Real world test of City Safety
I know Bill, and he’s far from an “inept driver” as one commenter alludes to regarding this “2011 Volvo S60: A drivers car that makes better drivers” Boston.com article. Stuff happens, and in this case we saved all the hassles associated with a minor accident.
Here is Clifford’s counterpoint to Bill’s piece.
To an extent, Clifford is correct: electronic systems are not perfect. But if we get it right 90% of the time – heck, even 50% of the time – isn’t it worth using technology that might save a life? Could we have waited couple more years to get an absolutely 100% perfect system? Sure, but what about those that we could have saved? Tough call for any company to make. I think we made the right decision. What Clifford missed is that City Safety and Pedestrian Detection with Full Auto Braking can be switched Off, with default On when car first started by driver. Clifford mentioned that the system could not detect our dummy during testing. Well, the system actually worked, it did not detect a human. There are vast differences between a dummy stuffed with foam and human stuffed with carbon and water.
dan
2 Comments
Thanks for the links Dan. I agree with your sentiment that it’s better to release a safety feature earlier rather than later, as long as it /never/ interferes with how the car should work without it. In this case, as long as it never brakes the car dangerously without a real reason to do so, etc. I don’t think I’ve heard one report of anything like this happening, despite Clifford’s experience with the traffic barriers.
It’s unfortunate that some of the “failed” tests are garnering so much publicity. I have a feeling that if those situations were examined more closely you’d find that the driver actually either chickened out or sabotaged the test or accidentally ruined the test by adding driver input (steering, brakes, accelerator) at the last minute turning the systems off. It seems to me it would be very difficult to build software that could tell the difference between an accidental input and a purposeful input leading up to an accident.
Then there’s the difficulty in accurately modeling a human to the radar/etc detection systems without actually using a real human….
I applaud Volvo for bringing this technology when and how it did. There are already announcements of copy-cat technologies coming and soon Volvo won’t be able to claim they are the only ones with this type of safety tech. Being first-to-market is worth something too.
Somehow I have the suspicion this tech will gain more popularity when it comes from a German maker though… too bad.
Hi Joel,
In all the ‘failed’ tests I’ve sat in on, being in a passenger, the driver tried to override the system. In reality, as Bill experienced, the typically doesn’t happen in real life. The system arms itself, and then takes action. Human could never be as fast in this kind of situation.
There are two ways to handle bad publicity, deny/ignore or just be up front about it. I’ve never seen us hide behind issues, much better to get it out and let people decide on their own. I don’t think it hurt us in those tests that went wrong. The one where an S60 went into back of a truck, an engineer was charging its battery and didn’t check system pre-production software. It happened.
Yes, others will follow, just like with side impact air bags, stronger roofs, and better energy absorbing technologies. What we want to do is force others to join what we are doing. Someone has to lead, might as well be bunch of Swedish engineers.
Kind regards,
dan