Geomatic: AI in the Service of Automotive Testing_
The Geomatic project addresses a crucial need in the automotive industry: automating the generation of typical test routes for vehicles.
Automating Route Creation_
Historically, the creation of these routes – combining specific percentages of urban, rural, and highway driving with slope constraints – was done manually, a time-consuming and complex process. The core of the Geomatics project lies in developing an algorithm capable of automatically generating routes that precisely match the requirements provided by engineering teams.
This innovative approach integrates:
Precise geographical data from OpenStreetMap.
Elevation data to incorporate slope information (Aster data).
User-defined criteria: starting point, total distance, proportions of city/road/highway, slope proportion and threshold.
Methodology and Technical Challenges_
Our main tasks in this project included:
Acquiring and processing a large volume of raw data from OpenStreetMap to generate a road network graph around the user-specified starting point.
Enriching the graph with key information: each road segment is categorized (city, rural road, highway) and assigned slope data. The road categorization, tailored to each country's specifics and based on speed limits, was a key topic of discussion with the project teams.
Simplifying and "cleaning" the graph: to ensure route consistency (avoiding unnecessary U-turns or dead ends), an in-depth simplification process was required. This step is critical to generate diverse, error-free routes.
Developing the route generation algorithm: Points are uniformly distributed across the map, shortest paths are calculated between them, and a simplified graph is built. This approach significantly reduces problem complexity and adapts to tolerance levels relative to distance.
These steps required advanced engineering to handle large volumes of geographical data and to design a robust algorithm capable of meeting performance and accuracy requirements. The complexity lay not only in calculating the routes but also in ensuring the generation of a diverse set of valid paths.
The Geomatics project demonstrates the potential of artificial intelligence in automotive testing. By automating route generation, it transforms a complex manual process into an efficient, reliable, and reproducible digital solution tailored to engineering needs.
Benefits of the OSE Solution_
The Geomatics project offers significant operational benefits to engineering teams:
Significant time savings: automation replaces hours of manual work
Standardized testing: algorithmic generation ensures consistent and reproducible routes
Optimized test campaigns: routes are tailored to the specific needs of engine testing
The expertise acquired by OSE Engineering during this project, particularly in handling and leveraging OpenStreetMap data and developing route generation algorithms, opens new possibilities for future applications in road network analysis and traffic statistics generation.
