System uses streetlights to provide accurate, low-cost traffic counts

Internet of Things (IoT) pioneer Echelon Corporation has completed initial field tests of a new application for its InSight Cognitive Vision System, which can provide accurate and low-cost traffic counts.

A pilot deployment in Spokane, Washington State, validated the ability of the intelligent camera to accurately and inexpensively count traffic on city streets. Echelon’s new application captured data that matched or exceeded the accuracy of conventional and manual counters, day or night. The new streetlight-based system is less expensive and easier to deploy than conventional roadway surface-based counters, and because the system communicates over the existing lighting network, it can be deployed on any streetlight, not just those at intersections adjacent to traffic control boxes. Spokane’s InSight system also included an adaptive learning capability to accommodate differences in each camera’s deployment for light levels, shadows and reflections.

Traffic counting is traditionally accomplished with pneumatic tubes temporarily laid across the pavement surface, with inductive loops buried in the roadway, or with microwave radar transmitters. Echelon’s new InSight application makes use of access to the company’s Lumewave lighting control network, allowing placement of the self-learning system on any streetlight instead.

Moving the traffic counter from the roadway surface to a streetlight simplifies deployment, and moves the sensor into a more versatile location. By using a city’s existing connected lighting platform, the traffic counting application based on the InSight technology offers smart cities a lower cost alternative in managing energy use, as well as a platform to broadly collect traffic data.

“While there are many open-source computer vision algorithms available today, most address one particular problem under a narrow set of conditions and are not effective for reliably addressing multiple real-world issues,” said Sohrab Modi, CTO and SVP of engineering at Echelon. “We developed this cognitive AI vision technology to enable a municipal streetlight system to host an increasingly broad array of smart city solutions.”

Adam Miles, Spokane’s associate traffic engineer, commented, “We are encouraged by these field test results that demonstrate that InSight provides reliable traffic flow information directly into the street lighting process. Over time, we believe the aggregated traffic data we collect will enable us to address resident and commuter needs in new ways we are just beginning to consider.”

Echelon’s CEO, Ron Sege, added, “With this new application, cities can see how InSight will provide the foundation for a family of AI-based tools leveraging the omnipresence of street lights.”