E-scooters can impede autonomous vehicle safety. Learn how Aya enabled a client’s AV models to accurately detect and classify e-scooters.
Overview
The electric scooter market is booming and is expected to reach $34.7 billion by 2028, growing at an annual rate of 7.6%. This surge is transforming urban transportation, affecting everything from public transit to city infrastructure, and notably, how self-driving cars are trained to ensure safety.
With more e-scooters on city streets, regulations are struggling to keep up. City planners are working to figure out the best ways to manage this new technology while keeping everyone safe.
As the number of e-scooters grows, it’s crucial for AI companies to adapt to the rise in short-distance vehicles. They need to stay updated on new laws, common human behaviours, and ensure that self-driving cars can navigate safely in urban environments.
The Challenge
The Client needed to address a new challenge in training their self-driving car systems: the growing presence of e-scooters on European roads. Existing open-source data lacked sufficient examples of e-scooters and had inconsistent labels.
The task was to create a comprehensive dataset that would help self-driving car models recognize and predict the movements of e-scooters.
Solutions
Aya Data’s experts labelled 10,000 images featuring e-scooters in various settings and situations. Each e-scooter was marked and categorised by type and other relevant details.
Results
The labelled images were used to train the Client’s computer vision system. This improved the system’s ability to detect e-scooters in real time with 95% accuracy.
Ready to enhance urban mobility with advanced technology? Contact us to discover how our solutions can improve the safety and efficiency of autonomous vehicles.
Disclaimer: Aya Data values client confidentiality and will not disclose any client or project details beyond what is permitted. As a result, names and specific details have been kept anonymous in our case studies.