MIMO Channel Estimation In Time-Frequency-Doubly-Selective Fading Environment

The technology of multiple input and output system (MIMO) can satisfy the demand of the growth of the mobile service because it can obviously enhance the channel capacity. In a reliable system the receiver must know the state of the channel (CSI) which can make the signal detection and the channel coding and decoding carrying on smoothly. So the MIMO channel estimation is important.Firstly we analyze the SISO system in the time-domain and frequency domain separately and analyze the channel characters of the frequency-selecting,time-selecting and tine-frequency-selecting channels based on the BEM which is more practical. We build the MIMO double-selecting channel model based on BEM and then analyze the channel capacity of the MIMO block fading system. The simulation experiment indicates that the capacity can improve about 2.5bits/s/Hz when the sending power is adapted assigned and the capacity is decreasing along with the increasing of the correlation coefficient when the sub-channels are correlative.Then we mainly analyze the PSAM channel estimation algorithm. The traditional algorithm sends the pilot sequence firstly and then sends the information sequence only, so the algorithm can't track the fast time-variable channel. So we insert the pilot sequence into the sending sequence and the entire block data is divided into many sub-blocks. So we can track the channel more precisely. We then deduce the LS estimating value,the lower bound of the capacity and the bit error rate. Then we deduce the optimal PSAM parameters which can minimize the MSE and the BER and maximize the capacity. The simulation experiment indicates the arithmetic can track the channel more precisely than the traditional algorithm.Because the PSAM algorithm must allocate the slots for the pilot sequence so the capacity is decreased. The superimposed training channel estimation algorithm(ST) can improve the capacity but the MSE is higher. So we propose the improved ST algorithm which is called DDST algorithm which greatly reduce the MSE and BER and can remain the capacity the same as the ST algorithm. We then estimate the MIMO double-fading channel using DDST algorithm and deduce the LS estimating value of the CSI array. We then deduce the lower bound of the capacity of the DDST and the PSAM respectively. The simulation experiment indicates the MSE and BER of DDST is slightly higher than the PSAM but much lower than that of the PSAM algorithm. The experiment also indicates that the capacity of the PSAM is sensitive to the number of the sending and receiving antennas but the capacity of the DDST is slowly increasing with the increasing of the number of the sending and receiving antennas.