The artificial neural networks-based smart number plate of vehicles with real-time traffic signs recognition and notification: a case study of public transport in EAST AFRICAN COMMUNITY (EAC)

Abstract

The world is advancing technologically in all sectors, including intelligent transportation, whereby various vehicles' movements are monitored and controlled remotely. These technologies simplify the tasks in traffic control and increase road safety. The previous related works implemented and designed provided different technologies that can identify, locate and detect the vehicle's speed. Even though these technologies have been implemented, there is still

a lack of assistance to drivers for earlier knowing and reminding the road situation and to real- time notify the dedicated authorities such as traffic police stations once an accident happens. In

this project, an Artificial Neural Network based smart number plate with real-time traffic sign recognition and notification was developed. The developed smart number plate comprises two parts, the smart plate and the display. The smart plate comprises a sensing and processing unit, while the display comprises a notification unit, and both communicate through wireless communication. The sensing unit contains a speed sensor, vibration shock sensors, a Global Position System (GPS), and an AI-Thinker camera. The processing unit comprises espressif board ESP32-CAM and ESP32-S that act as controllers. The notification unit contains the Liquid crystal Display (LCD), Global System for Mobile communication (GSM), and Buzzer. With the TensorFlow model for machine learning, the smart number plate classifies and recognizes traffic speed signs with real-time notification. This smart number plate had been tested on different vehicles, and it assisted drivers in obeying the traffic speed signs earlier, and the traffic police station had been alerted for emergency support. Moreover, the remained traffic signs such as informative traffic signs were not detected and was recommended to be added into the future machine learning models.