The Mobile station is connected to satellites that retrieve the information about coordinates using GPS and network
base station tracks the location from server base station of location database. The proposed scheme utilizes the two-step
Least Square method for estimating the three-dimensional position (i.e. the longitude, latitude, and altitude) of the
mobile devices. The Kalman filtering technique is introduced to eliminate the measurement noises and to track the
trajectories of the mobile devices. The simulation result shows the consistent location estimation accuracy under
different environments.
In order to achieve the minimum weight conditional variance of importance weight and get more accurate estimation
of sight condition, the optimal trial distribution is used. Then by applying decentralized Kalman filtering method, the
mobile state could be analytically computed. In the parameter learning step, sight conditions are updated according to
the measurement and the estimation and mobile state. Simulation results show that this method could achieve a good
tracking performance and the NLOS parameters can be effectively inferred.