On Noise Correlation in Interference Alignment for MIMO Power Line Communications

The power line is inherently a broadcast medium when used for data transmission, since the electricity grid has no well defined boundaries for communication signals. This broadcast nature of the channel leads to mutual interference of communication signals emitted in relative proximity. To overcome the rate-limiting effect of this interference, recently the concept of interference alignment (IA) has been applied to power line communications (PLC). In particular, the benefits of IA have been shown for transmission in multi-input multi-output (MIMO) PLC systems. However, the results were obtained under the assumption of independent noise at the different ports of the PLC receiver. Unlike for most wireless communications scenarios for which IA has usually been studied, this assumption is often not valid for MIMO PLC. In particular, due to the dominance of noise signals induced and broadcast from grid-connected devices as well as the signal cross-coupling in PLC networks, noise is often found to be spatially correlated. Therefore, in this paper we explore the performance of IA for MIMO PLC in the presence of correlated noise obtained from indoor measurements. To this end, we adopt the maximum signal- to-interference-plus-noise ratio IA algorithm, whose formulation includes the noise statistics. Numerical results quantify the effect of noise correlation on the achievable sum-rate for MIMO PLC with IA.