Smart Parking with Fine-Grained Localization and User Status Sensing Based on Edge Computing

Parking at an affordable place is the precedent task for all activities of the everyday life in urban environments such as shopping, working, exercising, etc. So, it is the most common and essential requirement of all users in a car park to fast search a preferred parking spot closely associated their current intent. Although modern parking lots have installed the sensing and display systems to inform drivers on the availability of parking areas, such systems are unable to tell drivers exact parking spots and make any recommendation to improve the traffic conditions and driver experiences. In this paper, a novel analytic- based smart parking system clustering Internet of Things, smart mobile devices and edge computing is proposed. This novel parking system aims at providing customized parking experience to users through highly accurate positioning and user status detection which are achieved by joint mobile sensing-machine learning based analytics as the edge intelligence. Based on the proof-of-concept implementation, the proposed scheme can achieve 99.1% positioning accuracy of a parking spot; in terms of user status sensing, especially getting in a car and out of a car, detection accuracy shows 96%; finally, it shows much shorter service consumption time of 15.6 times than the legacy approach.