In the conventional iterative superimposed pilot-assisted channel estimation (SI-PACE), simple averaging of the instantaneous channel estimates obtained by using the pilot over several single-carrier (SC) blocks (called the frame in this paper) is taken in order to reduce the interference from data symbols. Therefore, the conventional SI-PACE has low tracking ability against fading time variations. To solve the tracking problem, Wiener filtering (WF)-based averaging can be used instead of simple averaging. However, WF incurs high computational complexity. Furthermore, the estimation error of the fading autocorrelation function significantly degrades the channel estimation accuracy. In order to improve the channel estimation accuracy while keeping the computational complexity low, a new iterative SI-PACE using sliding WF (called iterative SWFSIPACE) is proposed. The channel estimation is done by sliding a WF having a shorter filter size than the measurement interval. The bit error rate (BER) and throughput performances of SC-FDE using iterative SWFSIPACE are investigated by computer simulation to show that the proposed scheme achieves good BER and throughput performances while keeping the computational complexity low irrespective of the fading rate (or maximum Doppler frequency).
- Autocorrelation function estimation
- Superimposed pilot
- Wiener filter