Dataset card: two parts (paper, traversing study)
Browse filesPart A keeps the paper card text and all load_dataset configs unchanged. Part B documents traversing/, pinned at 6620faead5225ac9aa5ae8ab19bc2ef2db38a863.
README.md
CHANGED
|
@@ -59,7 +59,21 @@ configs:
|
|
| 59 |
data_files: raw/tracked/episodes.parquet
|
| 60 |
---
|
| 61 |
|
| 62 |
-
# NeDM
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
|
| 64 |
High-fidelity [Project Chrono](https://projectchrono.org) trajectories used to train the
|
| 65 |
neural reduced dynamics models (NN-ROMs) in
|
|
@@ -77,7 +91,7 @@ are published (70 GB in total):
|
|
| 77 |
* **`processed/`** — the four training caches the deployed models read (`.npy`), so the
|
| 78 |
paper's training configs run without touching the raw data.
|
| 79 |
|
| 80 |
-
## Datasets
|
| 81 |
|
| 82 |
| Config | System | Terrain / task | Rate | Episodes (train / val) | Rows | Raw Parquet | Columns |
|
| 83 |
|---|---|---|---|---|---|---|---|
|
|
@@ -95,14 +109,14 @@ channel) using the collectors in the code repository (`src/nedm/hmmwv_data.py`,
|
|
| 95 |
`scripts/collection/collect_hmmwv_crm_dataset.py`, `src/nedm/arm_data.py`,
|
| 96 |
`src/nedm/tracked_vehicle_data.py`).
|
| 97 |
|
| 98 |
-
### Splits
|
| 99 |
|
| 100 |
Train/val is decided **per episode at collection time** and stored in the `split` column:
|
| 101 |
`sha1(episode_id)[:8] / 0xFFFFFFFF < validation_ratio → val` (ratio 0.20 for the HMMWV sets,
|
| 102 |
0.15 for `arm` and `tracked`). Whole episodes stay together; the assignment depends only on the
|
| 103 |
episode id, so it is stable under re-sharding. `train` and `val` files never share an episode.
|
| 104 |
|
| 105 |
-
## Layout
|
| 106 |
|
| 107 |
```
|
| 108 |
raw/<config>/train/<shard>.parquet transitions, one file per raw collection shard
|
|
@@ -121,7 +135,7 @@ physical channels are `float32` (`time_s` is `float64`; `sample_index`, `collisi
|
|
| 121 |
identifiers are dictionary-encoded strings). Files are zstd-compressed with BYTE_STREAM_SPLIT
|
| 122 |
float encoding and ≤ 262,144-row row groups.
|
| 123 |
|
| 124 |
-
### Column groups
|
| 125 |
|
| 126 |
**HMMWV (`hmmwv_flat`, `hmmwv_bumpy`, `hmmwv_crm` — identical 105 columns).** Units are in the
|
| 127 |
names (`_m`, `_mps`, `_mps2`, `_rad`, `_radps`, `_n`, `_nm`); world frame is Chrono's ISO
|
|
@@ -157,7 +171,7 @@ contact or joint-limit hit, so lengths are 9–500 steps (mean ≈ 58). The pape
|
|
| 157 |
tire channels, plus `left_sprocket_speed_radps`, `right_sprocket_speed_radps`. The paper's 3-D
|
| 158 |
state is `vel_body_x_mps, vel_body_y_mps, yaw_rate_radps`; action is the driver triple.
|
| 159 |
|
| 160 |
-
### `episodes.parquet` and the metadata bundle
|
| 161 |
|
| 162 |
`episodes.parquet` flattens each episode's `dataset_index.json` entry and its JSON sidecar
|
| 163 |
(nested values are JSON strings): `episode_id`, `split`, `scenario_family`, `rows`,
|
|
@@ -171,7 +185,7 @@ state is `vel_body_x_mps, vel_body_y_mps, yaw_rate_radps`; action is the driver
|
|
| 171 |
terrain and solver settings), every per-episode sidecar, and the shard-plan manifests. It is what
|
| 172 |
lets the release be turned back into the collectors' original directory tree (below).
|
| 173 |
|
| 174 |
-
## Loading
|
| 175 |
|
| 176 |
**Streaming with 🤗 `datasets`** (no download of the 44 GB flat set required):
|
| 177 |
|
|
@@ -207,7 +221,7 @@ PYTHONPATH=src python scripts/release/download_nedm_datasets.py --dataset arm --
|
|
| 207 |
The rehydrated CSVs carry the float32 values the trainer uses; caches rebuilt from them are
|
| 208 |
bit-identical to the ones in `processed/` (this is checked in the release validation).
