Channel Estimation Based On LS For MIMO-OFDM System

1. Research contentOFDM technology is a sort of multi-carrier transmitting technology which multi-carrier compartment perpendicularity in time domain that could utilize spectra resource efficiently. In addition, OFDM makes frequency selective fading channel became smoothness channel in frequency domain that the impact of multi-path fading became clear. Thusly, for exploitation of two potential technical fully, OFDM and MIMO should be hand in hand. That will reach two effects: One is the taunt transmission rate of the communication system, the other is very strong reliability by diversity, thereby realizing high speed and high capacity TOR in limited spectra resource. It would be the kernel technical solution of future broadband wireless communication.The performance of future broadband wireless communication would still be constraint by wireless channel. Because both transmitters and receivers working in complicated wireless channel environments which have got prodigious unpredictability and randomness result in distortion of amplitude, phase and frequency of signals. These issues make a very large challenge for MIMO-OFDM system design. Wherefore,channel estimation for MIMO-OFDM system should comprehensive consideration about multi-antennae and frequency orthogonal.Channel estimation for MIMO-OFDM system usually be divided three classes: Estimation base on pilot, blind estimation and semi-blind estimation. From the realizable of MIMO - OFDM system, based on pilot method should be the major research emphasis. This type of algorithm performance ably and simplicity, is able to realize and has field of application extensively. The rationale of channel estimation base on pilot channel algorithmic is estimating channel parameter by known discipline sign.Channel estimate algorithm base on least-squares estimation for MIMO -OFDM system is an emphases subject in this paper. And time-frequency domain co-estimation for ls algorithm is well studied. According to the lack of subsistent algorithms, a refinement PSM in LS-DFT algorithm will be introduced. Both two algorithms are using LS-DFT algorithm as rough estimator in frequency domain and refinement algorithms in time domain for decreasing operand and increasing precision. Finally veritable effect by simulation, the result proclaim ameliorating performance of channel estimation by taking full advantage of relevance between time domain and frequency domain in time variant wireless channel.2. PSM in LS-DFTThe structure of OFDM symbols decide that estimation in time domain and frequency domain can work at the same time in MIMO-OFDM system. But time-frequency co-estimation is complex. Wherefore for the moment method that estimating in frequency domain firstly and in time domain lately is in general using. In this paper, PSM is using in time domain with the co-estimation algorithm base on LS-DFT. Usual PSM adopts true path by threshold M. In this paper, method is improved by screening out signal energy paths for estimation which considering the impact of noise energy paths and energy-leak paths. This algorithm not only can reduce LS-DFT algorithmic operand of LS-DFT algorithmic,but also advance ability of anti-Gaussian-white-noise and anti-signal-energy-decadent. Emulation proved that this refinement algorithm had great performance than LS algorithm in middle and low ranks of SNR. This algorithm could be used as a rough estimator for other base on LS-DFT time-frequently co-estimation in middle and low ranks of SNR wireless channel environments. 3. LS-LMSLMS algorithm own merits, such as temperament of calculated amount and easy to realize, has major defect that rapidity of convergence is impacted by step. On the basis of such point, incorporation LS-DFT time-frequently co-estimation algorithmic structure, paper lodges a sort of refinement algorithm base on LS-LMS channel estimation, which utilizes LS-DFT as rough estimate, simply operand by PSM in time domain, and using LMS estimation which is trimmed by step setting dynamically that would make algorithm fast convergence. Such refinement algorithm could slash algorithmic complexity and improve the lack of LMS algorithm. By simulating this algorithm, experimental result proclaims that LS-LMS algorithm has fine performance and is still fine in inferior wireless channel environments.