Lunch today with Jeffrey Brantingham, an anthropologist/archaeologist interested in agent-based simulation. We talked about how to model the spatial distribution of auto theft.
Jeff mentioned that the real hot spots for auto theft are the huge parking lots at malls and park-and-drive stations, rather than curbside. That gave me an idea that was new to Jeff. Maybe some reader can tell me whether it’s in use or why it’s infeasible.
Imagine having a sensor under each parking space that notices when a car pulls in, and a keypad on a post beside each space. As you get out of your car, you use the keypad to enter a 4-digit code you make up on the spot. Before you get back in to drive away, you enter the same code.
If, when a car pulls out, the right code hasn’t been entered, information about the car (e.g., from cameras on poles) is relayed to the exit points from the lot, and the car is stopped and checked on the way out.
The details of how to identify suspect cars and what to do about them would have to depend on the details of the lot in question. Whether the program was worthwhile would depend on those details, and on the frequency of theft. But this sounds as if it might be worth some actual design work.
A reader comments:
Interesting idea. In Rockville, the new parking meters on the street have (IR, motion?) sensors that tell if a space is occupied. The meter is cabled to City Hall. When the time runs out, a meter maid is directed to the spot to issue a ticket.
The purpose is revenue, but the technology is in place. The meter sensor means no having to jack hammer the curb side for a pressure sensor. The cables went in before the sidewalks were poured. The meters have some input buttons for credit card choices (credit or debit? Receipt?) that could be used for security information. However, detection of auto theft would not produce revenue for the city. What’s its motivation?
Another points out that the problem could be solved at the car, rather than the parking space:
Once you enter a portable keypad into the equation, you don’t have to spend a fortune putting sensors in parking spaces and whatnot to end auto theft. Most (all?) new cars are controlled by several computers. The ECU, for example, controls the fuel flow and the timing of my Miata. Disable the ECU, and the car won’t go, no matter how good the thief, unless they brought their own ECU and they are willing to replace the ECU right there in the Walmart parking lot. Building a ECU with an encrypted, wireless communication port would be simple. Building a key chain that talked to it (commanding it to go on and off) would also be easy. All of the components, hardware and software, are off the shelf. In short, Detroit could make every car coming off the line theft-proof for perhaps $50.
The mystery isn’t how to do it. The mystery is why they don’t.
The key to the mystery, it seems to me, is that sales of replacement parts to people whose cars have been damaged in the course of being stolen are a non-trivial profit source for the auto companies.