CC BY-NC-SA 4.0 All datasets licensed under CC BY-NC-SA 4.0

Emirates Multi-Task (EMT) Dataset

Emirates Multi-Task (EMT) is a comprehensive dataset for autonomous driving research, containing 57 minutes of diverse urban traffic footage from the Gulf Region. The dataset provides rich semantic annotations across two agent categories: people (pedestrians and cyclists), vehicles (seven classes). Each video segment spans 2.5-3 minutes, capturing challenging real-world scenarios:

  • Dense Urban Traffic: Complex multi-agent interactions in congested scenarios
  • Weather Variations: Clear and rainy conditions
  • Visual Challenges: High reflections from road surfaces and adverse weather combinations (rainy nights)

More details can be found in the download page and paper.

Off-Road Open Desert Trail Detection (O2DTD) Dataset

O2DTD is the first dataset on desert freespace detection, collected with six different light conditions (dawn, morning, afternoon, sunset, twilight, and night), containing a total of 5,045 RGB images.

Khalifa University’s Autonomous Shuttle (KUAS) Dataset

The dataset includes around twenty minutes of unlabeled data (A set of 8 LiDARs, Monochrome Cameras, IMU, GPS) captured from an Autonomous Shuttle that is deployed and operated in Khalifa University, SAN Campus, UAE.