Maximising results of novel road safety research
On Thursday November 12th, The SAFE-UP consortium kicked-off a new series of knowledge transfer webinars with fellow R&D&I projects OSCCAR and HEADSTART.
The ultimate goal of this webinar series is to facilitate synergies between SAFE-UP and other related innovation projects, identifying State-of-the-Art results to build upon, as well as raising awareness of available tools and databases for partners.
The OSCCAR project (Future Occupant Safety for Crashes in Cars), which focuses on vehicle occupants, was represented by Werner Leitgeb (ViF, project coordinator) and Daniel Schmidt (Bosch).
Werner Leitgeb first presented a general overview of the project, focusing on:
The demonstration of new advanced occupant protection principles and concepts addressing future desired sitting positions enabled by Highly Automated Vehicles (HAVs);
contributions to diverse, omnidirectional, biofidelic and robust Human Body Models (HMBs);
and the project’s contribution to the standardisation of virtual testing procedures and promotion of HBMs acceptance, to pave the way for virtual testing-based homologation.
Daniel Schmidt then explored the project’s methodology to determine Future Safety-Critical Scenarios (FSCS). Specifically, he explained through combining a bottom-up approach (based on existing technologies) and a top-down approach (assuming a scenario with a full penetration rate of HAVs), OSCCAR foresees a reduction of the number of car accidents by 2025 (see slide 19 of this presentation).
Some specific project results of the publicly available OSCCAR deliverable D1.1 were also presented:
According to data stemming from GIDAS database (German In-Depth Accident Study), one-third of fatal accidents in Germany could be avoided, assuming full penetration of Autonomous Driving (AD) (see slide 20 of this presentation). Based on this data, OSCCAR has performed accident re-simulations while one participant is equipped with autonomous driving functions. In a next step, the OpenPASS framework will be used to confirm the first results of accident re-simulations.
In an exemplary case of urban intersections (not weighted to European level!), accident re-simulations show if the main accident initiator is the HAV, 71% of accidents could be inherently avoided (assuming obeying traffic rules, no effects of alcohol, etc.), 19% could be avoided and 10% could be mitigated (e.g. thanks to speed reduction before the crash). On the other hand, if the HAV is the opponent car and the human-driven vehicle is the initiator, 37% of accidents could be avoided by the HAV and 49% could be mitigated.
The main conclusion that can be taken from these unweighted examples is that casualties in the future could be greatly reduced, assuming a full penetration of HAVs in the transport network.
The HEADSTART project (Harmonised European Solutions for Testing Automated Road Transport), is focused on defining testing and validation procedures of Connected Automated Driving (CAD) to be standardised at EU level. The initiative was represented by Álvaro Arrúe (IDIADA, project coordinator) and Nicolas Wagener (IKA).
Alvaro Arrúe highlighted the multi-disciplinarity within HEADSTART, as different key enabling technologies (KETs) are pillars of the project: communication (e.g. vehicle connectivity), cyber-security, and positioning & localisation of the vehicle.
He also introduced HEADSTART’s five-step roadmap, which outlines the methodology set to achieve the overall goal of the project: to validate safety and security performance according to the needs of key user groups that would leverage project results, including technology suppliers (industry and academia), as well as consumer testing and type approval (regulation and legislation entities).
Nicolas Wagener focused his presentation on the testing and validation methodology of HEADSTART. When it comes to the validation and verification of AD functions, sample calculations show that a huge number of driving kilometres in field tests would be necessary. This is not feasible for modern AD functions. Therefore, the HEADSTART methodology builds upon a scenario-based data-driven approach by abstracting real-world drives to scenarios leveraging large databases. Further details on the methodology can be read here.
Finally, Alvaro Arrúe also explained how HEADSTART selected the three use cases (out of 5 initially pre-defined) to validate the project methodology. To do so, they combined internal investigation, surveys, interviews and workshops, aligning them also with aspects of the ERTRAC roadmap: truck platooning, highway pilot and traffic jam chauffeur.
HEADSTART will closely cooperate with other related initiatives to further enable the demonstration of the HEADSTART methodology, particularly with:
ENSEMBLE - related to Truck platooning.
CAVRide (via IDIADA) - related to Traffic jam chauffeur.
Automated Drive Demonstrator (via Virtual Vehicle) - related to Highway pilot and Traffic jam chauffeur.
This cooperation will also provide SAFE-UP with the opportunity to engage with a broader number of relevant R&D&I projects, and the consortium looks forward to building such collaborations in the near and distant future.
You can view both project presentations by clicking below.