|
| 209 |
|
| 210 |
-
## Processed caches
|
| 211 |
|
| 212 |
| Cache | Trained model | State | Action | Transitions (train / val) | Size |
|
| 213 |
|---|---|---|---|---|---|
|
|
@@ -222,7 +236,7 @@ Each cache holds contiguous `float32` arrays `{train,val}_{states,actions,target
|
|
| 222 |
(`state_fields`, `action_fields`, `dt_s`, train-split mean/std used for normalisation). Values are
|
| 223 |
raw physical units; the model applies the statistics.
|
| 224 |
|
| 225 |
-
## Known limitations
|
| 226 |
|
| 227 |
* `hmmwv_bumpy` episodes are short (mean 3.3 k rows) because 78 % end on the 0.9 × 500 m
|
| 228 |
keep-in guard; the regime is meant as a test set.
|
|
@@ -232,7 +246,7 @@ raw physical units; the model applies the statistics.
|
|
| 232 |
fluid–solid interaction force.
|
| 233 |
* Simulation is deterministic and noise-free; there is no sensor model.
|
| 234 |
|
| 235 |
-
## Citation
|
| 236 |
|
| 237 |
```bibtex
|
| 238 |
@article{zhang2026abstraction,
|
|
@@ -245,3 +259,388 @@ raw physical units; the model applies the statistics.
|
|
| 245 |
|
| 246 |
License: BSD-3-Clause (same as the code). Simulation assets are Project Chrono's HMMWV and M113
|
| 247 |
models; the LRV arm geometry is in the code repository (`src/arm_model/`).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
data_files: raw/tracked/episodes.parquet
|
| 60 |
---
|
| 61 |
|
| 62 |
+
# NeDM datasets: the paper and the traversing study
|
| 63 |
+
|
| 64 |
+
This repository holds two separate releases. **Part A** is the set of datasets behind the NeDM paper, unchanged since
|
| 65 |
+
its release (Hub tag `paper-v1`). **Part B**, everything under `traversing/`, holds the drives, training files, models
|
| 66 |
+
and evaluation records of the traversing study, follow-on work that is not part of the paper.
|
| 67 |
+
|
| 68 |
+
Contents:
|
| 69 |
+
|
| 70 |
+
- [Part A: datasets of the paper](#part-a-datasets-of-the-paper): `raw/`, `processed/`, `assets/` and
|
| 71 |
+
`release_manifest.json`
|
| 72 |
+
- [Part B: traversing study](#part-b-traversing-study): `traversing/`
|
| 73 |
+
|
| 74 |
+
## Part A: datasets of the paper
|
| 75 |
+
|
| 76 |
+
### NeDM — Neural Reduced Dynamics Datasets
|
| 77 |
|
| 78 |
High-fidelity [Project Chrono](https://projectchrono.org) trajectories used to train the
|
| 79 |
neural reduced dynamics models (NN-ROMs) in
|
|
|
|
| 91 |
* **`processed/`** — the four training caches the deployed models read (`.npy`), so the
|
| 92 |
paper's training configs run without touching the raw data.
|
| 93 |
|
| 94 |
+
### Datasets
|
| 95 |
|
| 96 |
| Config | System | Terrain / task | Rate | Episodes (train / val) | Rows | Raw Parquet | Columns |
|
| 97 |
|---|---|---|---|---|---|---|---|
|
|
|
|
| 109 |
`scripts/collection/collect_hmmwv_crm_dataset.py`, `src/nedm/arm_data.py`,
|
| 110 |
`src/nedm/tracked_vehicle_data.py`).
|
| 111 |
|
| 112 |
+
#### Splits
|
| 113 |
|
| 114 |
Train/val is decided **per episode at collection time** and stored in the `split` column:
|
| 115 |
`sha1(episode_id)[:8] / 0xFFFFFFFF < validation_ratio → val` (ratio 0.20 for the HMMWV sets,
|
| 116 |
0.15 for `arm` and `tracked`). Whole episodes stay together; the assignment depends only on the
|
| 117 |
episode id, so it is stable under re-sharding. `train` and `val` files never share an episode.
|
| 118 |
|
| 119 |
+
### Layout
|
| 120 |
|
| 121 |
```
|
| 122 |
raw/<config>/train/<shard>.parquet transitions, one file per raw collection shard
|
|
|
|
| 135 |
identifiers are dictionary-encoded strings). Files are zstd-compressed with BYTE_STREAM_SPLIT
|
| 136 |
float encoding and ≤ 262,144-row row groups.
