On the Tradeoff Between Multiuser Diversity and Training Overhead in Multiple Access Channels

TL;DR

This paper analyzes the balance between training overhead and multiuser diversity in a single antenna multiple access channel, deriving optimal parameters for system performance as block length increases.

Contribution

It provides closed-form solutions for optimal user selection and training parameters, and explores their asymptotic behavior as block length grows large.

Findings

Optimal number of users K and training length L derived

Expressions for system performance as L increases

Tradeoff analysis between training overhead and multiuser diversity

Abstract

We consider a single antenna narrowband multiple access channel in which users send training sequences to the base station and scheduling is performed based on minimum mean square error (MMSE) channel estimates. In such a system, there is an inherent tradeoff between training overhead and the amount of multiuser diversity achieved. We analyze a block fading channel with independent Rayleigh distributed channel gains, where the parameters to be optimized are the number of users considered for transmission in each block and the corresponding time and power spent on training by each user. We derive closed form expressions for the optimal parameters in terms K and L, where K is the number of users considered for transmission in each block and L is the block length in symbols. Considering the behavior of the system as L grows large, we optimize K with respect to an approximate expression for the achievable rate, and obtain second order expressions for the resulting parameters in terms of L.