On the Design and Tradeoff of Wireless Downlink Space Time Scheduler on Network Capacity and Coverage

It is well-known that wireless scheduling algorithm could exploit multi-user diversity to enhance the network capacity of wireless systems. However, the advantage of scheduling with respect to network coverage is a relatively unexplored topic. We consider a wireless system with an access point or base station equipped with transmit antennas as well as mobiles with single receive antenna. With multiple transmit antenna, there are additional degrees of freedom which could deliver either spatial multiplexing gain and/or spatial diversity gain. The spatial multiplexing allows simultaneous transmission to multiple clients at the same time and frequency and therefore contributes to network capacity gain. On the other hand, the spatial diversity offers robustness to channel fading and contributes to the network coverage. With cross layer scheduling, there is also multi-user selection diversity which contributes to both network capacity and coverage. In this paper, we propose a common analytical design framework for both capacity-optimized scheduler and coverage-optimized scheduler design. We show that both problems could be cast into a common optimization problem with an adjustable parameter . We found that tradeoff in the space time scheduler is required only for users close to the base station. For example, the space time scheduler should exploit spatial multiplexing ( ) and spatial diversity ( ) to optimize the network capacity and network coverage respectively. On the other hand, for users that are far from the base station, the space time scheduler should always exploit spatial diversity ( ) nomatter it is target for network coverage or network capacity. Due to huge complexity involved for the optimal solution, we propose two heuristic space-time schedulers, namely the greedy algorithm and the genetic algorithm, that allows flexible tradeoff between network capacity and network coverage. We show that the greedy algorithm, which has been widely used in today’s wireless data systems (3G1X, HDR, UMTS), offers the optimal coverage performance ( ). When , there is a large performance gap between the greedy algorithm and the optimal result. On the other hand, the genetic scheduler could fill in most of the performance gap at reasonable complexity.