|
| 137 |
|
| 138 |
+
#### Column groups
|
| 139 |
|
| 140 |
**HMMWV (`hmmwv_flat`, `hmmwv_bumpy`, `hmmwv_crm` — identical 105 columns).** Units are in the
|
| 141 |
names (`_m`, `_mps`, `_mps2`, `_rad`, `_radps`, `_n`, `_nm`); world frame is Chrono's ISO
|
|
|
|
| 171 |
tire channels, plus `left_sprocket_speed_radps`, `right_sprocket_speed_radps`. The paper's 3-D
|
| 172 |
state is `vel_body_x_mps, vel_body_y_mps, yaw_rate_radps`; action is the driver triple.
|
| 173 |
|
| 174 |
+
#### `episodes.parquet` and the metadata bundle
|
| 175 |
|
| 176 |
`episodes.parquet` flattens each episode's `dataset_index.json` entry and its JSON sidecar
|
| 177 |
(nested values are JSON strings): `episode_id`, `split`, `scenario_family`, `rows`,
|
|
|
|
| 185 |
terrain and solver settings), every per-episode sidecar, and the shard-plan manifests. It is what
|
| 186 |
lets the release be turned back into the collectors' original directory tree (below).
|
| 187 |
|
| 188 |
+
### Loading
|
| 189 |
|
| 190 |
**Streaming with 🤗 `datasets`** (no download of the 44 GB flat set required):
|
| 191 |
|
|
|
|
| 221 |
The rehydrated CSVs carry the float32 values the trainer uses; caches rebuilt from them are
|
| 222 |
bit-identical to the ones in `processed/` (this is checked in the release validation).
|
| 223 |
|
| 224 |
+
### Processed caches
|
| 225 |
|
| 226 |
| Cache | Trained model | State | Action | Transitions (train / val) | Size |
|
| 227 |
|---|---|---|---|---|---|
|
|
|
|
| 236 |
(`state_fields`, `action_fields`, `dt_s`, train-split mean/std used for normalisation). Values are
|
| 237 |
raw physical units; the model applies the statistics.
|
| 238 |
|
| 239 |
+
### Known limitations
|
| 240 |
|
| 241 |
* `hmmwv_bumpy` episodes are short (mean 3.3 k rows) because 78 % end on the 0.9 × 500 m
|
| 242 |
keep-in guard; the regime is meant as a test set.
|
|
|
|
| 246 |
fluid–solid interaction force.
|
| 247 |
* Simulation is deterministic and noise-free; there is no sensor model.
|
| 248 |
|
| 249 |
+
### Citation
|
| 250 |
|
| 251 |
```bibtex
|
| 252 |
@article{zhang2026abstraction,
|
|
|
|
| 259 |
|
| 260 |
License: BSD-3-Clause (same as the code). Simulation assets are Project Chrono's HMMWV and M113
|
| 261 |
models; the LRV arm geometry is in the code repository (`src/arm_model/`).
|
| 262 |
+
|
| 263 |
+
## Part B: traversing study
|
| 264 |
+
|
| 265 |
+
### What the study is
|
| 266 |
+
|
| 267 |
+
Follow-on work to the paper, not part of it. A vehicle in [Project Chrono](https://projectchrono.org) has to reach a
|
| 268 |
+
goal across generated hill-and-crater terrain, on rigid ground and on CRM deformable soil. A neural network trained on
|
| 269 |
+
thousands of recorded Chrono drives predicts, for a candidate route and speed profile, where along the route the
|
| 270 |
+
vehicle is likely to fail: roll back on a climb, stall, dig into the soil or tip over. A route search picks the route
|
| 271 |
+
with the lowest predicted risk, and Chrono's stock PID path follower drives it. Separately, a route tracker is trained
|
| 272 |
+
with PPO inside a learned neural reduced dynamics model (NRD) and compared with that PID follower. The milestones, how
|
| 273 |
+
they were measured and what they do not show are documented in
|
| 274 |
+
[`traversing/` of the code repository](https://github.com/uwsbel/NeDM/tree/main/traversing). This part holds what
|
| 275 |
+
reproduces those results: the recorded drives, the exact training files built from them, every trained model behind a
|
| 276 |
+
result-table column, and the evaluation inputs and per-drive records. Every file is a byte-exact copy of the study's
|
| 277 |
+
record, or a deterministic archive of such copies; nothing was re-encoded.
