Performance Analysis of Millimeter Wave Ad Hoc Networks

Ad hoc networks provide a flexible, infrastructure-free means to communicate between soldiers in war zones, aid workers in disaster areas, or consumers in device-to-device (D2D) applications. Ad hoc networks, however, are still plagued by interference caused by uncoordinated transmissions which leads to poor scaling due to distributed coordination. Communication with millimeter-wave (mmWave) devices offers hope to 

ad hoc networks through higher bandwidth, reduced interference due to directional antennas, and weaker interference power due to building blockage. 

 

To understand the potential of mmWave for ad hoc networks, WNCG graduate students Andrew Thornburg and Tianyang Bai with WNCG professor Robert W. Heath Jr. analyzed various performance metrics of mmWave ad hoc networks. They used a stochastic geometry approach to characterize the one-way and two-way signal-to-interference ratio (SINR) distribution of a mmWave ad hoc network with directional antennas, random blockages, and ALOHA channel access. The effect of random receiver location is quantified which shows that random receiver distances do not alter the SINR distribution beyond knowledge of the mean receiver position. A method for computing the distribution of mmWave ad hoc interference-to-noise ratio which shows that mmWave ad hoc networks can still be interference limited. Several reasonable simplifications are used to derive the transmission capacity and area spectral efficiency. The performance of mmWave is then analyzed in terms of rate coverage. The results show that mmWave networks can support higher densities and larger spectral efficiencies, even in the presence of blockage, compared with lower frequency communication for certain link distances. Due to the increased bandwidth, the rate coverage of mmWave can be much greater than lower frequency devices.

 

Parts of the work was accepted to be presented IEEE International Conference on Communications 2015 (ICC) and IEEE International Conference on Acoustics, Speech, and Signal Processing 2015 (ICASSP). A full journal version has been submitted to IEEE Trans on Signal Processing. A pre-print is available on Arxiv.

 

The authors would like to acknowledge support from Army Research Labs under Grant No. W911NF-12-R-0011 and the National Science Foundation under Grant No. 1218338.