A Novel Toa Location Algorithm Based On Dynamic Compensation Using Linear Location Prediction In Nlos Situations

Wireless location has recently become one of the most important issues in the 3G and B3G technologies. The problem of how to improve the accuracy of wireless location is ranked at the top among all other challenging problems. An intrinsic characteristic of the wireless network that significantly prevents wireless location from achieving accurate positions is the ubiquitous Non-Line-of-Sight (NLOS).Inspired by Linear Prediction and Dynamic Compensation, a novel Time-of-Arrival (TOA) location algorithm, Linear Location Prediction (LLP) , is proposed to address the aforementioned problem. In this dissertation, we use the initial values of the Mobile Station (MS) as the indispensable parameters in the following Linear Location Prediction stage. Given these initial parameters, the next state of the MS can be predicted via the predicting mechanism by the Mobile Location Centre (MLC). As a result, the most probable location and the corresponding parameters of the MS can be estimated in advance. The estimated data will not be deployed as the next state of the MS unless they entirely meet the constraint defined in the algorithm.In addtion, we also perform an in-depth analysis on the valid hand-off range issues which will probably be encountered in wireless location in order to guarantee the three BSs which paiticipate in positioning are always the ones which are closest to MS. By taking into account the situation that the MS may be staying around the border, a novel criterion for hand-off initiation is defined. A complete hand-off will not be finished until the consecutive time the MS does not trigger another hand-off exceeds a pre-defined value. Finally, the theoretical and simulation analysis concludes that the method proposed in this dissertation yields better performance in terms of higher accuracy and greater robustness to NLOS contamination compared to other algorithms.