Traditional antenna calibration methods for time division duplex （TDD） systems asSume that the flee-space channel remains the same during calibration, which is unreasonable under the high-speed rail and other time-varying channel scenarios, and will cause calibration error due to time variability. This paper proposes an antenna calibration method for time-varying channels. In the proposed method, the transceiver first sequentially sends a pilot signal to ob- tain equivalent do^vnlink and uplink channel responses. Then, by predicting the downlink （uplink） channel response fed back from the receiver using the channel prediction algorithm, the transmitter obtains the channel response correspond- ing to the channel response on uplink （downlink）. Finally, the transmitter calculates the transmission calibration factor through the prediction value. Compared with the traditional antenna calibration method, this method can improve the accuracy of the calibration factor. Simulation results show that the performance degradation of antenna calibration can be caused by time-varying channels and the proposed method can well compensate for the performance loss and sig- nificantly improve the antenna calibration performance for time-varying channels.
A signal detection algorithm is proposed for the orthogonal frequency division multiplexing （OFDM） sys- tem in the presence of fast time-varying channel. The channel is represented by a piece-wise linear variant model with normalized Doppler frequency of less than 0.2. The channel parameters are extracted through ISI/ICI （inter-symbol in- terference/inter-carrier interference） cancellation and circular convolution reconstruction. Meanwhile, an improved OFDM symbol detection algorithm is also proposed based on circular convolution reconstruction. The channel state in- formation in the OFDM symbol duration can be obtained accurately from the adjacent two block pilots in a linear model. The simulation results show that the proposed method can not only track the channel variation, but also promise better performance gain in the OFDM symbol detection. Furthermore, the bit error ratio （BER） is close to the perform- ance with the perfect channel state information.
