New cars, new toys

I’ve been playing with new cars again today – one of the few perks of the job which crop up from time to time. I can’t give too much away as it was a closed meeting, but most of the cars and their associated technologies are already on the market so I can talk in general terms. Suffice to say I got to drive several cars much more valuable than my own (some would even pay off my mortgage…), and each jam-packed with active safety gadgets. In fact, that was one of my first observations – the sheer amount of technology and computing power now being rammed into cars, with radars, cameras and sensors literally everywhere.

Anyway, that’s not what I really wanted to talk about. I’ve aired my thoughts about vehicle automation before on this blog, so I won’t go over old ground, but there was a couple of new observations I wanted to pick up on.

One or two of the vehicle manufacturers are using speed sign recognition cameras, sometimes coupled with a GPS map database, to display current speed limit information inside the car. I can only think this is a Good Thing – I don’t have evidence, but to my mind many speeding offences aren’t for lack of restraint, rather are to do with drivers missing or forgetting the last sign they passed. Sometimes the road environment even tries to trip us up, with outdated laws preventing the use of repeater signs for 30mph limits, for instance. I for one would really like one of these cars with a speed limit aide-memoire, and would much prefer that over an ISA system which just stops me from speeding. Don’t get me wrong – I’m not a speeder, it just makes me uncomfortable to take control from the driver in that way. And this is a personal and professional opinion, as I’ve written about in this paper.

Which brings me onto my next point – who’s in charge? The same paper talks about philosophies of ‘hard’ vs. ‘soft’ automation – basically regarding who has the last power of veto, the computer (hard) or the human (soft). We’ve seen aircraft manufacturers take different approaches based on these philosophies, and it’s now fascinating to see vehicle manufacturers doing the same thing. My view is the same as it’s always been – we should support drivers, not replace them. But I qualify that in the paper by talking about ‘below the line’ vehicle operations (such as ABS, ESC – and I could now include last-minute collision mitigation in that) which can be automated since they don’t consciously affect the driver’s task.

It seems that many of these technologies are going down well with the customers. My concern is whether that translates into better (safer) driving (for which we don’t have the data yet), and if there are any longer-term behavioural compensations that might cancel out any beneficial effects.

One thing I certainly didn’t expect from the day, though, was for the cars to be groping me – one of them had a seat massager (for passengers), while another had a ‘dynamic seat bolster’ to stop you sliding about in corners. Definitely a new experience for me…

An autopian future?

I don’t feel right snaffling someone else’s blog posts, but I just had to pick up on this gem from Tom Vanderbilt, as it’s so close to my own research interests. (Tom’s is an excellent blog for driving buffs, too, and he’s always got something interesting.)

The ‘Autonomobile’ from Mike and Maaike is a design exercise to come up with a driverless car. This kind of thing comes around from time to time, and the Autonomobile is reminiscent of Concept 2096 (not a great link, but the only one I could find with an image) – another one developed to mark 100 years of the automobile, and an early inspiration in my PhD work. Interestingly, as I remarked on Tom’s blog, these are very similar but the anticipated timeframe is coming down – by 50 years, as the Autonomobile is apparently set for 2040.

I’m less convinced. Yes, these guys are right that technically speaking, we’re not far off. In controlled, closed-loop systems it is possible – see the Heathrow Personal Rapid Transit on their website, which we’ve been looking at lately. But there’s a bunch more human factors which mean that last 20 per cent (or 10, or 5) will be the real struggle. Mainly it’s the reliability issue – until the systems become so good that we can literally sit back and read the paper, there’ll always be the expectation of a human being there to supervise and save the day. And if there’s one fundamental which has come out of all the research on automation – vehicle or otherwise – it’s that humans don’t make good supervisors. We’re meant to do, not watch.

Besides, as others have pointed out, such ‘fully automated’ cars already exist to let us work, socialise etc. during the drive – they’re called taxis. And that’s the other side of the human factors equation – the social acceptability of automated cars. For one, a lot of people drive because they enjoy it. But the harder one to crack will be trust in these systems – will people ever really let go and switch off? And what happens if(when) there’s an accident – who’s to blame?