Scaling Laws for Ergodic Spectral Efficiency in MIMO Ad-Hoc Networks

This paper considers a multi-antenna wireless net- work where the locations of transmitters are distributed as a Poisson point process, which is a natural model for ad-hoc and device-to-device networks. We show that, in such a network, the ergodic spectral efficiency scales linearly with respect to the network density under appropriate multiple antenna configurations and diversity assumptions. This scaling law is achieved by a simple zero-forcing decoder, which eliminates inter-stream interference using spatial multiplexing transmissions. We also show that when each receiver knows channel state information from some interferers, a higher scaling law holds for ergodic spectral efficiency than that without channel state information when using a partial zero-forcing method which eliminates dominant interference signals while boosting the desired signal power. Further, we show that spatial multiplexing transmission methods are essential for obtaining better scaling laws in certain regions of network parameters.