Analysis And Design Of Selecting Satellites In The Application Of GPS In Train Positioning

With the development of satellite navigation and positioning technology, GNSS (Global Navigation Satellite System) is widely used in train positioning system. Though GNSS technology has such advantages in positioning, the accuracy of location determined by onboard receivers need to be improved.For orbital error, ionosphere and troposphere delay error and multi-path error the accuracy of onboard receivers is affected. Multi-path effect occurs for the mask of obstacles along the track. GNSS signal arrives at the receiver through different paths, such as direct signal and reflected signal from trackside objects surrounding the receiver antenna. No identical model could be constructed to eliminate multi-path effect for different track circumstance. When train moves in the section with many tunnels, receivers work between unlocking and locking states. Re-acquiring high accuracy position cost much time.A method based on selecting satellites is proposed to be used in the specific receivers to eliminate multi-path effect and shorten the time of receivers locking satellites for train location. It is processed by acquiring the trackside circumstance based on stereo vision technology, estimating and selecting valid satellites in direct view of onboard receivers, excluding satellites affected by multi-path effect in order to reduce location errors. Obstacles along trackside are recorded by DC/DV. The 3D information of the objects is reconstructed via stereo vision technology. The validity of the satellites is determined through comparing the position between satellites and obstacles. The constellation combinations with the highest precision, as the reference input of software receiver, implement positioning calculation. The 3D information of trackside objects could be stored. After the whole line information determined, receivers select valid satellites directly from circumstance database. The receivers could estimate valid satellites fast in sections with tunnels, which may shorten positioning time and acquire high accurate position.The method is verified by simulation experiment. The objects surrounding a measured station were recorded by DC and processed based on stereo vision technology. The final results show that this method is effective in excluding the multi-path effect and improving the accuracy.