|
| 278 |
+
|
| 279 |
+
Terms used below:
|
| 280 |
+
|
| 281 |
+
- **Milestones** (the `milestone` labels of each item):
|
| 282 |
+
- `m1`: a Chrono depth camera in the planning loop (live multi-waypoint missions, rigid ground);
|
| 283 |
+
- `m2`: one risk model for rigid ground and soil, not told which one it is on (HMMWV on f104);
|
| 284 |
+
- `m3`: a learned route tracker against Chrono's PID follower;
|
| 285 |
+
- `m4a`: the HMMWV planner on arenas it never trained on;
|
| 286 |
+
- `m4b`: the pipeline on other vehicles (Gator, Polaris; the M113 only in a smoke test).
|
| 287 |
+
- **Arenas.** f104 is the 80 m x 80 m hill-and-crater arena where most training drives were recorded. g203, g217 and
|
| 288 |
+
g228 are sibling arenas from the same terrain generator: the depth-camera model of m1 trained on all three (g216 and
|
| 289 |
+
g231 were its held-out arenas), the m4a planners on g203 and g228 (g217 was their development arena). The unseen
|
| 290 |
+
arenas are eight more made with new seeds and never used in training.
|
| 291 |
+
- **Ground.** Rigid means a rigid heightmap. Soil means Chrono's CRM particle soil: a 0.24 m layer over a rigid floor,
|
| 292 |
+
one setting throughout.
|
| 293 |
+
- **Groups and route ids.** A start/goal pair is a group. Each group has several routes: designed routes (sideways
|
| 294 |
+
offsets x speed profiles, ids ending `_route_NN`) and planner-style routes (`_op_NN`), for example
|
| 295 |
+
`f104_v2_group_0000_route_03`. The soil collections and the other vehicles reuse these ids; the Gator and Polaris
|
| 296 |
+
drive folders carry a `gator__` or `polaris__` prefix.
|
| 297 |
+
- **Drive.** One Chrono run of one route with the stock PID path follower, after a 0.8 s braked settle, recorded every
|
| 298 |
+
50 ms. It stops at the goal (2.5 m radius), on rollover, on leaving the terrain, on prolonged blockage, on soil when a
|
| 299 |
+
wheel digs through the whole layer, or at the 120 s horizon.
|
| 300 |
+
|
| 301 |
+
### Layout
|
| 302 |
+
|
| 303 |
+
```
|
| 304 |
+
traversing/README.md short index: every item with bundle, milestone, packing, files, bytes, restore path
|
| 305 |
+
traversing/release_manifest.json every file (bytes, sha256, item, kind) and every item (bundle, milestone, ...)
|
| 306 |
+
traversing/models/<item>/... trained networks with their training records
|
| 307 |
+
traversing/evaluation/<item>/... evaluation suites, locked route picks, task lists, per-drive records and results
|
| 308 |
+
traversing/processed/<item>/... the exact training files the models were fitted on
|
| 309 |
+
traversing/raw/<item>/... the recorded Chrono drives the training files were built from
|
| 310 |
+
traversing/assets/<item>/... arena heightmaps, the f104 terrain grid, the soil setting and vehicle variants
|
| 311 |
+
```
|
| 312 |
+
|
| 313 |
+
Each item is packed in one of two ways:
|
| 314 |
+
|
| 315 |
+
- **`files`.** Every file is uploaded as is, at `traversing/<bundle>/<item>/<path>`, where `<path>` is its path under
|
| 316 |
+
the item's source folder. Used for models, single large training files and small sets of evaluation files.
|
| 317 |
+
- **`tar`.** Folders of many small files (mostly drive folders) go into `part-00000.tar.gz`, `part-00001.tar.gz`, ...
|
| 318 |
+
of about 1 GiB of uncompressed files each; a drive folder never spans two shards. Each member is stored under its
|
| 319 |
+
restore path relative to the code repository's root, so extracting a shard at the root of a checkout puts it in
|
| 320 |
+
place. The archives are deterministic (members sorted by path, time stamps 0, owner 0, mode 0644, regular files
|
| 321 |
+
only, PAX format, gzip level 6 without name or time). Next to the shards:
|
| 322 |
+
- `index.csv.gz`: one row per member, with columns `path, bytes, sha256, shard`;
|
| 323 |
+
- `ids.txt.gz`: the list of drive folders, for items selected by such a list;
|
| 324 |
+
- `episodes.csv.gz`: raw collections only, one row per drive (see below).
