As part of their autonomous vehicle projects, car manufacturers are looking for a geolocation solution that is both reliable and economically viable. In a context of strong ethical and economic issues, accurate and reliable positioning is an essential building block for these projects.




SYSNAV is developing an adapted version of its magneto-inertial technology for cost effective geolocation of autonomous vehicles. The current phase of the development of the SYSNAV solution aims to replace RTK (GPS) as it is too expensive for mass production, unreliable, and has limited availability. The SYSNAV solution integrates the following technologies into its Blue Force module: LIDAR, cameras, GNSS (e.g. GPS), odometer, cartography, software algorithms, and of course SYSNAV’s core magneto-inertial technology.


To correct the bias of the inertial sensors, the SYSNAV module regularly verifies its estimated position by using external information (detection of marks). For example, the RLAT uses visual recognition to identify the white lines on the roads and estimate the bias of the inertial sensors. In order to evaluate the performance of the data fusion algorithms, the trajectory calculated is compared to that of the RTK (GPS) integrated in the test vehicles as it is assumed to be the best information currently available on the actual trajectory when the GPS signal reception conditions are optimal.


Conversely, the sensor’s location information is used to improve the geolocation accuracy of objects used for landmarks detection and “map matching”. The algorithms designed to exploit a means of lateral registration based on the fusion between camera data and HD mapping make it possible to find performances comparable to what was obtained during the first phase of the project when RTK was exploited. When RTK is active and available, the positioning accuracy is ~ 3 cm. SYSNAV’s solution can ensure a 30cm accuracy at full speed (90 km / h) after the loss of any registration signal (GPS, lane matching, etc.) for 15 sec. (ie 400 m)