On The Sample-Shifted And Code-Aided Channel Estimation For Multicarrier CDMA Communication Systems

Multi-carrier communication systems will play key role in the future broadband mobile communication systems. However, when the transmission rate is high, accurate estimation to the channel parameters becomes a critical problem. This thesis focuses mainly on the channel estimation techniques for multi-carrier CDMA systems, including pilot-based channel estimation and code-aided channel estimation.The thesis first gives an outline of the multi-carrier systems, and then discusses the channel characteristics. The channel influences the signals in two ways, large scale fluctuation and small scale fluctuation, and the latter one is the emphasis of this thesis. Small scale fluctuation can be described in two scenarios: time dispersive and frequency dispersive. Delay spread causes inter-symbol interference, and the relative motion of the mobile station will cause Doppler frequency shift, resulting frequency spread.Next, the system structure of OFDM and multi-carrier system are discussed, followed by the pilot based channel estimation for the multi-carrier CDMA systems. Extensive simulation and comparisons are carried out for the four interpolation techniques: linear interpolation, second-order interpolation, cubic spline interpolation and low-pass interpolation. After investigating a sample shifted channel estimation technique, this thesis proposed an improved robust channel estimation technique.Then, a code-aided channel estimation technique for multi-carrier systems is investigated. It is shown that, in the presence of powerful error-correcting code, the conventional pilot-based and blind channel estimation techniques will have a poor performance. To overcome this problem, it is recommended by some researchers to exploit the information provided by the code, i.e. to make use of the iterative Turbo principle and EM algorithms. By making iterative operation between estimation and detection, it is possible to approach a higher performance. Based on this technique, an improved code-aided channel estimation algorithm is presented in this thesis. It is shown through computer simulation that the proposed improved algorithm provides more accurate estimation than the original algorithm.