Journal of Communications Vol. 15, No. 2, February 2020 ©2020 Journal of Communications 214 Performance of Leakage Based Precoding Scheme for Minimizing Interference Subuh Pramono 1 , Eddy Triyono 2 , and Budi Basuki Subagio 2 1 Department of Electrical Engineering, Universitas Sebelas Maret, Surakarta 57126, Indonesia 2 Department of Electrical Engineering, Politeknik Negeri Semarang, Semarang 50275, Indonesia Email: subuhpramono@staff.uns.ac.id; eddytriyono@gmail.com; budi.basuki2010@gmail.com Abstract—In this paper, we consider the performance of leakage based linear precoding scheme in downlink multiuser MIMO. This work focuses on minimizing the interference to improve the performance system. The proposed leakage based linear precoding will be investigated in term of bit error rate (BER), sum rate capacity and outage. Simulation results show that the leakage based linear precoding scheme (MMSE SLNR and BD SLNR) achieved better BER, higher sum rate capacity, and lower outage than non-leakage precoding scheme (BD). At specific 10 dB SNR, there are 2.3 10 -3 and 1.8 10 -2 of target BER for MMSE SLNR and BD SLNR, respectively. Furthermore, the MMSE SLNR yields a lower outage than BD SLNR, at specific 2 dB SINR, the probability of an outage is 0.65 for MMSE SLNR and 0.86 for BD SLNR. Overall, the MMSE has better performance than BD SLNR because it has capabilities to eliminate the IUI and ISI. Index Terms—Leakage based precoding, MMSE SLNR, BD SLNR, BER-outage, sum-rate I. INTRODUCTION Nowadays, mobile communication technology is moving into the fifth generation which can provide a gigabit data rate communication. One of the most important subsystems of the fifth mobile communication technology is the antenna subsystem, to support a gigabit data rate service, the fifth mobile communication technology applies a MIMO antenna system (multiple input multiple output). The MIMO antenna system has attracted attention due to its capability of improving link reliability and increasing data rate without extra power transmit and bandwidth. The number of user in mobile communication service is now increasing exponentially, a billion devices are connected through the mobile communication network. Multiple antennas are mounted at a single transmitter to serve multiple users (receivers) simultaneously. Early MIMO was addressed on SU MIMO (single-user MIMO) where the multiple spatial channels were used for a single user. Now, research focused on MU MIMO (multi-user MIMO) where multiple users were served simultaneously in the same frequency and time slot, it is shown in Fig. 1. MU MIMO increased significantly the spectral efficiency. Recently, the number antenna at the transmitter transforms into a Manuscript received August 2, 2019; revised January 9, 2019. doi:10.12720/jcm.15.2.214-220 large number antenna, it is usually called a massive MIMO. In term of multi-user MIMO that means the massive MIMO serves multiple users with a single or multiple receiver antenna. In multi-user MIMO, the transmitters serve several users in the same time slot and frequency that means the presence of multiple users cause interference in downlink multi-user MIMO system, the channel of downlink multi-user MIMO is depicted in Fig. 2. Several types of interference that arise in downlink multi-user MIMO: inter-user interference (ISI), inter-cell interference (ICI), and co-channel interference (CCI). The existence of noise and interference influences the performance of downlink massive MIMO. Interference will degrade the capacity and bit error rate of the downlink multi-user MIMO system. The elimination or suppression of the interference and noise can be accommodated by providing decoding at the receiver and precoding scheme at the transmitter. However, decoding at the receiver in the downlink MU MIMO system caused increasing of computational complexity. Generally, precoding schemes at transmitter were categorized into two types: nonlinear and linear precoding system. DPC (dirty paper coding) and Tomlison–Harashima precoding are categorized into nonlinear precoding [1]-[11]. The nonlinear precoding yielded high sum rate capacity with higher computational complexity. The linear precoding, such as block diagonalization (BD), minimum mean square error (MMSE), and zero-forcing (ZF) had a moderate sum rate capacity with lower computational complexity. In addition, several criterions of precoding schemes have been developed and proposed to handle the interference and noise, i.e., interference ignorant, orthogonal filtering, orthogonal precoding, leakage based precoding, minimal interference precoding, maximum interference precoding algorithm, and an iterative scheme [1], [9]. In previous work [2], zero-forcing (ZF) precoding scheme eliminated the co-channel interference but neglected the noise. The zero forcing brings a constraint in term of the number of transmit antennas, particularly the number of antennas in the transmitter must be larger than the number of antennas in the receiver. The work [3] investigated the performance of block diagonalization which effectively suppressed the intra-cell interference. In addition, the authors in [4] investigated the performance of whitening filter in BD precoding scheme