This event on 11-12 March was focused on applications and sensing technologies, and validation processes for AVs.
Radar:
Mercedes-Benz’s Dr Jürgen Dickmann gave interesting feedback on the state of the art of the radar technology and the lessons learned from the development and launch of the ‘Drive Pilot’ feature on the S- class (traffic jam pilot up to 60 km/h):
- Radar is a key sensor to build a robust L3 feature, and provide a 360° sensing information (10 radars on the S-class)
- There are two types of radar on the S-class: a high-resolution long range imaging radar (iLRR), a many mid-range radars (MRR)
- An imaging radars require a minimum network of 32 × 32 elements to support an L3 feature
- In the future, Mercedes expects to have only one type of radar—the MRR—and to achieve the performance of the current imaging radar through a virtual antenna combining many MRRs (distributed aperture radar)
- Advanced signal processing and AI allow to clean the detections to a level similar to the one achieved with the Lidar technology and its high-resolution imaging capability.
Robotaxis:
Waymo’s Senior AD Product Leader Shweta Shrivastava gave the status of the validation process and safety records of the Waymo L4 robotaxis:
- Waymo relies on camera, radar and lidar technologies to achieve a robust L4 feature and to optimize the ODD (Operational Design Domain), covering many environmental and road conditions.
- Waymo validations required Millions of miles on public roads and billions of miles of simulations. The current system is the 5th generation.
- After over 7 million customer rides, the Waymo Drive has a good safety record and allows to reduce, in theory, the number of crashes by 85 per cent.
- Waymo will continue to enhance the performance of their L4 features, with a step-by-step deployment to guarantee a safe approach of L4validations
- Road Safety could be one of the benefits using robotaxis, with each year 50 million injuries and 1.35 million deaths on the road, half of which are vulnerable road users (pedestrians, motorcycles, cyclists…).
L3 features
BMW introduced the ‘Personal Pilot L3‘ on the new iX7 (highway traffic jam pilot up to 60 km/h)
- The validation process was managed with Ansys as a partner
- BMW collected 24.5 million km of test data with simulations and performed thousands of acceptance test to validate the feature for the end user
- The functional tests were based on NCAP, ALKS scenarios
