Multiple-training bi-directional adaptive equalizers for TDMA digital cellular systems

Abstract

In this paper, two popular adaptive equalization methods, fractionally spaced decision feedback equalization (FSDFE) and maximum-likelihood sequence estimation (MLSE), are investigated for the design of digital mobile receivers for the IS-54 specifications of the NADC* system. A hi-directional equalization technique is incorporated and a multiple training LMS (MT-LMS) algorithm is used as the adaptive algorithm for both equalization methods. The results show that both MT-MLS and bi-directional techniques are effective in improving the receiver performance. However, the MT-LMS algorithm is more useful for MLSE than FSDFE whereas the bi-directional equalization technique improves FSDFE much more than MLSE.