Capacity of Ultra Wide Band Wireless Ad Hoc Networks

Throughput capacity is a critical parameter for the design and evaluation of ad-hoc wireless networks. Assuming a fixed per node transmission capability of T bits per second at a fixed range, it has been shown [1] that the information theoretic uniform throughput capacity per node r(n) is Theta (T/radicn log n), a decreasing function of node density n. However we consider an alternate communication model, with each node constrained to a maximum transmit power P₀ and capable of utilizing W Hz of bandwidth. Under the limiting case W rarr infin, such as in ultra wide band (UWB) networks, we show that the uniform throughput per node is O ((n log n)^alpha-1/2 (upper bound) and Omega(n^(alpha-1)/2(log n^(alpha+1)/2) (achievable lower bound). Thus we demonstrate that the throughput increases with node density n, in contrast to previously published results. The capacity problem is also considered from an optimization theoretic perspective. The novel information theoretic capacity bound is compared to the network optimization results, demonstrating its practical applicability to UWB networks. The dramatic effect of the UWB physical layer on the capacity justifies the promise of UWB as a physical layer technology for ad-hoc networks.