Iterative channel estimation for CDMA wireless systems

Code Division Multiple Access, a common platform on which 3G technologies are built, provide higher capacity for voice and data communications and allow more subscribers to connect at any given time. The first stage in the CDMA signal processing receiver comprises the estimation of complex channel coefficients. This channel estimation plays a major role in the successful detection of the user's signals and in this thesis we analyse some existing channel estimation techniques, as well as proposing a new scheme based on the APPA method. This thesis describ~s the WCDMA uplink physical layer in detail and explains the single user transmitter and receiver architecture. One of the main features of CDMA is the use of rake receivers, which incorporates maximal ratio combining techniques to mitigate the effects of fading caused by the multipath channel. A conventional rake receiver has been implemented and its perfonnance analysed in different channel conditions. The perfonnance of the rake receiver improved with an increase in diversity and when the paths that arrived were· separated by more than one chip duration. To estimate the channel parameters, that are needed to correct th.e phase rotations and attenuations caused by the fading channel, a Wiener filter was employed which produced estimates in agreement with the actual channel parameters. The perfonnance .of various spreading codes were also analysed which showed that codes with long spreading gain perfonned better than those with short gains due to the high cross correlation between the short code lengths. An iterative channel estimator based on Expectation Maxim~ation (EM) algorithm is iJ:lplemented and analysed in flat fading and frequency selective fading channels for S1S0 and MIMO systems. Turbo codes well known as powerful error control codes and capable of achieving Shannon bound for channel capacity produce extrinsic information during the decoding process whiCh is used to improve the decoding iteratively; This extrinsic information is exploited in the EM channel estimation process where the estimation operates iteratively and jointly with the turbo decoding thereby providing reliable channel estimates. The iterative channel estimator is found to perfonn significantly better than the non-iterative estimator. A major contribution of the thesis is the development of a new channel estimation algorithm based on the APPA scheme which is also based on the turbo princfpie. The algorithm involves derivation of the log-likelihood function and then the approximation of this function in order to derive maximum-likelihood estimates of the channel amplitude and phase. The log-likelihood function of the APPA scheme consists of the log of the expected likelihood function rather than the expected value of the log-' likelihood function as in the EM scheme. It' is shown to perfonn better than the EM algorithm in flat fading channel conditions for a BPSK turbo coded 'system. Its performance is measured by comparing the mean square estimation error with the Cramer-Rao bound and overall BERperformance against SNRs. A new technique was developed to estimate the phase of the carrier using join~ turbodecoding and carrier phase estimation for differentially encoded OQPSK system. This was done in order to show the benefits of including the turbo decoder in the iterative estimation process. The differentially encoded OQPSK modulation scheme has been adopted by numerous wireless communication systems especially satellite communication systems.EThOS - Electronic Theses Online ServiceGBUnited Kingdo