Channel Estimation For MIMO Wireless Communication

In order to accomplish the higher demand for future mobile communication, the method that adopt multiple transmit and receive antennas is calling more people's attention. As of now, the first Multiple Input and Multiple Output (MIMO) system named V-BLAST in the world has been implemented by Bell laboratory. However, a recognized MIMO standard hasn't been formed yet.The study on wireless channels is the key for mobile communication, including MIMO system.. In fact, any design for a reliable wireless communication system must take channel characteristic into account. For example, a receiver must depend on accurate channel estimation to detect information-bearing signals. Due to complexity of MIMO channel compared with SISO channel, one of the difficulties of MIMO communication is MIMO channel estimation. So a thorough study is necessary.In general, people believe that between any transmit antenna and any receive antenna, there exists a sub-channel and all sub-channels are not mutually independent but correlate to each other in some degree. In this paper, for simplification of MIMO channel estimation, the author assumes the optimal condition of fully independent sub-channels.On the basis of knowledge about MIMO system and characteristic of wireless channel, as the main work of this paper, the author gives several MIMO channel estimation algorithms for single carrier time-invariant FIR channel model. Simulation results are effective.In chapter 3, based on likelihood (ML) criteria and least square (LS) theory ,channel estimation is achieved by sending training sequence in conventional method, and a new scaled LS (SLS) approach to the channel estimation is studied. Though, these training sequences are not random ones, but the optimal choice of training sequences to obtain as accurate channel estimation as possible.Recently, some scholar mentioned implicit sequences, send the symbols that periodic sequences superimposed on information symbols. By utilizing the characteristic of periodic sequences, channel estimation can be achieved only based on first-order statistics of receiving symbols. The approach need not assign extra time-slot to training sequences and no loss in bandwidth, the speed of convergence is big. But a higher peak-to-average symbol power ration as its disadvantage arises. In chapter 4 , the author explores a fast channel estimation algorithm for SISO. In order to improve channel estimation's precision, in chapter 5, semi-blind channel estimation and detection using superimposed training's first-order statistics were studied.In this paper, author used non-blind and semi-blind approaches. The channel estimation arouse a great interesting by using superimposed training, owe it's high spectral efficiency and fast arithmetic. This paper emphasis on channel estimation using superimposed training's first-order statistics, adopted LS theory, compared the complexity of arithmetic in different training sequences. In simulation, find out some optimal sequences to improve channel estimation's precision.