Preparing For The Growing Popularity of E-Scooters


The global electric scooters market size is expected to reach USD 34.7 billion by 2028, expanding at a CAGR of 7.6%. In the US, people complete some 40 million e-scooter trips each year, and many European capitals now have some form of e-scooter rental scheme. There are some 15,000 rental e-scooters available in Paris alone. 

The rapid influx of e-scooters has transformed the transportation and highway infrastructure landscape, especially in urban areas. 

Critically to this project, the increasing prevalence of e-scooters on the roads is changing the way we train autonomous vehicles. Autonomous vehicles must be able to see e-scooters and react swiftly to ensure the safety of their users.


Autonomous Vehicles


Amsterdam, Netherlands

Company Size


E-scooter Safety: A Regulatory Conundrum 

The rapid proliferation of rideshare e-scooters has not been met by advances in regulation, and city councils are still trying to understand how to harness their benefits while keeping people safe.

The public use of e-scooters is permitted in many countries, including Germany, Belgium, Norway, France, Switzerland, Austria, the Czech Republic, some cities in the USA, Singapore, and Malaysia. There is seemingly no mutual agreement on how to regulate e-scooters, and the rules are frequently changing, leading to mass confusion amongst the public. 

For example, in Japan, you need a motorbike license to use an e-scooter. In the UK, e-scooters are currently illegal to use on pavements, cycle paths, or roads unless part of a government-backed trial scheme. In some countries, you can only use e-scooters on pavements, and in others, you can only use them on roads. 

But there is yet another fly in the ointment: how do e-scooters interact with self-driving autonomous vehicles?

Services: Polygon, Bounding Box, Semantic Segmentation

The Challenge

E-scooters are the most popular electric micro-transportation device used worldwide, and the benefits are evident; not only are e-scooters eco-friendly, but they’re easy to access, easy to use, and cheap to buy and repair. 

The primary drawback? Safety. The issue is threefold; e-scooters pose risks to both the rider and pedestrians, and there is also concern that self-driving vehicles are not yet equipped to recognize the unique form factor and behavior of e-scooters.

Our international client wanted to cover an unaddressed edge case in their AV training workflow - the increased prevalence of e-scooters on European roads. Open source datasets could not provide sufficient instances of this edge case, or were labeled inconsistently.

The challenge was to create a large and purpose-built dataset that would enable AV models to detect, classify and ultimately predict the movements of e-scooters.

The Solution

Aya Data’s AV labeling experts labeled 10,000 images of e-scooters in different contexts, positions, and situations. The e-scooters were bounded, and tagged with classes denominating the type and associated properties.

Creating this bespoke e-scooter dataset aided the client in testing the interaction between e-scooters and autonomous vehicles. Our dataset, complete with various e-scooter classifications, helped facilitate the development of testing models. 

Future-proofing Our Cities

As cities begin to integrate more autonomous vehicles into their fleets and permit the use of private self-driving cars, it’s imperative that every component of the streets and roads is properly labeled, including e-scooters.  

Increasing Road Safety

Computer vision is already used to auto-detect pedestrians via cameras fitted to e-scooters, triggering auto-braking in the event of a likely collision. The same process needs to be applied to self-driving cars to protect e-scooter users, particularly where riding on the pavement is not permitted. 

Reach Consensus on E-scooter Safety 

Developing reliable solutions to tackle the challenges of autonomous transportation helps governments and local authorities reach international standards of micro-transportation regulation.


The images labeled by Aya Data were used to train a computer vision model to detect e-scooters in real-time, with 96% accuracy.


images labeled


accuracy in model


of collisions avoided

Explore Case Studies

Medical Imaging

Medical imaging is a time-intensive task that is being revolutionized by AI and ML.

Content Moderation For Advertising and Marketing

Reducing the risk of reputational damage for brands and businesses.

Ethical Data Sourcing in Africa

Get inspired by AyaData's case studies on data sourcing. See how we've helped businesses optimize their processes and achieve success. Learn more.

Image Segmentation in Satellite Imagery

Using satellite images to measure land-use changes in South America.

Real-Time Transcription of US Police Radios

Private companies are turning to technology to help combat crime and reduce the burden on society.

Identifying Shoplifting Events in Retail Stores

Companies are looking for innovative ways to curb their losses with the use of technology.

Detecting Disease in Ghana’s Maize plants

As global demand for maize continues to rise, the opportunity for technology to improve yields is more important than ever.