De benodigde kennis wordt in samenwerking bepaald tijdens werkbijeenkomsten en Ronde Tafels. Dit leidt tot kennisvragen die worden beheerd door het Ministerie van Infrastructuur en Milieu (DGB).
De onderstaande vragen zijn al deels beantwoord. De antwoorden staan in de documenten uit de collectie en zijn hier verzameld. (stand per maart 2017) Voor meer vragen en antwoorden: zie het Kennisjaarverslag.
Chapter 3/ Conclusion
Self-driving vehicles could compensate for some but not all crashes caused by other traffic participants (Pedestrian error could be compensated by AV). Lighting failures might turn out to be irrelevant to safety from the perspective of being able to control one’s vehicle at night, because self-driving vehicles might not rely on visual input. / (1) The expectation of zero fatalities with self-driving vehicles is not realistic. (2) It is not a foregone conclusion that a self-driving vehicle would ever perform more safely than an experienced, middle-aged driver. (3)During the transition period when conventional and self-driving vehicles would share the road, safety might actually worsen, at least for the conventional vehicles.
17.4 Significance of Possible Predictions based on Accident Data
Hoe gaat mogelijk systeemfalen van de zelfrijdende auto ondervangen worden? Denk hierbij aan het uitvallen van communicatie, storing in elektrische systemen, maar ook in mechaniek van de zelfrijdende auto. Welke fall-back opties zijn te creëren zodat dit opgevangen kan worden? Hoe wordt dit gedaan in andere infrastructuurnetwerken?
A common element in all levels of automation is safety-critical electronic control systems. There are voluntary industry standards, such as ISO 26262, which establish uniform practices for specific levels of safety integrity in complex embedded systems. In the United States, NHTSA has the regulatory responsibility for performance standards for vehicle systems or sub-systems that address a specific type of safety risk. For AVs, NHTSA is focusing on developing functional safety requirements as well as potential liability requirements in the areas of diagnostics, prognostics and failure response (fail safe) mechanisms (NHTSA, 2014).
Techniek maakt voertuigen veiliger, Hoe kun je veiligheid aantonen als testen op de weg maar een beperkt aandeel metingen betreft?
Safety Benefits of Automated Vehicles: Extended Findings from Accident Research for Development, Validation and Testing
→ report goes into more detail
For testing methods in order to develop and validate safe automated vehicles with reasonable expenditure, the author recommends combining area-wide traffic, accident, weather, and vehicle operation data as well as traffic simulations. Based on these findings, a realistic evaluation of internationally and statistically relevant real world traffic scenarios as well as error processes and stochastic models can be analyzed (in combination with virtual tests in laboratories and driving simulators)to control critical driving situations in the future.
Self-Driving Regulation, Pro-Market Policies Key to Automated Vehicle Innovation
One important challenge, which is expected to be met by late 2014 or early 2015 , is providing sufficient evidence that road – tested autonomous vehicles are in fact safer than manually driven vehicles. As Bryant Walker Smith of Stanford Law School has noted, a high degree of statistical confidence must be reached in order for automakers and component developers to begin scaling up technology deployment beyond testing.
Google’s self- driving cars have logged over 500,000 miles on U.S. public roads to date. To demonstrate their safety over manually driven vehicles with 99 percent confidence, Google will need to log approximately an additional 200,000 miles of crash-free automated driving (see Table 2).
Introduction page 4 of document
Van de 33.000 verkeersslachtoffers die de VS jaarlijks betreurt, zijn er volgens deskundigen 22.000 te voorkomen als we de mens achter het stuur vandaan halen.
Point made by an online respondent of a survey:
“I am extremely concerned that proponents have little regard to or understanding of the level of reliability required to class any of these systems as safe . For example in regard to Google cars : ‘Ultimately, Google aims to provide a solution for the millions of car accidents that occur worldwide – 93 percent due to human error .’Statement is misleading/ wrong . Human factors contribute to 93 percent of crashes but many other factors also contribute. And the most responsible drivers cause a crash where someone is injured around once in 2,000,000 Miles. And public would expect autonomous cars to have a much lower rate-say once in 20,000,000miles.That requires a system that will not fail/malfunction more than once in ~ 80 vehicle lives or once in 1250 years of average driving.”
21.3 Requirements for a Test Concept
In order to discuss in the following section why full automation poses a particular challenge for safety validation, we will first describe the requirements for test concepts to assess safety. These are divided into effectiveness and efficiency criteria.
