ezharjanmit
KinHEC - Kinematic Trajectory Benchmark for Hand-Eye and Robot-World Calibration
A large-scale, fully synthetic, ground-truth-complete benchmark for hand-eye calibration (AX=XB) and robot-world calibration (AX=YB) problems, consisting of time-stamped pose streams from robot kinematics and camera observations with physically motivated noise models.
Downloads71
Episodes77628
Why This Matters for Physical AI
KinHEC provides a large-scale, systematically controlled benchmark for calibration problems fundamental to accurate robot perception and control, enabling robust comparison of classical and learning-based sensor calibration solvers under precisely specified conditions.
Technical Profile
- Modalities
- proprioceptionpose_estimation
- Robot Embodiments
- Universal Robots CB3Universal Robots e-SeriesFranka Emika PandaKUKA LBR iiwa 7KUKA LBR iiwa 14ABB IRB 120PUMA 560
- Environment
- simulation
- Task Types
- hand-eye-calibrationrobot-world-calibrationpose-estimationsensor-calibration
- Episodes
- 77628
- Data Format
- parquet
- Annotation Types
- pose_labelsoutlier_labelsvisibility_labelsground_truth_calibration
- License
- mit
Access
Need custom proprioception data?
Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.
Request a Sample Pack