August 01, 2011
Brain Reading Speeds Car Braking
Drive by thinking. Brain reading speeds car braking.
This is what German researchers have successfully simulated, as reported in the Journal of Neural Engineering. With electrodes attached to the scalps and right legs of drivers in a driving simulator, they used both electroencephalography (EEG) and electromyography (EMG) respectively to detect the intent to brake. These electrical signals were seen 130 milliseconds before drivers actually hit the brakes—enough time to reduce the braking distance by nearly four meters.
Problem: Drivers will start slightly delaying their decision to apply the brakes as the brakes respond more quickly. One can imagine this system would still improve safety when a kid runs out in the street. But a Wired piece finds researchers on semi-autonomous vehicles are concerned that improved computer assists for drivers make drivers more complacent and therefore more dangerous.
Clifford Nass, a communications professor at Stanford University who studies multitasking, put it more bluntly.
“People are always happy to be lazy, and it’s sort of a rule of safety design,” he said. “So if you give people the slightest opportunity to be lazy, they’ll take to it with great gusto and joy.”
This is especially true for frequent multitaskers — and most apparent with young people, whose brains have developed to crave new information.
The challenge becomes how to make all the computer driving assists really deliver a net improvement in driver performance. The article discusses how to keep the driver engaged even as computers do more of the driving work.
Car fatalities have been declining pretty steadily haven't they? So while some of the gains due to safety improvements may be given back by increased recklessness it's clear not all of them are.
IOW, risk homeostasis. DJ, I suppose we can distinguish between safety features which are evident to the driver, and which will cause the driver to adjust his behavior in a more risky direction, and safety features which can't be seen, and thus don't effect driver behavior. You wouldn't really expect a driver to factor in an enlarged crush zone, the way they might factor in belting themselves in place. And, of course, there are all those contributions to safety made by traffic engineers: Divided roads, bridge abutments guarded, grooved pavement.
Unfortunately, bypassing the peripheral nervous system and gross movement of limbs in activating car controls is going to be very noticeable in terms of car behavior. The driver will indeed simply adjust his driving habits to take the car on the same less safe than some regulator would prefer trajectory as they would have prior to the technological advance.
No, the general lesson should be to increase safety in ways the driver can't perceive, and thus compensate for. I'm still waiting for them to take the obvious step, and put air bags on the *outside* of the car, activated by sensors slightly in advance of a collision. Might even help pedestrians, which no amount of safety features for the driver will do.
It's time to allow humans the option of getting out of the loop altogether. The human reaction time doesn't matter if no human is controlling the car, and the risk homeostatis issue goes away (so long as speed and convenience are maintained).
Hackers can't wait for this.
Massively multiplayer online bumper cars.
I doubt people will have much luck hacking a LIDAR or a rail-based TDR system.