Adaptive Interference Avoidance for Dynamic Wireless Systems

Abstract

The paper presents an interference avoidance al- gorithm for codeword and power adaptation in the uplink of a Code Division Multiple Access (CDMA) system with variable number of active users. The algorithm tracks changes in the system configuration, and transitions incrementally from one optimal point corresponding to a given number of active users to a new optimal point corresponding to the new number of active users. The transition uses gradient-based adaptive interference avoidance in which users adapt their codewords to optimize their corresponding spectral efficiency subject to constraints on the signal-to-interference plus noise-ratio (SINR). The algorithm is illustrated with numerical examples obtained from simulations. I. INTRODUCTION Interference avoidance has emerged in the literature as a new technique by which transmitters in a wireless communication system are optimized in response to changing patterns of interference to better suit the environment in which they operate. Currently, interference avoidance algorithms are static in the number of users (1), and each time this changes the algorithms must be reiterated in order to determine the socially optimal solution for the new number of users. We note that other related algorithms for codeword adaptation for uplink CDMA systems (2)-(7) have the same characteristic, and are not adaptable to changing numbers of active users in the system. In order to overcome this limitation, recent research (8) proposes using Grassmannian signatures for use in dynamic systems with variable number of users in the system. These are designed to support a maximum number of active users in the system subject to a given interference level, and have the nice property that interference among users does not change when less users are active in the system. As noted in (8) the disadvantage associated with equiangular Grassmannian signatures is that they may not exist for any desired system configuration specified by given number of users and processing gain value. In this paper we present an alternative approach to dealing with variable number of active users in the uplink of a CDMA system, and propose an adaptive algorithm that moves the system incrementally from an optimal configuration with a given number of active users, to a new optimal configuration with a different number of active users. The transition between the two optimal configurations is based on an adaptive inter- ference avoidance procedure: when a change in the system status occurs this translates in a change of the SINR of active users which will employ a greedy gradient-based technique to optimize their corresponding spectral efficiency subject to