A sequential test for autonomous localisation of map errors for driving assistance systems

Driving assistance systems are progressively introduced to enhance safety and comfort in passenger vehicles. They increasingly rely on information stored in digital navigation maps. However maps can be obsolete or contain errors, resulting in malfunctions of context based driving assistance systems and possibly generating hazardous situations. This paper aims at making the vehicle able to detect and localise map errors in an autonomous manner using its embedded sensors. The proposed approach relies on the sequential generation and monitoring of residuals. The vehicle estimated trajectory is compared statistically with the geometric data in the map. The approach allows driving assistance functions to know if they can rely on the map in real-time and to store this information for future journeys. The method is very efficient in terms of computational load which makes embedded applications possible. Performance is assessed using vehicle data acquired in real traffic conditions, which is then compared with an outdated navigation map.