AI-driven LED streetlights adapt to traffic patterns

April 25, 2017 //By Julien Happich
IoT systems provider Echelon Corporation has developed a patent-pending cognitive vision-based technology that can enable a wide range of smart city and smart campus applications, including traffic-adaptive lighting.

Using artificial intelligence in vision-enabled edge devices, Echelon's InSight technology collects traffic data and processed at the edge of the network instead of on a central server. It then uses the Echelon's Lumewave lighting platform to transmit traffic information, reducing response time and improving reliability.  This architecture is said to enable faster action in response to changing conditions and minimizes network bandwidth requirements.

The first application of Echelon's new technology will be to provide traffic-adaptive lighting in Spokane, Washington. The cognitive vision system will be deployed on traffic intersection streetlights where it will analyze traffic flows and automatically adjust light levels to enhance safety while reducing energy consumption and maintenance costs. 

With InSight, each unit analyzes video streams locally and makes decisions about what light levels to set based on traffic volumes and conditions, triggering higher levels during peak hours and lowering light levels during non-peak hours. The solution leverages Echelon's connected streetlight control system, along with the trend by cities and campuses to upgrade to LED lighting.

While LED streetlight conversions typically reduce electricity consumption by 50 percent (for equivalent light levels), Echelon estimates that intelligent traffic-adaptive lighting could reduce consumption by an additional 30 to 40 percent. Traditionally, cities have used a wide variety of traffic monitoring systems such as in-ground loops, cameras, radar, or infrared to detect traffic for timing traffic signals. These are typically closed loop systems and provide limited information for traffic signal timing only.