| Summary: | 碩士 === 國防大學中正理工學院 === 電子工程研究所 === 93 === A multicarrier code-division multiple access (MC-CDMA) system combining orthogonal frequency division multiplexing (OFDM) and code-division multiple access (CDMA) has been recently proposed for implementation of high data rate system in the future. While MC-CDMA is robust to frequency selective fading channel and supports high capacity services, it is very sensitive to frequency and time offsets compared with single carrier systems. Minor carrier frequency and time offsets not only result in intersymbol interference (ISI), but also cause the loss of orthogonality between subcarriers leading to intercarrier interference (ICI). The performance of MC-CDMA systems may suffer significant degradation.
This thesis proposes two new solutions to the carrier frequency offset and time offset estimation problems without using reference symbols, pilot carriers, or excess cyclic prefix for MC-CDMA communications. The techniques developed here provide a high-accuracy offset estimate by taking advantage of the inherent structure of MC-CDMA signals. First, we take advantage of the presence of virtual carriers in MC-CDMA signaling and use blind estimation reminiscent of spectral analysis in array processing, i.e., minimum variance distortionless response (MVDR), multiple signal classification (MUSIC) and root multiple signal classification (ROOT-MUSIC) algorithms. However, these algorithms also suffer slight performance degradation when the statistic information of symbols and number of users are insufficient. To mitigate these effects, the MVDR and MUSIC estimators with a spreading code-aided technique for carrier frequency offset and time offset estimations are also proposed in this thesis. Finally, performances of the proposed estimate methods are accessed by simulations for a downlink MC-CDMA system.
|
|---|
