Weighted Sum-Rate Optimization for Rate and Reach Enhancement in G.fast

Abstract

Digital subscriber line technologies such as G.fast face strong crosstalk between different twisted pair copper wires on higher frequencies. With crosstalk cancellation and interference management techniques, the performance is optimized with respect to high crosstalk. Maximum sum-rate is a widely used objective for groups of multiple subscribers, but does not reflect the subscriber demands and may lead to a high spread of data rates. This paper presents an approach to utilize weighted sumrate optimization to improve data rates for a subset of subscribers with higher demand, resulting in a longer reach for a certain data rate while using free resources from other subscribers. Results for linear and non-linear zero-forcing precoding as well as a gradient approach for linear MMSE precoding are investigated.