A novel deconvolution approach for high accuracy LOS estimation in WCDMA environments

A prerequisite for reliable mobile positioning in WCDMA systems is the estimation with high accuracy of the line- of-sight (LOS) propagation delay between 3 or more base stations (BSs) and the mobile station (MS). LOS component may be followed by non line-of-sight components arriving within less than one chip interval. Correlation based methods are limited in accuracy by the width of the main lobe of the signal autocorrelation function (ACF). In the presence of square root raised cosine (RRC) pulse shape and strong interference from neighbour BSs, traditional deconvolution methods, such as least-squares (LS) and projection onto convex sets (POCS) algorithms fail to provide good accuracy of LOS estimates. By incorporating additional constraints in the iterative steps of POCS algorithm, we show here that increased resolution in LOS estimation can be achieved even under severe interference conditions. The link-level simulation model was built in conformity with current WCDMA specifications.