Connect With Us


Miovision: Smart City roadside AI

William Payne
May 18, 2017


Canadian smart cities AI firm Miovision is using dedicated NVIDIA GPUs to create neural nets to provide machine learning solutions for traffic analysis. Video of vehicle, cyclist and pedestrian traffic is transformed into cloud based data analytics that can be used by city planners to improve traffic flow. The company is working with a number of universities, including Waterloo, Toronto and Sherbrooke, to develop deep AI solutions for traffic analysis. Miovision believes that in time, deep AI processing and analysis can be performed at the roadside.

Earlier this year, Miovision launched Miovision Labs to be the research and testing ground for the next generation of traffic technology. Part of that work is helping cities use traffic data to make better decisions, and study how emerging technologies such as computer vision, deep learning, big data analytics and embedded device design can play roles.

Miovision Labs is exploring new ways to make intersections more responsive to real-time traffic conditions. They do this using NVIDIA graphics processing units (GPUs) to quickly analyse volumes of data.

The company believes that in future, AI will be built into devices to process data at the roadside. That information will then be fed back into the system to alert the right departments in real time. A sensor might detect a double-parked car blocking a lane of traffic and alert traffic managers to adjust nearby traffic lights to improve vehicle flow. They could also alert police in case of an emergency.

“Miovision is creating the world's first neural network for traffic,” said Kurtis McBride, Miovision CEO and co-founder. “In the near future, the combination of our technology will make it easier for cities to monitor traffic and address problems to help people get around more quickly and safely.”

“Today, we're helping cities with information and technology to build better transportation systems, but we've only scratched the surface,” McBride said. “Through the research we're doing at Miovision Labs, using NVIDIA GPUs, with academic partners like the University of Waterloo, University of Toronto, and University of Sherbrooke, we're building a neural network for transportation that will be the foundation for smart cities of tomorrow.”