Voi tests e-scooters with built-in pedestrian detection

By this point, most people know electric scooters aren’t the safest form of transportation you can use to get around a city. Frequently, people don’t have the necessary experience to know how to respond to certain road situations. Voi, a Swedish startup that offers electric scooter rentals throughout the UK and continental Europe, thinks it may have a solution. The company is partnering with another startup called Luna to add a computer vision machine learning algorithm and high-end cameras to its scooters that will assist them with detecting pedestrians.

In practice, the companies claim the tech will help Voi’s scooters better navigate areas densely populated with people. It will also help them understand what kind of surface or lane they’re driving over — whether that be a bike lane, sidewalk or road — and respond appropriately. Another part of the partnership will see Luna lend Voi GPS technology that will allow it to control the parking of its scooters down to the centimeter in select areas. They claim this functionality will help reduce street clutter, as well as streamline operating costs.

A year-long test of the tech will get underway this week in Northampton, UK. The initial phase will see Voi employees ride scooters equipped with the technology around the city to allow the AI to learn its surroundings. Once the training is complete, the public will be able to use those scooters.

Voi isn’t the first company in the space to turn to machine learning as a potential way to address some of the inherent problems with electric scooters. At the start of 2020, Lime began testing a feature that kept track of how often people rode its scooters on the sidewalk. If the accompanying AI determined someone spent more time on the sidewalk than on the road, it would send them a push notification to encourage them to ride on the street more often next time. Whether software alone can make scooters safer to ride is hard to say. But there’s clearly a need to do something about the issue. A recent UCSF study found that between 2014 and 2018, electric scooter-related injuries in the US jumped by 222 percent, with approximately 39,000 people hurting themselves in one way or another.