PFI-TT: Smart City Curbside Parking Management

University Of Washington, Seattle WA

Investigators

Abstract

The broader impact/commercial potential of this Partnerships for Innovation - Technology Translation (PFI-TT) project in benefitting cities and drivers by enhancing scientific and technological understanding of smart mobility. The cities will benefit from accurate curbside parking management system, which can help them optimize curb lane space, parking planning, and parking pricing. The curbside parking inventory database will also allow map and navigation service providers to lessen the amount of time that vehicles spend driving around looking for parking spaces. With the parking search navigation solution, drivers will spend less time on curbside parking search, avoid illegal parking, save gas, and enjoy peace of mind during parking search. The cities will further benefit from reduced traffic congestion and pollution. The proposed research will foster new research and education opportunities. Students will be trained to conduct machine learning (ML) and IoT research in smart mobility. This project is being undertaken by a team consisting primarily of female researchers, female entrepreneurs, and minorities. The proposed project will build a smart curbside parking management system by jointly utilizing the ML and IoT technologies. The goals of the proposed applied research are to develop technologies to correctly read and locate the curbside parking signs from the street videos taken by dashcams and to detect and track the parked vehicles and open parking spaces from roadside deployed cameras. The proposed research will address the technical hurdles in (i) accurately reading curbside parking signs from the street videos taken in different cities by fine tuning the team’s AI models for sign detection, sign tracking, and sign text recognition; (ii) accurately locating the curbside parking signs from the street videos by fusing the results/data from video odometry technology, map, and video GPS footage; (iii) detecting open parking spots by jointly utilizing the team’s curbside parking inventory database and the AI models for lane detection and vehicle detection; and (iv) tracking curbside parking occupancy by studying ML-based parked vehicle detection and tracking technologies. The anticipated technical results include a series of AI models and algorithms, a curbside parking rule database, and a system for curbside parking inventory management. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.