The system’s plug-and-play AI sensors are installed at each intersection approach to provide road user detection and classification via a fusion of machine vision and radar. Users are classified as bikes, pedestrians, trucks, buses and more. Anonymised processed data is streamed to the cloud for city-grid coordination. Connected vehicle capabilities are integrated directly into the sensor units.
Meanwhile, the AI optimisation engine, installed in the traffic light cabinet, aggregates sensor data from each approach, calculates optimisation variables and monitors safety. It also acts as a gateway to cloud-based data analytics software and the user portal.
And it works! A project in Chandler, Arizona, showed a 20% reduction in traffic delays – down from 33.45 seconds to 26.62 per vehicle.