Today I was awaiting a green light in traffic when my Valentine One started bleeping crazily, in the way it does when it detects a police laser. Obviously I wasn’t moving, so I wasn’t concerned about incurring any laser-based citations, but I began looking around the intersection for a Metro cruiser.None of the city’s finest were visible. But the V1 kept insisting there was a laser on us. Odd, I thought–but I have 100% faith in that little black box, so my mind kept churning, wondering what it could be. As the spotlight went green and I started underway toward my destination again, the warning stopped, and the proverbial light bulb appeared above my head. A new Volvo S60 had been behind me at the light, and was now passing me on my left. The Volvo’s infrared collision-avoidance system was what had antagonized my little electronic friend.
I became even more certain that the Volvo was indeed the culprit just a few hours later, when the exact same thing happened–new Volvo (this time an XC60) stops behind me at a light, V1 goes crazy. Pull away from the light and put some distance between V1 and Volvo, and V1 settles down.
So how about that? Not only does Volvo’s collision-avoidance system not work (see the video below), it also sets off laser warnings in detectors in front of it.
Let me get some more information about that for you…
Admittedly in my case the laser warning was triggered only when the Volvos were very close to my rear bumper, but this wasn’t exactly a scientific test. What if, say, in heavy, fast-moving traffic, a Volvo sets off a V1 in a car in front of it, causing the already task-saturated driver to instinctively brake, and the Volvo’s system/driver fails in a manner similar to the way it did in the video above? Now you have an accident caused by a system that is supposed to prevent one.
In the October 2011 issue of Car and Driver
, Editor-at-Large John Phillips and former senior editor Phil Berg attempted to drive three different cars on interstates at interstate speeds while essentially blindfolded with their noble intent being to test the cars’ advertised ability to automatically maintain their lane. You can read the article here
, but it boils down to this: while the hardware does
exist to make auto-lane-keeping possible, the software is not good enough to make the systems safe/effective.
In my opinion, there are simply too many variables for lane-keeping to be left to a computer–it’s just like trying to ask a computer to forecast the weather more than a day or two in advance. For example, is there snow on the road, or just off of it? What about gravel or sand or leaves? Does the road have lines painted on it or not? How does sudden bright sun glare or sudden deep shade affect the system? What about rain/spray from vehicles in front? What if the pavement is cracked or sealed or resurfaced or alternating between asphalt and concrete? All of these things matter, and it would be next to impossible to engineer software to take every conceivable situation into account.
“The 1%,” I think, dream of the day that they can sell autonomous vehicles to the masses. On that day, we poor hapless souls will suddenly have oodles more free time as we are whisked from place to place to rack up cell phone data overages while surfing Amazon to find more useless widgets to spend money on. They’ll make so much money when this plan finally comes together that you know what, even if they have to settle a few autono-mobile wrongful death cases out of court, it’ll still be worth it, WELL worth it, to them.
So my advice to you, dear reader, is this: learn to drive, and learn well. Do an autocross and/or get some track time so you know your vehicle’s limits and how it behaves when it goes beyond them. Teach those dearest to you all that you possibly can. Never drive while distracted, talking on a phone, texting, or watching anything other than the road. And never buy a vehicle (or pay for an option) that tries to do the driving for you, because no one cares more about what’s in your car than you do.