Using Edge Devices and Machine Learning for Controlled Access to a Smart Campus

University campuses can employ Fourth Industrial Revolution technologies to improve traffic flow and increase security. We provide one such solution that uses inexpensive edge devices (in our case, the Jetson Nano device from NVIDIA) to scan the license plates of cars entering a closed campus and to automatically determine whether or not entry should be allowed. Our solution eliminates the need for a guard to perform this check and speeds up the verification process, ultimately reducing the entry delay of incoming cars. We can also use the data captured from these devices to provide a wide range of services to the campus population such as, reducing traffic congestion, restricting access to parking lots and reducing car theft. These services will be inexpensive, thus requiring no additional fees to students and staff. We developed a custom license plate dataset for this study since our license plates are different to those in public datasets. In the proposed solution, the captured image is processed on the edge device but queries information from a centralized database to make a decision. A copy of this database is cached on the device but that copy is only used if communication to the central database is lost. We trialled the system under various conditions and present a subset of our results. Once sufficient data is collected we also plan to use AI tools to provide further enhancements.