You are viewing 1 of 2 articles without an email address.


All our articles are free to read, but complete your details for free access to full site!

Already a Member?
Login Join us now

Motional expands dataset for safer autonomous driving

In the 18 months since launch, more than 10 new datasets have been made publicly available across the industry, helping to build “safer, more educated” autonomous vehicles.

LinkedInTwitterFacebook
Lidar segmentation provides a significantly more detailed picture of a vehicle’s surroundings
Lidar segmentation provides a significantly more detailed picture of a vehicle’s surroundings

Driverless technology company Motional has announced an expansion to its publicly available nuScenes dataset to help enable a “safer, smarter” autonomous driving industry.

 

According to Motional, nuScenes, created in March 2019, was the first publicly available dataset of its kind, and pioneered an industry-wide culture of safety-focused data-sharing and collaboration.

 

Self-driving vehicles

 

It launched as a collection of 1,000 urban “street-scenes” in Boston and Singapore. Motional reports, the scenes, composed of millions of photos and data-points collected from the vehicles’ full sensor suites, were then meticulously hand-annotated, and used to inform and advance machine learning models with the aim of building the safest possible self-driving vehicles.

 

The expanded dataset now includes nuScenes-lidarseg and nuImages and Motional cited it as a demonstration of its commitment to making driverless vehicles a “safe and reliable reality”.

  • nuScenes-lidarseg is the application of lidar segmentation to the original 1,000 Singapore and Boston driving scenes, making it the largest publicly available dataset of its kind. Lidar segmentation provides a significantly more detailed picture of a vehicle’s surroundings than the original nuScenes’ bounding boxes, and adds an 1.4 billion annotated lidar points. Motional claims it’s a significant step forward for the industry – allowing researchers to study and quantify novel challenges such as lidar point cloud segmentation and foreground extraction
  • nuImages is a brand-new dataset comprising nearly 100,000 annotated images, selected to generate a wide range of unpredictable, challenging driving conditions. nuImages was created in response to user-demand and will help self-driving vehicles safely navigate unusual scenarios like jaywalkers, large intersections, and challenging weather conditions.
NuImages offers a new dataset that comprises nearly 100,000 annotated 2D images
NuImages offers a new dataset that comprises nearly 100,000 annotated 2D images

“Safety transcends competition,” said Karl Iagnemma, president and CEO, Motional. “The belief that passenger safety must take priority over any competitive advantage is at the heart of nuScenes.”

 

Since release, Motional claims more than 8,000 researchers have used nuScenes, with some 250 scientific papers published using the data. Furthermore, nuScenes gave rise to a culture of information-sharing in the autonomous vehicle industry.

“The belief that passenger safety must take priority over any competitive advantage is at the heart of nuScenes”

In the 18 months since launch, more than 10 new datasets have been made publicly available across the industry, helping to build safer, more educated autonomous vehicles.

 

In 2019, Motional was part of the team that released Safety First for Automated Driving white paper, an organised framework for the development, testing and validation of safe automated passenger vehicles.

 

Since 2018, the company has also operated a public robotaxi fleet in Las Vegas, providing more than 100,000 rides.

 

You might also like:

LinkedInTwitterFacebook
Add New Comment
You must be a member if you wish to add a comment - why not join for free - it takes just 60 seconds!