21.5.1 Validity of the current test concept for autonomous driving
At present, real driving is the most important method for the approval; the reason for this, in particular, is the validity combined with the justifiable economic overhead. However, along with the economic overhead, autonomous driving also presents a systematic challenge for the known methods. At present, real driving stands for driving in public road traffic with test drivers. The task of the test driver is to drive or supervise the vehicle in every situation in accordance with the task of the vehicle user. Transferred to autonomous driving, the use of a test driver in the driver’s seat would be non-real behavior of a user, as the user does not have to supervise the vehicle and the environment anymore and intervene.
Point 64, page 27 in report
Witnesses discussed the research evidence for the effectiveness of different systems. Professor Sampson said that it was very difficult to research which technologies were most effective in terms of reducing accidents, because of the difficulties in running controlled trials of different features, with sufficient numbers of vehicles. Professor Carsten explained that the key struggle was with the continual monitoring and evaluation of technology, and developing an understanding of how casualty rates were affected over time by different technologies.
He explained that while it was statistically possible to show the safety benefits arising from car impact regulations, it was “really hard” to do this in relation to other safety approaches.
AVs are capable of providing large amounts of data that could assist investigation in case of a crash. By recording the actions and forces involved in the minutes before and after a crash, they may help determine the cause of the crash and assist in resolving any liability dispute.
Point 33, page 15 in report
Telematics also known as ‘black boxes’ monitor the location of a driver and driving performance.
However, a safety prognosis of highly or fully automated vehicles depends on assumptions, as so far no series applications of such features exist. For testing methods in order to develop and validate safe automated vehicles with reasonable expenditure, the author recommends combining area-wide traffic, accident, weather, and vehicle operation data as well as traffic simulations.
Two questions discussed in paper:
–What significance do analyzes and findings from road-accident research hold for the introduction of automated vehicles?
–How can the potential safety benefits of automated vehicles be established?
The validity of accident data regarding potential safety benefits varies considerably depending on the collection method.
Hieronder een lijst met een aantal documenten die enkel betrekking hebben op het subdomein. Er zijn nog meer documenten gerelateerd aan dit subdomein maar deze zijn ook gelinkt aan andere subdomeinen. De volledige collectie beschikbare documenten in onze online bibliotheek (catalogus en dropbox)
|Titel/Title||Auteur(s)/Author(s)||Product van/Produced by||Opdrachtgever product/Product Requested by||Publicatiedatum/ Publication date|
|ROAD SAFETY WITH SELF-DRIVING VEHICLES: GENERAL LIMITATIONS AND ROAD SHARING WITH CONVENTIONAL VEHICLES||MICHAEL SIVAK BRANDON SCHOETTLE||The University of Michigan Transportation Research Institute Ann Arbor, Michigan 48109-2150 U.S.A.||The University of Michigan Transportation Research Institute Ann Arbor, Michigan 48109-2150 U.S.A.||31-1-2015|
|Accident rates of self-driving cars: A critique of the Sivak/Schoettle||Dr. Alexander Hars||BLOG||BLOG||23/1/2015|
|Motoring of the future||.||The Transport Committee is appointed by the House of Commons to examine the expenditure, administration, and policy of the Department for Transport and its Associate Public Bodies.||Ordered by the House of Commons||3-6-2015|
|Tomorrow’s Road Infrastructure for Automated Driving||Philippe Nitsche, MSc.||Austrian Institute of Technology (AIT)||Austrian Institute of Technology (AIT)||14/08/2014|
|Self-Driving Regulation, Pro-Market Policies Key to Automated Vehicle Innovation||Marc Scribner||the Competitive Enterprise Institute.||the Competitive Enterprise Institute.||23/04/2014|
|Future of automated Vehicles Build Environment Workshop||Steven E. Underwood, Ph.D.||Graham Institute, TRB, AUVSI, and SAE Survey Results||Graham Institute for Sustainability University of Michigan||29/07/2015|
|Safety Benefits of Automated Vehicles: Extended Findings from Accident Research for Development, Validation and Testing||Thomas Winkle||Department of Mechanical Engineering, Institute of Ergonomics, Technische Universität München||Springer Berlin Heidelberg||22/05/2016|
|The Release of Autonomous Vehicles||Walther Wachenfeld, Hermann Winner||Institute of Automotive Engineering, Technische Universität Darmstadt||Springer Berlin Heidelberg||22/05/2016|
|Do Autonomous Vehicles Learn?||Walther Wachenfeld, Hermann Winner||Institute of Automotive Engineering, Technische Universität Darmstadt||Springer Berlin Heidelberg||22/05/2016|