Any city that has purchased and installed wireless parking sensors may be the latest victim of rapid technological change, because a new app called PocketParker may make them obsolete. The app is still in development, but a paper about its progress was presented last week at the UbiComp computing conference in Seattle.
Cruising for parking could become a thing of the past with the new PocketParker app. From Torquay Palms.
Researchers from the State University of New York in Buffalo described the technology on which the app is founded as “pocketsourcing” — smartphones morph into passive sensors that track the location and movements of users who install PocketParker, which currently only addresses parking lots versus on-street parking.
It works with parking lot data culled from OpenStreetMap to determine the number of spaces in a lot based on its dimensions. PocketParker then uses a smartphone’s accelerometer to identify a user’s location. It next analyzes a user’s movements to decide whether or not she’s searching for parking. For example, if the user drives slowly through a parking lot without stopping, then that suggests the lot is full. If the user exhibits movements typical of walking and then quickly accelerates and leaves the lot, she likely got into her car and left the lot.
The researchers tested the app with 105 smartphone users in Buffalo over six weeks for a total of 10,827 car arrivals and departures. They checked their work against cameras they installed at participating lots and discovered that they correctly predicted how many parking spaces were open 19 out of 20 times. They were also able to forecast the number of spots at the lot to within 6 percent of the actual number.
The figures suggest that PocketParker could offer owners of parking lots a cheaper way of tracking capacity. For instance, it runs between $200 and $250 just to install wireless parking sensors and $3 to $10 monthly for the networking and data services, says Worldsensing, which manufactures and installs sensors.
PocketParker needs none of that infrastructure but does have drawbacks, the biggest being that it can’t account for motorists who don’t download the app. It may be able to tell when a space opens up, but if there are several drivers in the lot when that happens and none have PocketParker on their phones, then the app doesn’t know that they’re at the lot or if they’ve taken the available parking space.
“Until you have enough people using it, apps like this tend not to work well,” said Geoffrey Challen, one of the paper’s coauthors and a computer science professor at SUNY Buffalo.
One solution, he says, is linking the app to a mapping app the way Google Maps incorporates traffic data. If PocketParker were similarly integrated, then it would be able to quickly gather enough information to make better predictions about parking availability.
“Our goal is to prevent people from circling,” said Challen.
Related Posts
