Simaihub Mobile Robot Navigation Training Dataset Sample

Simaihub provides licensed expert navigation trajectories for training autonomous mobile robots. The dataset includes data from multiple indoor scenes like hospitals and warehouses, with domain randomization and dynamic obstacles. It originates from NVIDIA Isaac Sim and Isaac Lab simulations.

Use Cases

• Training offline reinforcement learning agents based on expert navigation trajectories.
• Developing imitation learning policies for mobile robots based on multi-scene indoor data.
• Benchmarking robot navigation algorithms in environments with dynamic obstacles and recovery behaviors.
• Studying domain generalization for robotics using data with randomized lighting and sensor noise.

Strengths

• Data provenance from NVIDIA Isaac Sim and Isaac Lab simulation platforms.
• Includes multiple indoor scene types: Hospital, Office, Warehouse, and Simple Room.
• Covers complex navigation scenarios like U-turns, narrow passages, and recovery behaviors.
• Incorporates domain randomization techniques such as dynamic obstacles and sensor noise.

Limitations

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

Provenance

Source

Simaihub

Collection Method

Simulated using NVIDIA Isaac Sim and Isaac Lab.

Freshness

Last updated 2026-07-12 09:27:51; freshness should be verified.