Research On Blind Equalization Algorithms Of MIMO Systems

In recent years, the demand for wireless data and multimedia service promotes the development and applications of high-speed wireless communication techniques. Multiple antennas adopted in Multiple Input and Multiple Output (MIMO) system can reduce correlation between multi-channels, which greatly increase system capacity. With increasing data speed, it appears time-variant multi-path fading in channel and decline performance in system. Blind equalization can effectively eliminate Intersymbol Interference (ISI). Blind equalization can balance channel characteristics by sending statistical channel information without additional training sequence. It has been widely used in digital communications, radar, seismic and image processing systems.In this thesis, the studies focus on constant modulus algorithm for MIMO system. Two proposed algorithms are dithered signed error constant modulus algorithm and modified normalized constant modulus algorithm, which are based on CMA algorithm:Firstly, combined with raditional dithered-sign error constant modulus algorithm for SISO systems, the modified dithered-sign error constant modulus algorithm is proposed. With introduction of cross-correlation items between outputs of different equalizers in the cost function, channel distortion is compensated, ISI and MUI are restrained. Computer simulation results illustrate that transmission signals is effectively recovered.Secondly, since direct application of normalized constant modulus algorithm in MIMO systems, one-to-many problem may arise. The promoted modified normalized constant modulus algorithm adds cross-correlation items between outputs of different equalizers. Computer simulation results illustrate the proposed algorithm can overcome the problem of one-to-many, be adaptive to channel changes and recover effective inputs.