Study On Blind Equalization Of MIMO Communication System

In the future, mobile communication and Wireless application require next generation wireless communication system, a more reliable connection, the higher frequency spectrum efficiency, a better mobility as well as the low emissive power, in addition must overcome various disturbances, but which the existing single input single output system (SISO) are unable to satisfy. The multi-input multi-output (MIMO) technology, which increases communications system's capacity greatly by using many antennas, is a key technology in new generation mobile communication system. At present various countries' scholar, conducts the extensive researches on various aspects' theory performance and the algorithm of MIMO. Because the wireless mobile communication MIMO system is one that suffers from the frequency decline influence, its receiving end's performance will worse seriously as the inter symbol interference caused by multi-channel transmission. Blind equalization and Blind estimate play important role to improve the system performance, especially in the system unable to obtain enough long or lacks the training sequence.Second-order blind equalizations, compared with the higher order blind algorithms, have the merit of low algorithm complexity. So this article only researches the second-order MIMO blind equalization algorithm. Based on the frequency selectivity MIMO channel model, has studied several main second-order MIMO blind equalization algorithms, including SSA (Signal Subspace Algorithm), OPDA (Outer Product Decomposition Algorithm), LPA (Linear Prediction Algorithm) and TXK the algorithm, and has given their performance simulation results. The result indicated that the four methods can recognize the channels, and SSA is best performance but the highest algorithm order of complexity. Under the low signal-to-noise ratio (<10dB), is SSA, LPA, TXK, OPDA in turn. Under the high signal-to-noise ratio (>10dB), is SSA, LPA, TXK, OPDA in turn. The existed MIMO blind equalization algorithms need to know the channel orders, this condition has restricted the algorithm flexibility. The goal of the paper is to propose one kind method that may carry on the channel equalization algorithm in the channel order unknown condition. In view of the outer product characteristic of OPDA, added zero before each data frame, this may make use of the main axle element of the correlation matrix of the received data to carry on the estimation of the channel order. This paper, combined the characteristic of the construction of data, revised the original OPDA algorithm, proposed the BFOPDA (based on frame structure OPDA algorithm) model, and carried on the simulation comparison with the existed second-order blind algorithms. The result indicated that although BFOPDA algorithm fails to identify the channel in the condition of quite a few received data (receives 7000 data), it can identify the channel in the condition of mass received data (receives above 35000 data), especially under low signal-to-noise ratio (<12dB), its performance enhanced 1 dB Compared to SSA, and 3 dB compared to OPDA. Next step, unify original outer product decomposition method to revise the Shortcoming of the new algorithm, which can't identify the channel under the condition of low received data quantity, and unify this method with the concrete modulator approach to do the practical application research.