Lyft recently released a Level 5 predictive data set for autonomous driving that includes more than 1,000 hours of driving. The company also launched the Autonomous Driving Sport Prediction Challenge, with a $30,000 prize pool.

Lyft has released a new data set.

Last July, Lyft released an L5 level of autonomous driving awareness data set containing more than 55,000 3D annotation frames tagged by humans. It was officially described at the time as the largest public data set of its kind.

Just over a year later, Lyft has released another data set of autonomous driving predictions at the L5 level.

To download the application, click here

170,000 scenarios, more than 2,500 kilometers of road data

The data set released by Lyft focuses on exercise prediction. Officials say one of the problems in long-term research in autonomous driving is creating models that are robust and reliable enough to predict traffic movements.

The data was collected by a fleet of 23 autonomous vehicles on a fixed route in Palo Alto, California, over a period of four months and includes travel logs of cars, pedestrians and other obstacles encountered.

The data set specifically includes:

  • 1,000 hours: More than 1,000 hours of recorded self-driving car movements;
  • 170,000 scenes: Each scene lasts about 25 seconds and includes traffic lights, aerial maps, sidewalks, etc.
  • 16,000 miles: 16,000 miles of data from public roads;
  • 15242 Annotated Maps: Includes a high-definition semantic map of the tagged elements as well as a high-definition aerial view of the area.

A sample semantic graph with a bird’s eye view of the dataset

The motion data is collected by an array of sensors mounted on the Lyft’s roof, which capture lidar, camera and radar data as the vehicle travels tens of thousands of miles.

In the data set, each scene encodes the state around the vehicle at a given point in time, with red for self-driving cars and yellow for other vehicles

Lyft said the collection, together with the provided toolkit, makes up the largest, most complete and detailed data set to date for developing autonomous driving, machine learning tasks such as motion prediction, planning, and simulation.

Currently, only a subset of the dataset is available for download, including:

  • Sample data set (53 MB)
  • Training data set (divided into three parts, 69.4GB in total)
  • Aerial view (2 GB)
  • Semantic Map (2 MB)

Download Address:

Prediction

Launch a challenge with a $30,000 prize pool

Lyft, meanwhile, plans to launch a challenge that will start on the Google Kaggle platform in August and award prizes totaling $30,000.

Last year, Lyft launched a self-driving 3D target detection competition with a $25,000 prize pool

Key points of this challenge:

  • Competition requirements: Participants predict the movement of vehicles;
  • Preparation: Researchers and engineers can download training data sets and Python-based software packages to experiment with the data from now on, officials warn. Because the test and validation suite will be released as part of the competition;
  • The ultimate goal: to empower the research community and accelerate innovation through data sets and competitions.

Sacha Arnoud, senior director of engineering at Lyft, and Peter Ondruska, director of audio and video research, wrote in a blog post, “Data is the driving force behind trying out the latest machine learning technologies, and while access to large-scale, high-quality automated driving data is limited, But that should not prevent us from experimenting with this research.”

“We believe autonomous driving will become a more convenient, safer and sustainable part of the transportation system,” Arnoud and Ondruska said. “By sharing data with the research community, we hope to identify important and unsolved challenges in autonomous driving.”

Visit Hyper. ai for more quality data sets!

Blog address:

https://medium.com/lyftlevel5/fueling-self-driving-research-with-level-5s-open-prediction-dataset-f0175e2b0cf8

Thesis Address:

https://arxiv.org/pdf/2006.14480.pdf

Making address:

https://github.com/lyft/l5kit/