Graph Signal Processing-Based Interference Suppression for Multi-User Massive MIMO Systems

Abstract

The contribution of this paper is a new interference suppression structure in the multi-user, massive MIMO communication systems by utilization of Graph Signal Processing (GSP). Massive MIMO plays a central role in the next-generation wireless systems since it provides high spectral efficiency/spatial multiplexing. Nevertheless, there is a severe problem in dense deployments, which is the inter user interference combined with hardware impairment and sub-optimal channel state information (CSI). To overcome this, the spatial relationships between the users and antennas are modeled as a graph particularly with interference modelled as a graph signal under the proposed framework. The system is able to achieve extensive attenuation of interfering components as a result of the graph filters applied to the spectral domain, keeping the desirable signal attributes intact. The suggested solution is used and tested with simulation where it was applied in both uplink and downlink conditions, and these experiments are carried out with regular Rayleigh fading channels. Comparison is done with conventional techniques in zero-forcing (ZF) and minimum mean square error (MMSE) beamforming methods. The findings are an up to 35 percent gain in signal-to-interference-plus-noise ratio (SINR) and up to 25 percent gain in bit error rate (BER) at high user loads. Besides, GSP-based approach provides an even lower-computational complexity and this makes it applicable in dynamic and resource-limited settings. This paper establishes that GSP is a scalable, data-efficient, and robust interference mitigation framework of signal processing in massive MIMO systems, and carries considerable potential of being implemented in future wireless systems, 6G.