|
| 325 |
+
|
| 326 |
+
`release_manifest.json` is the reference for everything under `traversing/`:
|
| 327 |
+
|
| 328 |
+
- `source_commit`: the experiment commit the files come from,
|
| 329 |
+
[`901d6c9`](https://github.com/uwsbel/NeDM/tree/901d6c9423a16c0fafc3d60056065415d5a725f2);
|
| 330 |
+
- `items`: per item its bundle, milestones, description, `needed_for` (`rerun`, `recount` or `retrain`), packing,
|
| 331 |
+
`restore_root`, member count and bytes, Hub files, parameter counts of its checkpoints (`known_params`), and the other
|
| 332 |
+
items it needs (`requires`);
|
| 333 |
+
- `files`: per Hub file its bytes, SHA256, item and kind (`file`, `tar_shard`, `index` or `episodes`). Shards also
|
| 334 |
+
record `tar_bytes` and `tar_sha256` of the uncompressed stream, since the compressed bytes depend on the zlib
|
| 335 |
+
version;
|
| 336 |
+
- `path_remap`: how paths written inside released task files map to restore paths, and the symbolic links the
|
| 337 |
+
planners expect (the release holds no links; `ln -s` recreates them from this list).
|
| 338 |
+
|
| 339 |
+
The code repository keeps a copy pinned to a Hub commit, `traversing/manifests/hf_release_manifest.json`, whose
|
| 340 |
+
`hf_revision` is `6620faead5225ac9aa5ae8ab19bc2ef2db38a863`.
|
| 341 |
+
|
| 342 |
+
### Items
|
| 343 |
+
|
| 344 |
+
<!-- BEGIN fill_card.py: items (generated from traversing/release_manifest.json; do not edit by hand) -->
|
| 345 |
+
|
| 346 |
+
**119 items, 732 files, 56.6 GB** (models 30 items, 206 MB; evaluation 33 items, 5.8 GB; processed 28 items, 33.9 GB; raw 25 items, 16.8 GB; assets 3 items, 8.1 MB). The raw collections with an episodes table hold 189,335 drives. Sizes are Hub bytes (compressed for tar shards); `release_manifest.json` has every file's exact size and SHA256. The descriptions are the manifest's: they name the column of the result tables in `traversing/results/` that each model or record stands behind, and the study's own file and model names.
|
| 347 |
+
|
| 348 |
+
#### Models
|
| 349 |
+
|
| 350 |
+
| Item | Milestone | What it is | Files | Size |
|
| 351 |
+
|---|---|---|---|---|
|
| 352 |
+
| `m1_nav_direct_depth_risk_ensemble` | m1 | Navigation risk ensemble: 3 seeds (s0-s2) of the direct-depth corridor model (channels range_abs, sec1, speed, valid) that scored every candidate route in all 120 navigation missions (all 4 decision-timing arms). | 4 | 3.1 MB |
|
| 353 |
+
| `m2_joint_station_history_transformer` | m2 | Joint station+history transformer trained on the same rows (deploy_a3_haux_txjoint, 5 seeds s0-s4) behind transformer_1s, transformer_0p5s and transformer_0p5s_grad (soil). | 6 | 8.4 MB |
|
| 354 |
+
| `m2_oracle_world_label_model` | m2 | Shared model given the true world label (T, 5 seeds s0-s4) behind oracle_tag_standing, oracle_tag_3s, oracle_tag_1s, oracle_tag_0p5s. | 6 | 5.3 MB |
|
| 355 |
+
| `m2_pooled_no_label_no_history_model` | m2 | Shared model with neither history nor label (P, 5 seeds s0-s4) behind pooled_3s, pooled_1s, pooled_0p5s. | 6 | 5.3 MB |
|
| 356 |
+
| `m2_rigid_only_planner_standing_start` | m2 | Earlier deployed rigid-only planner (night-2 N2, 5 seeds s0-s4) behind column specialist_rigid_standing (study arm Srigid, sampling arm B) in both m2 tables. | 6 | 5.2 MB |
|
| 357 |
+
| `m2_rigid_specialist_same_rows` | m2 | Single-world rigid model trained on the same re-anchored rows (Sp_rigid, 5 seeds s0-s4) behind specialist_rigid_3s. | 6 | 5.3 MB |
|
| 358 |
+
| `m2_shared_history_early_rows_final_model` | m2 | Model of the FINAL label-free configuration: the history design retrained with extra early-decision and moving-branch rows (deploy_a1_haux_gru, 5 seeds s0-s4). Behind shared_hist_early_rows_1s, shared_hist_early_rows_0p5s (soil and rigid) and shared_hist_early_rows_0p5s_grad (the 780/800 soil and 800/800 rigid headline). | 6 | 6.2 MB |
