Challenge Phase1 Dataset: Bimanual Robot Manipulation Trajectories

Posttraining-RFM-RSS2026 provides real-robot bimanual manipulation trajectories for three benchmark tasks. Data was collected on a bimanual YAM follower teleoperated by a GELLO leader arm, with timestamp-aligned frames across joint state and action. The dataset was released for the RSS 2026 Workshop & Challenge on Post-training for Robotics Foundation Models.

Use Cases

• Training or evaluating robot learning models based on real-world bimanual manipulation trajectories
• Benchmarking imitation learning algorithms on the three described manipulation tasks
• Developing temporal alignment methods for multi-modal robot state and action data
• Studying teleoperation strategies for bimanual robots based on the leader-follower setup

Strengths

• Contains real-robot data, not simulated, which likely increases its realism for model training
• Includes timestamp-aligned frames across multiple data modalities, which suggests precise synchronization
• Covers three distinct benchmark tasks, providing a basis for comparative evaluation

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment for large-scale training
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

Posttraining-RFM-RSS2026

Collection Method

Collected via teleoperation on a bimanual YAM follower robot controlled by a GELLO leader arm.

Freshness

Last updated 2026-05-11 13:12:32; freshness should be verified