In an extreme case of WTFism, while I was writing this post, news broke of the first fatal collision between an Uber vehicle in full-auto mode and a bicyclist near ASU in Tempe, AZ. The Uber vehicle had a safety driver in place. On the face of it, the accident seems to illustrate humans' poor ability to monitor fully-autonomous systems. Would this accident have happened if Uber had been developing an Augmented, as opposed to Autonomous vehicle? Highly doubtful. (And, by the way, Uber of all companies, is the one that should be working on Augmented Driving tech, which would make all its human drivers, who get no additional training and little vetting, safer.)
Over the last week or two, I’ve been following the auto writer and general iconoclast Alex Roy’s posts on the topics of a ‘Human Driver's Manifesto’ and on the auto industry’s misguided effort to usher in higher levels of automation before they’re really ready for prime time.
Roy interviewed Captain Chesley ‘Sully’ Sullenberger, who is a member of the Department of Transportation’s Adivisory Committee on Automation in Transportation, and got an earful on the fundamental problem with SAE Level 2 & 3 automation: the need for occasional transition back to human operation.
As a motorcyclist, I have a low opinion of most drivers’ skills. But as has been demonstrated, the transition from inattention to attention takes a long time. Automakers’ half-hearted attempts to build in systems that encourage drivers to keep a hand on the steering wheel, or that monitor drivers’ eye movements don’t really address the larger issue, which is that it’s not a matter of whether the driver has his hands on the wheel (at least occasionally) or has his eyes on the road, it’s whether he has his mind on the road.
As Roy points out, the big investments (and lobbying efforts) in automated driving have often been made by companies like Uber, Google, and Apple, which have a financial incentive to remove steering wheels altogether from cars, ASAP. Phantom Auto is an example of a company that provides remote driving capability for otherwise-autonomous vehicles when they meet a situation their algorithms can’t resolve. Needless to say, Phantom’s human operators at remote controls are not any better than drivers in situ when it comes to suddenly assessing hazardous conditions.
Sullenberger and Roy discussed the advantages of systems used in commercial aviation, which keep pilots engaged, and provide improved flight information while preventing (or at least discouraging) pilots from providing control inputs that are outside their aircraft’s performance envelope.
Roy is a fan of ‘Augmented Driving’—basically redirecting all the efforts that go into lower levels of automation, to keep drivers engaged while improving their situational awareness, rather than lulling them into complacency, and to provide feedback that would discourage or prevent behavior that would cause accidents.
As a group, motorcyclists are skeptical of the ‘safety’ of autonomous cars and trucks, and vehemently opposed to the very idea of autonomous motorcycles (although Yamaha has built a robot rider, and BMW’s R&D chief once admitted to me that BMW had built a self-riding bike for research purposes.)
Augmented Riding would be a far more acceptable proposition to most motorcyclists, and a lot of the technology is already available. The best available traction control and anti-lock braking systems which use IMU data are already impressive; protective clothing that deploys airbags when a crash is detected—often even before an impact—is available; Bosch has a blind-spot warning system; helmets with HUD systems already exist and will soon be more common and more useful.
An integrated approach, even to just the existing technology available to motorcyclists, would improve motorcycle safety. Layering in automotive style FCW systems and aircraft style haptic feedback would only involve repackaging existing tech (which, I realize, is not that easy, but still...) Soon, hyper-accurate mapping, and the kind of networked connectivity promised by 5G evangelists, could cut our currently dramatic rates of death and injury, without taking any of the fun out of riding.
As a motorcyclist, I hate the idea of people cruising along, inattentive, while their Level 2 (or, “it’s Level 3 but we’re not calling it that for liability reasons”) system drives their car. I would prefer drivers to have augmented driving tech at their disposal, that will make both of us safer. As a motorcyclist, I’m dead set against the Phantom Auto solution where the same ‘transition time’ lag will occur, but thousands of miles away. As a motorcyclist, I’m uninterested in automated riding, and doubt that there will ever be an effective Phantom Moto company ready to take over the control of my motorcycle.
However as a motorcyclist, I love the idea of Augmented Riding, and I am sure it will be easy to sell. Bosch? Connected Motorcycle Consortium? Let’s not just blindly ride down the road being mapped by Uber, et al.