|
| 359 |
+
| `m2_shared_history_model` | m2 | Shared rigid/soil history model H (CNN-GRU + 2 s history encoder, 5 seeds s0-s4) behind shared_hist_standing, shared_hist_3s, shared_masked_3s (history masked at decision), shared_hist_1s and shared_hist_0p5s. | 6 | 5.5 MB |
|
| 360 |
+
| `m2_soil_only_planner_standing_start` | m2 | Earlier deployed soil-only planner (CRM_N2, 5 seeds s0-s4) behind column specialist_soil_standing (study arm Scrm, sampling arm B) in both m2 tables. | 6 | 5.2 MB |
|
| 361 |
+
| `m2_soil_specialist_same_rows` | m2 | Single-world soil model trained on the same re-anchored rows (Sp_crm, 5 seeds s0-s4) behind specialist_soil_3s and specialist_soil_1s. | 6 | 5.3 MB |
|
| 362 |
+
| `m3_tracker_learned_dynamics_model` | m3 | Tag-conditioned NRD vehicle-dynamics model (causal transformer, 17-D state, 50 ms steps, 8x8 terrain crop) inside which the round-2 PPO policy was trained; not used when driving. | 2 | 19.7 MB |
|
| 363 |
+
| `m3_tracker_round2_numpy_actor` | m3 | Deployed round-2 learned tracker: the PPO actor exported to NumPy (158-D observation, 3 actions), run without torch or the NRD model; behind column nrd_policy_v2 of m3_tracker_routes.csv. | 2 | 2.0 MB |
|
| 364 |
+
| `m3_tracker_round2_ppo_checkpoint` | m3 | Final PPO torch checkpoint (iteration 999) of the round-2 tracker, the source of the NumPy actor. Not needed to re-drive (the actor is enough) and not a training input, but it is the only artefact with which the actor export and parity check can be repeated. | 3 | 5.9 MB |
|
| 365 |
+
| `m4a_rigid_f104_only_ens1` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M1a_rigid_deploy, behind columns f104_only_ens1_fixed2mps and _speedfree (rigid, M1a). The same bytes are aliased as H, the HMMWV-trained rigid model of the Gator study. | 8 | 7.8 MB |
|
| 366 |
+
| `m4a_rigid_f104_only_ens2` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s5-s9), M1b_rigid_deploy, behind columns f104_only_ens2_fixed2mps and _speedfree (rigid, M1b). | 8 | 7.8 MB |
|
| 367 |
+
| `m4a_rigid_three_arenas_all_data` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), A3_rigid_deploy, behind columns three_arenas_all_data_fixed2mps and _speedfree (rigid, A3). | 8 | 12.5 MB |
|
| 368 |
+
| `m4a_rigid_three_arenas_same_total_ens1` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M3a_rigid_deploy, behind columns three_arenas_same_total_ens1_fixed2mps and _speedfree (rigid, M3a). | 8 | 8.0 MB |
|
| 369 |
+
| `m4a_rigid_three_arenas_same_total_ens2` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s5-s9), M3b_rigid_deploy, behind columns three_arenas_same_total_ens2_fixed2mps and _speedfree (rigid, M3b). | 8 | 8.0 MB |
|
| 370 |
+
| `m4a_rigid_two_arenas_same_total` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M2_rigid_deploy, behind columns two_arenas_same_total_fixed2mps and _speedfree (rigid, M2). | 8 | 7.9 MB |
|
| 371 |
+
| `m4a_soil_f104_only_ens1` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M1a_soil_deploy, behind column f104_only_ens1 (soil table, study M1a). The same bytes are aliased as H_soil, the Gator study's stage-1 HMMWV soil model. | 8 | 6.3 MB |
|
| 372 |
+
| `m4a_soil_f104_only_ens2` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s5-s9), M1b_soil_deploy, behind column f104_only_ens2 (soil table, study M1b). | 8 | 6.3 MB |
|
| 373 |
+
| `m4a_soil_three_arenas_all_data` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), A3_soil_deploy, behind column three_arenas_all_data (soil, A3). | 8 | 7.1 MB |
|
| 374 |
+
