Channel estimation and signal detection with low complexity for time-varying frequency-selective fading channel

Abstract

In this thesis, we consider channel estimation and signal detection based on the low computational complexity model. This has an advantage of low computation, due to the linear approximation of channel for the OFDM block. This model consider for the ICI terms. Since the existing detection methods ignore the ICI term, they have some losses due to the time variation of channels. Using this channel model, we provide channel gain for the time-varying and eliminate the ICI term from the received signals. Based on this low computational complexity model, we improve the performance for the high Doppler frequency. We review the physical characteristics of channel. By using this physical characteristics of channel, we modulate two adjacent sub-carriers. But original coding method does not consider the channel estimation scheme. We modify the transformation matrix, which is adaptive for the pilot-tone set. Because pilot-tone placement has a great impact on the performance of channel estimation, we provide the optimal pilot-tone set. We propose the channel estimation and the signal detection algorithms for the time-varying frequency-selective fading. Both the channel estimation and the signal detection have better performance than the original low computational complexity model. By using the coding algorithms, we compensate the linear approximation for the OFDM symbol. Moreover, we still have the low computational complexity. We propose the ICI cancellation method at the receiver. Due to the modified transformation matrix, we obtain more exact channel gain. This reduces the potential error, when we detect the transmitted signal by using iterative ICI cancellation method. The proposed method has better performance with low computational complexity for the fast Doppler frequency.