Self-driving cars will increase the need for driver monitoring, not reduce it

Face and voice analysis technology deployment in the automotive cabin is growing fast, and the first use for this that people think of is driver safety monitoring. BLUESKEYE AI is one of only a very few Tier 2 suppliers of this technology to Tier 1 companies, who in turn sell this to the companies who actually assemble cars. And improving driver safety is certainly one way that Blueskeye’s face and voice behaviour analysis can be used, although Blueskeye’s real aim is to improve people’s long-term health and wellbeing. But hang on, aren’t we all supposed to stop driving soon? So is there any future in driver safety monitoring?

That’s the second thing that people think when they hear about driver safety monitoring. Surely, with the advent of self-driving cars, there will be no need for driver monitoring at all. Besides convenience, the strongest case for self-driving cars is that they’re safer than humans because they make fewer mistakes and don’t get tired or distracted. So surely, if cars can do all the driving, that makes driver monitoring pointless? An oxymoron, even, as it’s not the driver that’s driving?

And perhaps in a not too distant future, say in the year 2035, this may actually be the case. In the meantime, the number of journeys that will be guaranteed 100% driven by level 5 autonomy will be very low indeed (defined as full automation, where the car can drive itself in all conditions without human intervention). All automotive Tier 1s and OEMs we work with plan for a whole period of time where the human driver will have to take over regularly, if hopefully less and less frequently.

And then there’s the rather obvious issue that many people simply love to drive, and will insist that they can drive themselves, even if they have to pay extra for that. Whole generations will have to live out their lives before that desire so strongly linked with freedom will cease to put demands on car design.

So, in all reality, we are looking for a whole period of time, decades really, where there will be frequent switching between who’s driving: you or the AI in your car. And this is a good thing, really. It does not mean that the safety gain is lost. Because level 4 autonomy capabilities mean that when you are temporarily distracted, or incapacitated for a longer time due to a medical emergency, the AI can take over. And in many cases where people currently drive when they know they shouldn’t, they can let the AI drive (say, after a few drinks, or when really tired but you have to be home in time to put the kids to bed).

Self-driving capabilities are steadily improving, and that’s what reaches the headlines. What’s missing though is the ability of the car to know when it’s safe to hand over to a human, when it should wrest control from the human for their safety, and in particular, how to measure what the (mental) state of the driver is so that it can make these decisions. And this is because this is absolutely state of the art technology, and by no means complete. It also happens to be exactly what BLUESKEYE AI focuses on. Already measuring driver emotion, drowsiness, and attention, Blueskeye is the perfect company to transfer their capabilities of objectively measuring medical conditions currently used in the clinical domain to be used in the automotive domain.

EURO NCAP currently spells out the need to measure alcohol intoxication, but it vaguer about what medical conditions should be measured to assess whether a person can safely drive, and it only hints at the possibility of measuring cognitive load and capacity. My prediction is that this will rapidly see some very exciting additions and clarifications.

The automotive industry is a complex ecosystem, and the introduction of new technologies is shaking up how OEMs, Tier 1s and Tier2s work together. And in order to make this vision of safe driving possible, we must work together.

BLUESKEYE AI is ready to play our part. If you’re interested in learning more, get in touch!