OPTIMUM COMBINING WITH THE LMS ALGORITHM FOR INDOOR MULTIPATH CHANNELS
Date
1994-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Degree Level
Doctoral
Abstract
The capacity of wireless communication systems must be increased to accommodate ever increasing demand. The capacity issue is especially acute for indoor systems where high user densities are possible. Optimum combining—which exploits multipath fading by weighting and summing signals received via multiple antennas to suppress interfering components—can be used to increase system capacity.
Winters has examined the performance of optimum combining under ideal conditions—perfect power control (such that the power received from all co-users is equal) and ideal optimum combiner weights. In a real system, perfect power control is difficult to achieve, and the optimum combiner weights must be estimated using techniques such as the LMS algorithm. This thesis extends the work of Winters by examining the performance of optimum combining under more realistic conditions— imperfect power control and LMS-generated weights. Rayleigh fading and perfect desired signal - reference signal correlation are assumed.
In this thesis, a method to adaptively control the LMS algorithm step size, and a method to improve system performance under high power interference, are developed. The average output SNR, output ISR, and BER of systems employing optimum combining with LMS-generated weights are compared to ideal weight results for many input SNR, input ISR, co-user number and antenna number combinations. As well, BER distributions and failure (periods of over-threshold BER) duration distributions are examined. Finally, the capacity increases achieved by frequency sharing within cells and frequency sharing between adjacent cells of a cellular system are compared to the capacity increases achieved by cell size reduction. It is shown that optimum combining with LMS-generated weights can be an effective method to increase the user capacity of wireless systems.
Description
Keywords
Citation
Degree
Doctor of Philosophy (Ph.D.)
Department
Electrical Engineering