ParkNet: Drive-by Sensing of Road-Side Parking Statistics
Suhas Mathur, Tong Jin, Nikhil Kasturirangan, Janani Chandrashekharan,
Wenzhi Xue, Marco Gruteser, Wade Trappe
WINLAB, Rutgers University, 671 Route 1 South, North Brunswick, NJ, USA
Using sensors placed on cars, a system was created to collect data on parking spaces in an urban area. A key motivation for this work was a study that showed billions of dollars lost annually in time and gas due to traffic congestion. The idea of tracking parking spaces has generally been held to immobile sensors placed on individual parking spaces. This system shows that results similar to those from immobile sensors can be achieved with less hardware than current systems.
The study shows that their system worked well (95% accuracy in space counting, 90% in determining location) under the conditions it was placed in. Interesting points covered, are the problems they faced in determining how to handle the task of accurately collecting data from a moving object, specifically:
1. Providing accurate readings for unslotted parking areas. The implementation presented uses as estimation technique for areas with unmarked spots.
2. ParkNet cannot handle data that comes from anything but the rightmost lane. A system for gathering input from a more chaotic environment would be needed.
3. Not all vehicles were able to power the system without interfering with sensors.