Digital route model aided integrated satellite navigationudand low-cost inertial sensors for high-performanceudpositioning on the railways

摘要

The basis of all railway signalling activities is the knowledge of the position and velocity of all trainsudin the system. The railways traditionally rely on train detection systems for this knowledge. However,udthe dependence of these systems on railway infrastructures limits their ability to cope with theudadvent of new high-speed lines and the development of freight networks across the Europe. Hence,udthere is a need for the introduction of modern positioning technologies into the railways. Unfortunatelyudrailways provide an unfriendly environment for satellite-based radio positioning systemsud(GNSS). For this reason it is common to integrate GNSS with low-cost inertial sensors (INS) butudsuch systems cannot meet all railway positioning requirements. This thesis examines the potentialudof enhancing such an integrated GNSS/INS system with a digital route model (DRM).udThe study is carried out through a series of simulations of typical railway positioning scenes.udA simulated database of GNSS, inertial and DRM data is built from real GPS data collected onuda rail line between Norwich and Lowestoft. Several tests are first performed to test the validityudof the database. Simulations are then done with a number of traditional INS/GPS integrationudarchitectures to test the possible performance of each system in the railway environment using lowcostudINS sensors. The DRM-aiding is then realised through an integration with the GNNS/INSudsystem via an extended Kalman Filter.udResults from the study confirm the need for additional positioning information for an integratedudsystem with low-cost inertial sensors to deal with difficult satellite signal situations such as tunnels,uddeep cuttings and covered stations. It is shown that a DRM leads to significant improvements in theudoverall system positioning performance. Also the optimal configuration, in terms of point spacingudand accuracy, for a digital route model is selected from amongst simulated candidates.