Uppsala universitet

Reuse Partitioning and System Capacity
in the Adaptive OFDM Downlink
of the Wireless IP Project Target System.

Mikael Sternad, Uppsala University

Technical Report, Signals and Systems, Dept. of Engineering Sciences,
Uppsala University, Vers. 3.1, July 2003.

How can we, in our proposed adaptive OFDM downlink, simultaneously obtain good coverage within cells and a high average spectral efficiency? The present report outlines a solution to this problem. It also gives a first estimate of the attainable system capacity for our target radio interface.

The proposed method uses coordinated scheduling among sectors on the same site (radio access point) to suppress interference. Furthermore, one frequency band (with reuse 1) transmits to near users, while another band, with orthogonal resource sharing among clusters of 3 base stations, transmits to far-off users in the sectors.

The key idea is to apply a frequency reuse factor >1 only where it is needed most, in the outer part of the sector, where the signal from the base station is weak. It becomes possible to attain coverage over the whole sector, with an effective resource reuse 2 and an average capacity of 1.24 bit/s/Hz/sector for one Rayleigh fading user, including overhead, without assuming multiuser diversity or multiple receiver antennas. At least 16 QAM can be used in approximately 33 % of the sector area.

The resulting signal-to-interference ratio and spectral efficiency has been evaluated as a function of the position within the sector, by summing over all relevant interferers. The traffic density is assumed constant over the area, and hexagonal coverage areas are assumed for the base stations. Both triangular 60 degree sectors and diamond-shaped (30 degree rotated) sectors are considered.

The antenna pattern and an exponential path loss is taken into account. The spectral efficiency is first calculated for one user per sector (no multiuser diversity) with adaptive modulation. Results are presented for static channels and for flat Rayleigh-fading channels, both with path-loss. Situations with K active users within the sectors who each have L antennas and use maximum ratio combining are then investigated, for Rayleigh fading channels with path loss.

The presented estimates neglect shadow fading and noise. They therefore represent only a first approximation of the true, much more complicated, situation, which we will investigate in our system simulator under construction. Various ways of improving this basic solution are discussed briefly, including the use of coordinated scheduling over several base stations, slow power control, the use of interference rejection by multiple antennas in user equipment, and transmission by Trellis-coded adaptive modulation instead of uncoded adaptive modulation. The methods and performance measures presented here can serve as benchmarks for such more elaborate solutions.

Related publications:
An overview of the Wireless IP Project
The uplink of this system proposal, presented at WWRF March 2002, Phoenix, Arizona.
Licenciate Thesis by Nilo C. Ericsson (June 2001), on predictive scheduling combined with adaptive modulation.
VTC01spring-paper on channel prediction over 1-10 ms horizons.
Optimizing adaptive modulation, by taking the power prediction error into account.


| The Wireless IP Project |
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors. All persons copying this information are expected to adhere to the terms and constraints invoked by each authors copyright. This work may not be reposted without the explicit permission of the copyright holders.