standard-cognition2025cc-by-4.0

Tracks Dataset

Captures continuous real-world human movement in retail environments, providing 3D pose sequences sampled at 10 Hz across normalized store coordinates for robotics and embodied AI research.

Downloads143
Episodes2.3 million unique shoppers
Hours60000
Likes1

Why This Matters for Physical AI

Provides the largest structured pose-based trajectory corpus from real human behavior in retail environments, enabling development of human-aware navigation, motion planning, and humanoid gait control for embodied AI systems.

Technical Profile

Modalities
rgbpose_estimationkeypoint_detection
Robot Embodiments
humanoid
Environment
retail
Task Types
navigationmotion_planninghuman-aware_navigationgait_learning
Episodes
2.3 million unique shoppers
Total Hours
60000
Data Format
CSV
Annotation Types
pose_keypointstrajectory_labels
License
cc-by-4.0
Part of the Tracks Dataset family

Access

Need custom rgb data?

Claru builds purpose-built datasets for retail applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets