New Car Anti-Theft Device Profiles Your Rear End

Hugh Pickens writes "A car-seat identifier developed at Japan's Advanced Institute of Industrial Technology by Associate Professor Shigeomi Koshimizu can recognize a person by his or her rear end with 98 percent accuracy when the person takes a seat in his car. The bucket seat's lower section is lined with 360 pressure sensors that measure pressure on a scale from 0 to 256, sending information to a laptop, which aggregates the information, generates the key data and produces a precise map of the seated person's rear profile. Researchers say traditional biometric techniques such as iris scanners and fingerprint readers cause stress to people undergoing identity checks, while the simple act of getting seated carries less psychological baggage. Koshimizu wants to see his work available commercially as an anti-theft product in two to three years if automakers agree to collaborate. He sees possibilities of this device being used beyond auto-theft identity protection to a device for security identification in office settings, where users log on to their PCs as they sit down."

Comments

Detect differences between full and empty load?

[InterestingFella]

Some people apply their rear pressure differently based on if their rear load is full, empty or something in between. Not only does your overall weight change, but also the formation of rear pressing against the seat will be different, especially depending on your nutrition and different days. Is it going to be able to detect such load changes without many problems? Obviously there needs to be some kind of threshold, but if your rear pressure varies a lot the device could even lock you out from using your car.

Drive 49 out of 50 days!?

[Mitreya]

Oh yeah, who wants to drive to and from work EVERY day. With 2% failure rate, you can expect 8 failures-to-drive a year.
I mean, seriously, 98% is a good rate, but putting it on something one uses every day is just an accident waiting to happen.
Plus you won't be able to lend your car easily.

Re:Drive 49 out of 50 days!?

[vlm]

Assuming a random statistical distribution of failure, and a short "timeout" between tests, the overall system success rate of a 98% success on first try followed by 98% on second try means you'll only have to try a third 98% success trial something around once for each owner of the car, assuming you own a car "about 4 years" or so, and assuming I did the math in my head correctly for 2500 days. I figure every 125000 times you boot up the car, you'll need a 4th try, that's booting up the car twice a day for 171 years. So "lock out after 6 attempts" seems safe enough to only happen accidentally a couple times in the lifetime of the product run?

Biggest problem is going to be embarrassment at having to get out of the car and try it again, if people see you doing that they're going to make all kinds of interesting assumptions about what happened to your rear last night, rough time doing doggie style or things are still too stretched out or sore back there today or whatever, so I'm thinking women would be waaaaaaay too embarrassed to buy one of these systems, even if it only happens once in a while.

"added" benefit:

[tchdab1]

you're walking until you loose those extra pounds.