Subspacer-Based Blind Channel Estimation For MIMO-OFDM Systems

Multiple-input multiple-output orthogonal frequency division multiplexing(MIMO-OFDM), which uses the diversity of time、frequency and space to assure thequality of the transmission signal, is quite a promising technology implementationscheme in the next generation mobile communication system. Channel estimation, as akey technology of the wireless communication system and the necessary condition forthe coherent detection of MIMO-OFDM system, has been paid more and more attentionby researchers.This paper firstly introduces the basic principle and the simplified matrix model ofMIMO-OFDM system, and then gives a brief description of the classification and keytechnologies of the two channel estimation algorithms. We mainly focus on the noisesubspace-based blind channel estimation algorithm in this paper because of its simplestructure and good performance. A noise subspace method based on repetition index,which utilizes repetition index rather than cyclic prefix to add data redundancy, isproposed at first. It can greatly decrease the receiving samples needed for channelestimation and relax the requirement that the unknown channel must remain timeinvariant during estimation. In order to reduce the complexity of the proposed method,fast data projection method (FDPM) among subspace tracking algorithms has beenfurther improved. The new proposed algorithm can not only reduce the complexity ofthe channel estimation, but also decrease the required number of received OFDMsymbols. Simulation results show that, compared to the conventional noise subspacemethod, the value of the normalized root mean square error has increased, while the biterror rate has been almost the same in the new improved algorithm.