| `m4a_soil_three_arenas_same_total_ens1` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M3a_soil_deploy, behind column three_arenas_same_total_ens1 (soil, M3a). | 8 | 6.3 MB |
|
| 375 |
+
| `m4a_soil_three_arenas_same_total_ens2` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s5-s9), M3b_soil_deploy, behind column three_arenas_same_total_ens2 (soil, M3b). | 8 | 6.3 MB |
|
| 376 |
+
| `m4a_soil_two_arenas_same_total` | m4a | CNN-GRU route-risk ensemble, 5 seeds (s0-s4), M2_soil_deploy, behind column two_arenas_same_total (soil, M2). | 8 | 6.3 MB |
|
| 377 |
+
| `m4b_gator_rigid_own_model` | m4b | Gator-trained rigid ensemble on the Gator's 24,000 f104 rigid drives (CNN-GRU, 5 seeds s0-s4, G_rigid_deploy). Behind the README bar-table Gator rigid row (800/800 on f104; arms G_free at free speed and the G_fixed2 addendum); not a compact-table column. | 8 | 7.6 MB |
|
| 378 |
+
| `m4b_gator_soil_own_model_full_data` | m4b | Gator-trained soil ensemble on all 15,235 Gator f104 soil drives (CNN-GRU, 5 seeds s0-s4, G_full_soil_deploy), behind gator_own_model_sampling (Gfull_free_gator) and gator_own_model_sampling_grad (Gfull_grad_gator). | 8 | 6.8 MB |
|
| 379 |
+
| `m4b_gator_soil_own_model_tiers0to6` | m4b | Gator-trained soil ensemble on only the first 7 routes of each group (8,399 drives; CNN-GRU, 5 seeds s0-s4, G_soil_deploy), behind gator_own_model_tiers0to6_sampling (G_free_gator). | 8 | 6.2 MB |
|
| 380 |
+
| `m4b_hmmwv_soil_model_full_data` | m4b | HMMWV-trained soil ensemble on the full HMMWV f104 soil file, the same ids as the Gator (CNN-GRU, 5 seeds s0-s4, H_full_soil_deploy). Behind gator_hmmwv_model_sampling (Hfull_free_gator), hmmwv_own_model_sampling (Hfull_free) and hmmwv_own_model_sampling_grad (Hfull_grad_hmmwv). | 8 | 6.9 MB |
|
| 381 |
+
| `m4b_polaris_soil_own_model_full_data` | m4b | Final Polaris soil ensemble on all 15,229 valid Polaris f104 soil drives (CNN-GRU, 5 seeds s0-s4, polaris_full_soil_deploy). Behind polaris_own_model_sampling_grad and polaris_own_model_sampling on f104 and on the 8 unseen arenas; polaris_corrected_driveline_grad_routes re-drives the polaris_grad routes. | 7 | 5.4 MB |
|
| 382 |
+
|
| 383 |
+
#### Evaluation
|
| 384 |
+
|
| 385 |
+
| Item | Milestone | What it is | Files | Size |
|
| 386 |
+
|---|---|---|---|---|
|
| 387 |
+
| `nav_corrected_mission_outcomes` | m1 | Corrected navigation read-out after the rescue-route fix: summary.json and the 120 per-mission outcome files (30 missions x 4 decision-timing arms) behind m1_navigation_missions.csv. | 121 | 377 kB |
|
| 388 |
+
| `nav_corrected_run_folders` | m1 | Per-mission decisions, candidate routes and trajectory of the 120 corrected navigation drives. | 3 | 33.8 MB |
|
| 389 |
+
| `nav_missions_and_task_list` | m1 | The 30 frozen waypoint missions and the 120-row task list (mission x decision-timing arm). | 31 | 67.1 kB |
|
| 390 |
+
| `f104_eval_suite_800` | m2, m4a, m4b | The frozen 800-pair f104 start/goal suite: suite definition, lock file, 800 case files, the designed route of each group and the locked standing-start task lists. | 2 | 5.7 MB |
|
| 391 |
+
| `planner_map_f104` | m2, m4a, m4b | Whole-arena overhead depth map of f104 on which every f104 planner scores its candidate routes. | 2 | 4.1 MB |
|
| 392 |
+
| `shared_model_decision_states` | m2 | Inputs of the moving-start decisions: straight approach routes, the approach drives, and the recorded decision poses and 2 s history arrays at 3 s, 1 s and 0.5 s, with the approach task files. | 2 | 53.5 MB |
|
| 393 |
+
| `shared_model_drive_folders_rigid` | m2 | Per-drive folders behind every cell of m2_shared_risk_rigid.csv (9,469 distinct drives for 9,600 cells). | 3 | 435 MB |
|
| 394 |
+
| `shared_model_drive_folders_soil` | m2 | Per-drive folders behind every cell of m2_shared_risk_soil.csv (18,027 distinct drives for 18,400 cells). | 4 | 903 MB |
