Sign up for our weekly emails to stay up to date on the latest news!
NYU Tandon School of Engineering, in collaboration with May Mobility, has introduced a groundbreaking dataset that promises to accelerate the development of autonomous vehicle (AV) technology. This new dataset, named MARS (MultiAgent, multitraveRSal, and multimodal), provides unprecedented real-world driving data, offering valuable insights for researchers and advancing the field of AVs.
Key Highlights:
- MARS dataset offers unique real-world driving data captured from multiple autonomous vehicles (AVs) over repeated trips.
- The dataset was introduced by NYU Tandon in collaboration with May Mobility and was presented at the IEEE / CVF Computer Vision and Pattern Recognition (CVPR) Conference.
- MARS is publicly available and enables the study of AV interactions and environmental understanding over time.
- The collaboration showcases the potential of industry-academic partnerships in advancing AV research.
The MARS dataset, curated by NYU Tandon’s Automation and Intelligence for Civil Engineering (AI4CE) Lab and May Mobility engineers, was collected using a fleet of four autonomous Toyota Sienna Autono-MaaS vehicles. These vehicles operated within a 20-kilometer zone in a U.S. city, encompassing residential, commercial, and university areas. The dataset includes sensor data such as LiDAR, Camera, GPS/IMU, and vehicle state information, which are crucial for developing advanced AV algorithms.
“Datasets for autonomous vehicle research have typically come from a single vehicle’s one-time pass of a certain location. MARS offers many more opportunities because it captures real interactions between multiple AVs traveling over fixed routes hundreds of times, including repeated passes through the same locations under varying conditions,” said Chen Feng, lead researcher on the project and assistant professor at NYU Tandon.
The MARS dataset stands out due to its “multitraversal” nature, with data collected from thousands of passes through 67 specific locations at different times of day and under varying weather conditions. This allows researchers to study how AVs can use prior knowledge to enhance their real-time understanding of the environment.
May Mobility’s FleetAPI subscription service enabled NYU Tandon researchers to access over 1.4 million frames of synchronized sensor data. This collaboration sets a precedent for industry-academic partnerships that benefit the entire AV industry.
“The MARS dataset allows us to study both how multiple vehicles can collaboratively perceive their surroundings more accurately, and how vehicles can build up a detailed understanding of their environment over time,” added Feng.
The release of MARS comes at a crucial time as the AV industry strives to move beyond controlled testing environments and navigate the complexities of real-world driving. This dataset, collected from multiple commercial vehicles in actual use, plays a vital role in training and validating the artificial intelligence systems that power AVs.
May Mobility and NYU Tandon‘s collaboration underscores their commitment to creating safer mobility solutions and improving autonomous driving algorithms, furthering the development of AV technology.
Sign up for our weekly emails to stay up to date on the latest news!
