ARCADE Project: Shared driving responsibility in highly automated vehicles
Jo-Ann Pattinson, Postdoctoral Research Fellow at the Institute for Transport Studies, and responsible of legal issues in autonomous vehicles, shares with ARCADE some main issues in the domain of connected and automated vehicle.
The legal transfer of risk and responsibility between driver and vehicle is more complex than a delineation between ‘the car is in control’ or ‘the driver is in control’, and this is an issue which must be addressed before highly automated vehicles take to our roads. International law will require Highly Automated Vehicles to operate a digital interface to provide information regarding the operation of the vehicle, and driver responsibilities. This interface is also likely to request confirmation that the driver has understood the information provided, by seeking confirmation of a driver’s agreement or consent. It is at this point the legal framework is prone to fall apart, and represents a major barrier to manufacturers achieving real-world roll-out of automated vehicles. A problematic digital interface may look something like this:
There will be no difficulty with convincing drivers to press ‘Confirm’. Human beings are almost programmed to do this. Unfortunately, human beings are also programmed to ignore or skip over the text which immediately proceeds ‘Confirm’. When that text contains vital information relevant to their safety, the safety of their passengers and all other road users, this is a problem. It becomes a potentially tragic problem when the driver causes an accident after ignoring vital information. It is a fundamental problem for manufacturers when there is existing and compelling research to support the following:
When accessing a technical service:
I. People do not read or efficiently process information presented to them in typical digital interfaces
II. People will readily agree to terms and conditions in order to quickly access the service
The act of pressing ‘Confirm’ communicates a legal consent to take on responsibility and accept a certain degree of risk. However, if we are aware the driver is likely to do this in ignorance, the validity of this consent is in serious doubt. Drivers have a limited capacity to understand and process information and instructions. An interface which overburdens the driver may fail in law to transfer driving responsibility.
If drivers ignore important information, they are more likely to have an accident. In addition to the implications for safety, this is likely to lead to complex litigation, as the legality of the transfer between driver and vehicle will be called into question. This will have negative implications for user acceptance, public confidence and insurance regimes. The use of a digital interface which does not give sufficient agency to drivers has repercussions for corporate responsibility, liability and ethical responsibility.
Addressing the problem
Further research is required to study the interactions between the driver and the vehicle, including how well drivers absorb information from different and assess how well they are able to apply that information to a takeover situation. This data will be crucial for future interface design. This is an urgent step to take before highly automated vehicles are deployed on public roads. The Institute for Transport Studies at the University of Leeds is committed to data-driven research which will address knowledge gap, and formulate solutions to this problem.
Source: ARCADE Project
In: CLEPA News, Connectivity & Automation, Safety