OpenBot
Back to Explore
Manipulation datasetOpen

CognitiveDrone Dataset

CognitiveDrone is a simulated UAV vision-language-action dataset from Skoltech's Intelligent Space Robotics Laboratory, built to train drones on tasks that require reasoning rather than pure flight control. Each trajectory spawns the quadrotor before a set of virtual gates and asks it to fly through the correct one by solving an embedded cognitive task, across three categories: Human Recognition, Symbol Understanding, and Reasoning. Episodes are generated in Gazebo with ArduPilot SITL, recording 256x256 first-person RealSense frames plus 4D velocity commands (Vx, Vy, Vz, ω) at each step. *Two caveats: the release declares no license, and only 30 of the intended 128 RLDS shards were actually uploaded (with no TFDS metadata), so roughly a quarter of the 8,062 trajectories are retrievable from HuggingFace. The 436 complete raw validation episodes in the collector repo are fully usable — the sample frames here come from those.*

Scale
8062 episodes
Formats
rlds
License
unknown
Published
2025-03-03

Decision summary

Best for

scripted

Main blocker

Metadata requires review against the official source before publication.

Next check

Read the dataset manifest and feature schema.

Catalog assessment

Selection evidence

73/100

Provisional · confidence 48 · 4/6 evaluated

Access and governance

Access and license are declared by the source.

usefulfit 75 · confidence 85

Schema and signal coverage

No machine-readable schema has been verified yet.

unknownnot scored · confidence 15

Policy training readiness

Observation and action/state signals are declared; alignment quality still depends on sample verification.

usefulfit 85 · confidence 55
View 3 more dimensions

World-model readiness

Temporal observations plus geometry or semantic context are declared; sample alignment remains to be audited.

usefulfit 68 · confidence 50

Failure and recovery readiness

No verified failure/recovery annotation evidence is available yet.

unknownnot scored · confidence 15

Download and processing readiness

Scale is declared; transfer and processing estimates are not measured.

usefulfit 65 · confidence 65
Review unresolved evidence and next checks

Signal gaps

  • Feedback / correction / failure. Not enough evidence is available to classify this signal.

Next checks

  • Read the dataset manifest and feature schema.
  • Run a bounded sample audit before assigning Strong readiness.

Record specifics

Dataset facts

Source
Skoltech — Intelligent Space Robotics Laboratory
Evidence
secondary claim
Formats
rlds
episodes
8062
tasks
3
bytes
2466317963
Read paper

Metadata coverage

Loop signals

6/7 present or partial

Feedback / correction / failure

No decision-grade evidence captured yet.

unknown
View 6 more signal categories
Observation / ego video

video

present
Action / hand pose / robot state

CognitiveDrone is a simulated UAV vision-language-action dataset from Skoltech's Intelligent Space Robotics Laboratory, built to train drones on tasks that require reasoning rather than pure flight control. Each trajectory spawns the quadrotor before a set of virtual gates and asks it to fly through the correct one by solving an embedded cognitive task, across three categories: Human Recognition, Symbol Understanding, and Reasoning. Episodes are generated in Gazebo with ArduPilot SITL, recording 256x256 first-person RealSense frames plus 4D velocity commands (Vx, Vy, Vz, ω) at each step. *Two caveats: the release declares no license, and only 30 of the intended 128 RLDS shards were actually uploaded (with no TFDS metadata), so roughly a quarter of the 8,062 trajectories are retrievable from HuggingFace. The 436 complete raw validation episodes in the collector repo are fully usable — the sample frames here come from those.*

present
Gaze / attention

CognitiveDrone is a simulated UAV vision-language-action dataset from Skoltech's Intelligent Space Robotics Laboratory, built to train drones on tasks that require reasoning rather than pure flight control. Each trajectory spawns the quadrotor before a set of virtual gates and asks it to fly through the correct one by solving an embedded cognitive task, across three categories: Human Recognition, Symbol Understanding, and Reasoning. Episodes are generated in Gazebo with ArduPilot SITL, recording 256x256 first-person RealSense frames plus 4D velocity commands (Vx, Vy, Vz, ω) at each step. *Two caveats: the release declares no license, and only 30 of the intended 128 RLDS shards were actually uploaded (with no TFDS metadata), so roughly a quarter of the 8,062 trajectories are retrievable from HuggingFace. The 436 complete raw validation episodes in the collector repo are fully usable — the sample frames here come from those.*

present
Language intent / task phase

language · CognitiveDrone is a simulated UAV vision-language-action dataset from Skoltech's Intelligent Space Robotics Laboratory, built to train drones on tasks that require reasoning rather than pure flight control. Each trajectory spawns the quadrotor before a set of virtual gates and asks it to fly through the correct one by solving an embedded cognitive task, across three categories: Human Recognition, Symbol Understanding, and Reasoning. Episodes are generated in Gazebo with ArduPilot SITL, recording 256x256 first-person RealSense frames plus 4D velocity commands (Vx, Vy, Vz, ω) at each step. *Two caveats: the release declares no license, and only 30 of the intended 128 RLDS shards were actually uploaded (with no TFDS metadata), so roughly a quarter of the 8,062 trajectories are retrievable from HuggingFace. The 436 complete raw validation episodes in the collector repo are fully usable — the sample frames here come from those.*

present
Sim-real pairing

simulation · CognitiveDrone is a simulated UAV vision-language-action dataset from Skoltech's Intelligent Space Robotics Laboratory, built to train drones on tasks that require reasoning rather than pure flight control. Each trajectory spawns the quadrotor before a set of virtual gates and asks it to fly through the correct one by solving an embedded cognitive task, across three categories: Human Recognition, Symbol Understanding, and Reasoning. Episodes are generated in Gazebo with ArduPilot SITL, recording 256x256 first-person RealSense frames plus 4D velocity commands (Vx, Vy, Vz, ω) at each step. *Two caveats: the release declares no license, and only 30 of the intended 128 RLDS shards were actually uploaded (with no TFDS metadata), so roughly a quarter of the 8,062 trajectories are retrievable from HuggingFace. The 436 complete raw validation episodes in the collector repo are fully usable — the sample frames here come from those.*

present
License / format / access

Open · unknown · rlds

present
Evidence details and provenance

Official signal claims

LanguageProprioceptionVideo

Schema and annotations

No machine-readable schema facts are captured.

Sample verification

Pending. Metadata does not prove sample coverage, alignment, or file integrity.

curated official source

Integration notes

  • Metadata requires review against the official source before publication.

Catalog links

Related records