ATTI’s legal specialist questions the relevance of traditional cabin indications in such vehicles.
‘If you see me doing something dumb, say something.’ This is a well-known mantra in the airline industry, which encapsulates cockpit resource management (CRM). An equivalent system has been used in the automotive industry for decades whereby, if the vehicle notices the driver doing something dumb, it will ‘say something’ in the form of audible and visual warnings. This begs the question, what will happen in Levels 4 and 5 of autonomy where there is no pilot, driver or other human in charge of the driving tasks? If the car sees its own internal logic doing something dumb, will it say something? If so, to whom? More to the point, if the vehicle is going to have this dialog with itself, how will auto makers develop and test the necessary technical solutions?
Let’s start with the present-day approach to these alerts. Systems in modern cars discriminate between visual and audible information signals. If, say, the direction indicators are active, this will prompt a soft informational sound. If danger is detected, the audible signal will be a warning chime. There is robust and mature science around warning chimes to alert the driver without distraction from the driving task, including tone, modulation, cadence and of course volume. These well-established systems are subject
to evaluation, both to validate their compliance with the design intent, and to meet human factors goals. Finally, human user groups confirm the effect on the human ear and brain. This has been state-of-the-art technology for some time. The industry has reflected on and reached strategic solutions for attributes such
as automatic lane keeping. Handover strategies had to be developed and tested for this task, including chimes, haptics and visual displays.
Now apply the same challenge to Levels 4 and 5 of autonomous driving, where the human has no driving responsibility. How this will operate is anyone’s guess, and my own is that there will be very few audible warnings – maybe none at all. Is there any point in the car making a reassuring audible tone to tell the passenger that the car has indicated, for example? Does the passenger really need to be disturbed from watching the latest Harry Potter remake by this noise? If not, why develop and test a signal at all?
What about averting potential danger? If the vehicle sees itself ‘doing something dumb’, then of course it
will have to tell itself and correct itself. But why alert the passenger with a chime? This just leaves one variable. What if the passenger does something dumb? Suppose the car recognizes the passenger doing something unsafe, such as unbuckling their seatbelt while being driven. What happens then? Automotive manufacturers will have to decide on their strategies for this. One response might be to bring the vehicle to a safe stop, but car makers might conclude that this introduces more risk than it solves. For all these reasons, audible signal systems and the associated tests might be obsolete by the time Level 4 and 5 autonomy are introduced. After all, when it comes to humans in vehicles, there is an abiding truth: Even a foreseeable misuse is still a misuse.