VW and Bosch are using data from the Golf 8 in Europe and a digital twin model to improve the performance of self-driving cars.
While on the road, the vehicle fleet uses surround sensors to generate information about landmarks such as road signs, guardrails, curbs, and lane markings to help improve high definition maps. The vehicles, starting with the Golf 8 in 2023 and moving to other models, send the data completely anonymously via the VW cloud to the Bosch cloud.
The road signature is created in the Bosch cloud, producing a digital twin of the real environment. Information from radar and video sensors, as well as vehicle motion data, augments common navigation maps with additional layers for vehicle localization and control. These additional layers are compatible with typical map formats.
- Europe to build a digital twin of the earth
- Siemens looks to digital twin, IoT to drive growth
- First international digital twin standard
- ST builds digital twin of Crolles fab
- Nexans in 100 day digital twin move with Microsoft
Using the Bosch road signature, vehicles can determine their position with a high degree of accuracy: in real time, comparing the information currently provided by its surround sensors with that of its digital twin. This comparison enables the cars to accurately determine their position in the lane down to a few decimeters relative to the highly accurate map. The use of radar means that localization works reliably even in adverse weather conditions such as fog, rain, and snow – conditions which make it difficult, if not impossible, for a camera to perceive its surroundings.
Bosch plans to use current data to continuously expand the signature and keep it up to date says Dr. Mathias Pillin, president of the Bosch Cross-Domain Computing Solutions division. “The more vehicles that provide information now and in the future, the larger and more robust the database will be for automated and assisted driving,” Pillin says.
In addition, the road signature enables safer and more convenient lateral and longitudinal guidance of automated vehicles, as it contains information not only about landmarks, but also about road geometry, lane layout, road signs, and speed limits. It even includes typical driving behaviour at specific locations, providing data on how human drivers approach a curve or when they brake for an intersection. All of this will help driverless cars improve their algorithms says Pillin.
Other articles on eeNews Europe