Unlocking the Smartphone's Sensors for Smart City Parking
Résumé
Studies have shown that drivers often spend over 20 min cruising for parking in city centers, accounting for as much as 30% of the traffic congestion. In response, cities like San Francisco have deployed systems capable of providing drivers real-time parking availability information. However, such systems rely on specialized infrastructure whose installation and maintenance costs in the tens of millions of dollars, unaffordable for many cities. We present SmartPark, a system for real-time parking information that relaxes the requirement for specialized infrastructure, relying instead on the smartphone’s sensors and the ubiquitous Wi-Fi and cellular infrastructure. To accomplish this, SmartPark addresses two major challenges, under the constraint of minimum impact on battery life: transportation mode detection and location matching. To minimize initial deployment cost and risk, SmartPark introduces an analytical approach for estimating parking availability even when only a small fraction of users adopt the application. We evaluate SmartPark using simulations and in the wild. Simulation results show that SmartPark, benefiting from as little as 20% adoption rate, can estimate parking availability with accuracy above 90%. Experimental results with the help of 12 volunteers show that SmartPark detects unparking events 97% of the time while triggering zero false positives.