Derq sells a platform that combines data from traffic sensors, signal controllers and other sources, running the information through AI processes to help with traffic control and other tasks.
The two companies tested “the capacity of the Derq platform in terms of number of simultaneous video streams that can be processed in real-time per server,” according to the paper describing the results.
More specifically, the tests demonstrated how the platform can run a “large number of sensor streams with minimal processing and hardware constraints,” according to that paper.
The tests showed that the Derq platform can run what the paper called “high-performance deep learning models with minimal processing and low latency leading to highly accurate data outputs on both edge and cloud-based architectures.”
That matters for public agencies interested in buying and deploying new traffic control technology, Karl Jeanbart, Derq’s co-founder and chief operating officer, told Government Technology via email.
The tests show “that Derq’s algorithms can achieve cutting-edge performance using off-the-shelf cost-effective compute platforms,” Jeanbart said. “Public agencies can now deploy and benefit from Derq’s analytics solutions in a cost-effective manner, leveraging their existing infrastructure such as cameras and servers.”
The tests also mean that Derq can target a larger customer base, he said, with the company’s technology able to help traffic operation centers with automated traffic monitoring and faster incident response times.
“Derq can also provide agencies’ IT departments with more options and flexibility in procuring and deploying systems,” Jeanbart said.
