A low rank LMMSE(linear minimum mean square error) channel estimation with variable rank is proposed for MB-OFDM(multi-band OFDM) system in this paper. A comparison is performed with an LS(least square), LMMSE channel estimation, respectively. In the real time system, it is difficult to implement LMMSE channel estimation because of its high complexity though which superior channel estimation performance. On the other hand by using the low rank LMMSE algorithm with proper rank size, we can reduce the complexity of the LMMSE but without channel estimation performance degradation. We have analyzed the system performance of LS, LMMSE and low rank LMMSE with variety of the number of rank under the IEEE 802.15.3a channel model. The performance analysis provides an appropriate number of rank depending on the channel characteristics. Low rank LMMSE can be implemented in the real time with obtained number of rank without performance degradation, which is shown by our simulation results.
Published in:
Ultra-Wideband, 2007. ICUWB 2007. IEEE International Conference on
Date of Conference: 24-26 Sept. 2007