demand for high speed data is increasing at fast pace as the number of wireless based applications are increasing day by day. Next generation wireless communication systems have to achieve very high throughput and provide security in order to fulfil the desire for faster and reliable communications. But present generation systems cannot achieve higher data rates because of scarcity in band limited spectrum for wireless communication. So the technique called Multiple Input Multiple Output (MIMO) comes for the rescue from such situation. MIMO is the most promising technique for providing outstanding reliability and exceptionally high data rates. In this technology multiple antennas are used at the transmitter as well as receiver side which helps to exploit the various properties of the systems. Further, this system has been extended to the multiuser scenario. As in today’s world, there has been very large number of users for the wireless communication system. But, there is large number of problems like reduced capacity, interference, security etc while accommodating all users in the system. So, there needs to be system which can accommodate large number of end users by providing them the reliability and the high data rates along with reduction of interference. This can be done by using large number of antennas at the base station and by using some user grouping algorithms, which group the users based on the various properties. Along with that, the precoding technique called Zero Forcing can be implemented to further improve the system by reducing the interference. In this dissertation work multiuser MIMO system using Zero Forcing precoding along with the K-mean grouping algorithm has been used to improve the sum rate of the system for the downlink case. The total of 100 antennas at base station and 100 users having single antennas has been considered. For SNR of 15 dB, the sum rate for proposed precoder for (g = 2) is 11 bits/sec/Hz which is almost 1.7 times more than the conventional zero forcing algorithm. Further in this dissertation work, the optimization of the user grouping has been done. The sum rate keeps on increasing while the user groups are increased from two to four. But suddenly when the number of groups created reaches five, the performance of the system decreases.