The Economics of Smart Parking Leveraging IoT and AI While Being Cost-Effective

With the advent of car-centric city planning, we find ourselves in a dire situation regarding parking availability. This is causing a tangible economic loss to multiple parties: the businesses losing out on revenue due to people avoiding crowded spaces and the driver of the car due to waste of time circling for a parking spot. To alleviate this issue, researchers are now encompassing parking as part of a wider Intelligent Transportation System, a topic for which research has been ongoing for more than 30 years. In particular, the Smart Parking System component already has real-world applications in large US cities such as San Francisco with their own SFPark, an award-winning program, and ongoing research in the integration of Internet of Things (IoT) concepts to accelerate the availability of such systems. This paper aims to survey existing systems, both real-world applications and research proposals, while also pointing out some of their shortcomings. Iterating on this existing knowledge, the proposed Smart Parking System that integrates IoT and AI could generate an estimated $7 Million daily from worker productivity alone, along with increased revenue generated from parking and overall better quality of life for the citizens due to lower congestion levels, saving an estimated 32.25 minutes of circling daily per parking space. All of this can be achieved while also being cost effective regarding the implementation and running costs, where we will be doing a cost analysis for the different sensing scenarios that a parking spot might require such as camera recognition using Computer Vision, overhead Time of Flight sensors, induction loop or even novel solutions such as bollards with embedded infrared sensors. This system is leveraging state-of-the-art Internet of Things components while also performing most of the calculations on the edge of the network, which improves overall security and privacy.