|
|
| Publication: S. Venkataraman, S.A. Mahmoud, and R.A. Goubran, "Performance and Complexity of Equalization and Diversity Combining Methods in Wireless Communications", Wireless 94 Conference, Calgary, Ata., July 1994, pp. 537-544. |
| Abstract: This work is aimed at studying the issues involved in systems that use bot h diversity combining and equalization. We study several filter structures for equalization and diversity combining which differ significantly in their computational complexity in terms of the impact of several variables on the bit error rate performances. The variables include training length, signal to noise ratios, length of equalizer, tracking properties at different values of Doppler frequency and presence of co-channel interference. the results indicate that filter structures differ significantly with respect to each of the above parameters and choice of the filter structure should be made with these effects taken into consideration. A number of new filter structures are introduced which offer the same performance as that of well known structures but with significant reduction in the computational complexity when used in a slow fading channel with long equalizers and sufficiently long training sequences. The RLS algorithm is used for adaptation. The new filter structure under study is shown to offer significant enhancements in performance at a lower complexity when short training sequences are used. The paper is divided into the following parts: Part II explains the different filter structures, Part III gives a summary of the simulation results, Part IV is a discussion the merits an d demerits of the different structures with respect to the above effects and finally Part V concludes the work of this paper. |