Micromobility operator Tier has unveiled advanced AI capabilities designed in-house which it says can accurately identify e-scooter pavement and tandem riding.
Dubbed ‘Parrot’, the IoT module that connects to Tier’s e-scooters can detect anti-social riding behaviours without relying on camera or computer vision systems. It works by combining upgraded sensors, 4G connectivity and a “step change” in processing power.
“The Parrot works like a brain,” Tier’s CTO Nuno Simaria told Zag Daily. “It allows us to deploy more sophisticated algorithms on the e-scooter, enabling it to analyse rider behaviour without needing to resort to an expensive computer vision based system.”
Parrot e-scooter brain
Tier said the module can recognise how and where an e-scooter is being ridden, based on data from on-board sensors that are analysed by AI.
These e-scooter sensors measure changes in the texture of the road surface.
“By understanding changes from a smooth to bumpy surface, the e-scooter can identify if it feels like pavement,” said Simaria.
“It draws on a collection of data, such as the accelerometer, wheel ticks and more. It is the same principle for tandem riding, just with a different outcome.”
Since every city has different pavement and road surfaces, the model has to be trained specifically for the area it operates.
“Not all cities want to stop pavement riding across the whole city,” said Simaria. “They want to stop pavement riding when it disturbs people. Our module helps us to make sure that in those places that matter, it is possible to have a really high sense of certainty.”
The raw input from the sensors then goes through Tier’s AI model to make the decision in real-time. “If it is very sure, then it has options as to how that will feel for the end-user, such as slowing the vehicle down during the ride or sending an automated warning alert after their ride, depending on what cities are looking for.”
Simaria said the technology enables Tier to reach a superior level of performance and reaction time in tackling pavement and tandem riding, while the tech can also be deployed at a low cost.
The module is currently being tested on 60 e-scooters in London, with plans to retrofit Tier’s entire UK e-scooter fleet later this year. With the upgraded computing power, Tier expects to be able to extend the lifespan of its UK e-scooter fleet by an additional three to five years.