Role of Large Scale Channel Information on Predictive Resource Allocation

TL;DR

This paper demonstrates that future large-scale channel information can nearly achieve the performance gains of perfect channel knowledge in predictive resource allocation, simplifying practical implementation.

Contribution

It analytically shows that large-scale channel information suffices for near-optimal resource allocation, reducing the need for complex small-scale channel estimation.

Findings

Large-scale channel info captures most performance gains.

Simulation confirms the analytical results.

Estimation errors impact energy savings.

Abstract

When the future achievable rate is perfectly known, predictive resource allocation can provide high performance gain over traditional resource allocation for the traffic without stringent delay requirement. However, future channel information is hard to obtain in wireless channels, especially the small-scale fading gains. In this paper, we analytically demonstrate that the future large-scale channel information can capture almost all the performance gain from knowing the future channel by taking an energy-saving resource allocation as an example. This result is important for practical systems, since large-scale channel gains can be easily estimated from the predicted trajectory of mobile users and radio map. Simulation results validate our analysis and illustrate the impact of the estimation errors of large-scale channel gains on energy saving.