TNO introduces Streetwise scenario-based methodology for validation of automated driving

07 Jul 2017

Automated Driving is developing rapidly. Full evaluation of the safety of automated driving systems is difficult or even impossible, as the system needs to deal with ‘all’ situations on the highway or city streets. This ranges from normal every day driving conditions to near-critical or critical conditions. Moreover, there is no consensus on when an automated vehicle is safe enough to be launched onto the street.

At the Autonomous Vehicle Test and Development symposium in Stuttgart (20-22 June 2017), TNO introduced the Streetwise methodology and analysis pipeline for validation of automated driving. At the heart of this methodology is a real-life scenario database which will contain typical driving situations and their variants. The focus is on interaction between road users (vehicles, pedestrians, cyclists), road elements (tunnels, bridges, traffic signs, trees, trash cans, etc) and environment (such as weather and lighting conditions). The scenarios will be based on public road driving data.
Scenario-based  safety validation of automated driving is broadly supported by the automotive community. This is reflected in a draft standard of NHTSA and the ISO 26262 working group on SOTIF (safety of the intended function). Related projects in Germany (Pegasus) and EU (ENABLE-S3) strongly support this approach.

Streetwise features a completely data driven approach. Scenarios are not just created by good engineering but primarily from using real driving data from cars with radar and/or camera. This turns the cars into sensors for scenarios. The driving data is classified into elementary events of one car (like lane departure) and then combined into scenarios (like overtaking on a highway). The variation observed across recorded scenarios will be statistically characterized so that the full variety of driving on the road becomes available, e.g. for the generation of relevant test cases.

A key issue is to estimate the completeness of the scenario database: how well do the collected scenarios cover the variety on the road? Did we capture ‘all’ situations? TNO introduces completeness metrics on various scenario levels. It tells you how much new information is in the last 1000 km of driving. In this way, the progress in completeness is monitored. This is an important step in proving the safety and comfort of automated driving.

Interested? Please contact us.


Research on integrated vehicle safety

TNO is recognized by OEMs and TIERs all over the world as a valuable knowledge partner with unique expertise, tools and facilities. The activities within the research department Integrated Vehicle Safety... Read more

Sytze Kalisvaart, MSc

Media questions?


Stay up to date with our latest news, activities and vacancies collects and processes data in accordance with the applicable privacy regulations for an optimal user experience and marketing practices.
This data can easily be removed from your temporary profile page at any time.
You can also view our privacy statement or cookie statement.