Interference-Aware Scheduling and Routing in Unstructured Wireless Networks

A distributed admission-control paradigm for scheduling and routing in unstructured wireless networks which exploits each node´s estimates of channel gains and interference levels, and hence of signal-to-interference-plus noise ratios (SINRs), is presented. In contrast with the collision- avoidance approach, the present algorithms allow multiple concurrent transmissions and make transmission power assignments to admitted sessions based on maximizing the least of the resulting sustainable additional interference levels in order to maximze the possible number of future admissions and hence network throughput subject to an energy constraint. A nontrivual generalization to the case of multiple (elastic) traffic classes (where session priorities play an important role in admission decisions and power assignments) is developed. A routing mechanism that chooses the hop-sequence with the best sample statistics of the hop-SINRs is proposed and used in conjunction with the above scheduling algorithm to provide throughputs that are several-fold those of conventional schemes.