An integrated map-match algorithm with position feedback and shape-based mismatch detection and correction

Global Positioning System (GPS)-based vehicle navigation systems have been rapidly developed over recent years. In urban areas, due to the problems of satellite signal blockage and severe multipath effects, GPS may not be available for a long period of time. Integrated GPS/Dead Reckoning (DR) systems are commonly used to bridge GPS position gaps. Moreover, map-matching (MM) techniques can be implemented to improve positioning accuracy by locating an estimated vehicle position (from GPS/DR) on road; however, the reliability of map-matching methods is still a problem because vehicle position may be located to an incorrect road section (mismatch) due to large vehicle positioning errors from GPS/DR and complexity of road network, which is a typical environment in urban areas. In this article, we present a tightly integrated vehicle navigation system to overcome these problems. Performance of any MM algorithms is closely linked with position data quality. In our approach, we integrate MM results with GPS/DR output at the sensor level for data quality control and optimal estimation of sensor errors and vehicle position. Thus quality of position input to MM unit is significantly improved. Furthermore, GPS/DR output, together with road network, is used to detect and to automatically correct any mismatches occurred. We have implemented this integrated approach to a prototype vehicle navigation system and extensive tests have been carried out in Hong Kong urban areas. Testing results demonstrate that the performance of the new system is significantly improved in terms of accuracy, coverage, and reliability.