Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -51,8 +51,6 @@ Intent (setpoint) → Action (effort) → Outcome (feedback)
|
|
| 51 |
### Core Columns (Tier 1 - Universal)
|
| 52 |
```python
|
| 53 |
setpoint_pos_0..N # Commanded joint positions (rad) - for robots
|
| 54 |
-
setpoint_doc # Depth of cut (mm) - for CNC
|
| 55 |
-
setpoint_feed # Feed rate (mm/rev) - for CNC
|
| 56 |
effort_torque_0..N # Motor torque/current (Nm / A)
|
| 57 |
feedback_pos_0..N # Actual joint positions (rad)
|
| 58 |
timestamp # Seconds since episode start
|
|
@@ -101,12 +99,14 @@ print(f"REASSEMBLE: {df['episode_id'].nunique()} episodes")
|
|
| 101 |
|
| 102 |
## Minimum Viable Episode (MVE)
|
| 103 |
|
| 104 |
-
|
| 105 |
- ≥1 setpoint signal (commanded intent)
|
| 106 |
- ≥1 effort signal (motor response)
|
| 107 |
|
| 108 |
This enables causal analysis: if `effort` doesn't follow `setpoint`, something is wrong.
|
| 109 |
|
|
|
|
|
|
|
| 110 |
## Fault Types
|
| 111 |
|
| 112 |
| Code | Description | Dataset |
|
|
@@ -128,8 +128,9 @@ Forgis/factorynet-hackathon/
|
|
| 128 |
├── voraus/
|
| 129 |
│ └── voraus_ad_100hz_factorynet.parquet # 461 MB - Yu-Cobot time series
|
| 130 |
├── nasa_milling/
|
| 131 |
-
│ ├── nasa_milling_factorynet.parquet # 15 MB - CNC
|
| 132 |
-
│
|
|
|
|
| 133 |
├── rh20t/
|
| 134 |
│ ├── rh20t_factorynet.parquet # 31 MB - Franka manipulation
|
| 135 |
│ └── rh20t_metadata.json # Episode metadata
|
|
|
|
| 51 |
### Core Columns (Tier 1 - Universal)
|
| 52 |
```python
|
| 53 |
setpoint_pos_0..N # Commanded joint positions (rad) - for robots
|
|
|
|
|
|
|
| 54 |
effort_torque_0..N # Motor torque/current (Nm / A)
|
| 55 |
feedback_pos_0..N # Actual joint positions (rad)
|
| 56 |
timestamp # Seconds since episode start
|
|
|
|
| 99 |
|
| 100 |
## Minimum Viable Episode (MVE)
|
| 101 |
|
| 102 |
+
Most datasets in FactoryNet satisfy the **Minimum Viable Episode** constraint:
|
| 103 |
- ≥1 setpoint signal (commanded intent)
|
| 104 |
- ≥1 effort signal (motor response)
|
| 105 |
|
| 106 |
This enables causal analysis: if `effort` doesn't follow `setpoint`, something is wrong.
|
| 107 |
|
| 108 |
+
**Exception: NASA Milling** - This dataset contains only effort signals (spindle current, vibration, acoustic emission) with no dynamic setpoints. Machining parameters (depth of cut, feed rate) are constant per episode and stored in metadata only. For anomaly detection on this dataset, "normal" behavior should be inferred from healthy episodes rather than comparing to commanded setpoints.
|
| 109 |
+
|
| 110 |
## Fault Types
|
| 111 |
|
| 112 |
| Code | Description | Dataset |
|
|
|
|
| 128 |
├── voraus/
|
| 129 |
│ └── voraus_ad_100hz_factorynet.parquet # 461 MB - Yu-Cobot time series
|
| 130 |
├── nasa_milling/
|
| 131 |
+
│ ├── nasa_milling_factorynet.parquet # 15 MB - CNC effort signals (current, vibration, acoustic)
|
| 132 |
+
│ ├── nasa_milling_extensions.parquet # Extensions with vibration/acoustic data
|
| 133 |
+
│ └── nasa_milling_metadata.json # Episode metadata with wear labels + machining params
|
| 134 |
├── rh20t/
|
| 135 |
│ ├── rh20t_factorynet.parquet # 31 MB - Franka manipulation
|
| 136 |
│ └── rh20t_metadata.json # Episode metadata
|