Working closely with local authorities, Cycling Scotland has installed AI sensors in four Scottish cities, providing historic and real-time multimodal data at key locations of transformational active travel infrastructure.
The novel technology brings numerous potential benefits over traditional bike counting technologies, including the ability to;
- Identify 10 classes (forms of traffic) and 32 sub-classes of mode
- View path movements across the road space by class
- Calculate speed and journey times between sensors
- Update software periodically to refine performance

About AI sensors
How do our sensors work?
The sensors are small, light enclosures containing a camera and a processor which allows us to collect real time, anonymous, highly accurate multimodal traffic data.
They are mounted and take small amounts of power from either a lamp post or a traffic light signal. Once installed, virtual countlines are created within the sensor’s field of view, and different road users are counted as they cross these countlines.
The data collected is sent to the cloud and the video is discarded at source.
How does the technology differentiate between different modes?
The sensors use artificial intelligence, developed by reviewing hundreds of thousands of images, to accurately classify different vehicle types based on typical features.
Bikes for example are thinner and more transparent than motorbikes and these characteristics help them to be differentiated from one another.
