On The Channel Estimation Design For TD-LTE System

It is well known that3GPP Long Term Evolution (LTE) project is able to improve high bandwidth efficiency and to provide high-quality broadband services, thus making itself one of the most important candidates for the next generation of broadband mobile communications. Multiple input multiple output (MIMO) technology and orthogonal frequency division multiplexing (OFDM) technology are two critical enabling techniques to increase the system capacity and enhance the data transmission rate for LTE system. However, accurate channel state information (CSI) is the pre-requisite to realize coherent MIMO-OFDM transmission and to secure the reliable reception at the receiver side. Meanwhile, the accurate CSI is also important for the reasonable precoding design at the transmitter side. Therefore, in this thesis, we laid our emphasis on the channel estimate scheme suitable for the uplink shared channel as well as the downlink shared channel forFirstly, the thesis derived the equivalent frequency domain channel model of MIMO-OFDM system according to three reference channel models specified by TD-LTE standard. Based on the derived frequency domain channel model, the pilot based channel estimate schemes are presented and analyzed. More specifically, this thesis presented and compared some widely utilized channel estimate algorithms, such as the LS estimate, the MMSE estimate, the LMMSE estimate, the SVD-MMSE estimate as well as the DFT-based channel estimate schemes. As for the channel estimate at data position, this thesis analyzed the interpolation-based channel estimate based on the pilot position channel estimate. And different interpolation based channel estimate schemes, such as the linear interpolation, Gaussian interpolation, DFT filtering based interpolation and Wiener filtering interpolation scheme are compared. Then based on the pilot characteristics of both the physical uplink shared channel (PUSCH) and the physical downlink shared channel (PDSCH) channel, the thesis presented the suitable channel estimate schemes for TD-LTE system, simulation results are presented to validate the achieved estimate performance.Finally, based on the time-varying characteristics analysis and the resultant channel correlation between successive subframes, a joint Wiener filtering based interpolation was proposed in this thesis to improve the channel estimate performance by exploiting pilot placement regularized by the LTE system frame structure. It is shown through simulation that, the proposed joint Wiener filtering based interpolation is capable of significantly improve the channel estimate at the data position without increasing the system realization complexity. Our analysis unveils that, based on the channel estimate from pilot symbols, the proposed joint Wiener filtering based interpolation estimate can achieve a mean square error estimate performance which is more close to the theoretical lower bound, especially in the context of the reference ETU channel.