|
| 395 |
+
| `shared_model_drive_task_lists` | m2 | Task lists that drove the milestone-2 table arms (standing start, 3 s, 1 s, 0.5 s and final), with the two standing-start lock files; those two locks cover unreleased route copies and cannot be checked from the release. | 14 | 25.8 MB |
|
| 396 |
+
| `shared_model_results_early_and_final` | m2 | Per-pair results of the 1 s and 0.5 s early-decision arms and of the final gradient-refined configuration on soil and rigid ground. | 4 | 16.7 MB |
|
| 397 |
+
| `shared_model_results_standing_and_3s` | m2 | Per-pair results of the standing-start arms and the 3 s-approach arms, soil and rigid. | 4 | 10.6 MB |
|
| 398 |
+
| `shared_model_route_picks_3s` | m2 | Locked planner output of the 3 s-approach arms. | 2 | 45.8 MB |
|
| 399 |
+
| `shared_model_route_picks_early` | m2 | Locked planner output of the 1 s and 0.5 s early-decision arms (sampling search only), including the rigid 0.5 s arm. | 2 | 52.7 MB |
|
| 400 |
+
| `shared_model_route_picks_final` | m2 | Locked planner output of the final gradient-refined arms after the 0.5 s approach (soil and rigid). | 2 | 38.3 MB |
|
| 401 |
+
| `shared_model_route_picks_standing` | m2 | Locked planner output of the standing-start arms: per-pair pick records with candidate scores, route files, task list, run summary and lock file. | 2 | 47.6 MB |
|
| 402 |
+
| `f104_reference_routes_and_cases` | m3, m4b | f104 start/goal case files and reference routes that released drive tasks point at: the 423 routes and 55 cases the tracker drives follow, and the 147 routes (88 designed, 59 on-policy) and 27 cases of the vehicle smoke test: the 144 sample-A routes (86 designed, 58 planner proposals) plus the 3 routes of the bit-identity check. | 2 | 1.4 MB |
|
| 403 |
+
| `tracker_drive_folders` | m3 | Drive folders of the 2,538 tracker table rows (native PID, held PID and learned tracker on 423 routes, both grounds). | 3 | 141 MB |
|
| 404 |
+
| `tracker_results_suite_tasks` | m3 | Per-route tracker results (both grounds), the 423-route tracking suite with strata, and the task lists that drove the PID and learned-tracker arms. | 7 | 6.2 MB |
|
| 405 |
+
| `planner_maps_new_arenas` | m4a, m4b | Overhead depth maps of the 11 other planner arenas (g203, g217, g228 and the 8 unseen test arenas) and the map checks. | 23 | 45.2 MB |
|
| 406 |
+
| `unseen_and_vehicle_picks_lock_files` | m4a, m4b | Top-level pick lock files and job lists of the unseen-arena, Gator and Polaris evaluations. | 23 | 567 kB |
|
| 407 |
+
| `unseen_arena_drive_folders_rigid` | m4a | Drive folders behind the 37,100 cells of m4_unseen_arenas_hmmwv_rigid.csv (36,653 distinct drives). | 5 | 1.9 GB |
|
| 408 |
+
| `unseen_arena_drive_folders_soil` | m4a, m4b | Drive folders behind the 9,000 non-empty cells of m4_unseen_arenas_hmmwv_soil.csv (8,746 distinct drives). | 3 | 411 MB |
|
| 409 |
+
| `unseen_arena_drive_task_lists` | m4a, m4b | Task lists that drove the HMMWV soil and rigid unseen-arena evaluations (including the Gator rigid f104 drives). | 18 | 31.4 MB |
|
| 410 |
+
| `unseen_arena_eval_indexes` | m4a, m4b | Per-drive evaluation indexes of the unseen-arena study (soil and rigid; the rigid index also holds the Gator rigid f104 rows) and the task-to-arm mapping files. | 11 | 80.5 MB |
|
| 411 |
+
| `unseen_arena_route_picks` | m4a, m4b | Locked planner picks of every HMMWV arm of the two unseen-arena tables, on the 8 unseen arenas, f104, the held-out and dev arenas (236 arm x arena folders). | 3 | 228 MB |
|
| 412 |
+
| `unseen_arena_suites` | m4a, m4b | Frozen start/goal suites of the 8 unseen test arenas, held-out g203/g228 and dev g217: case files, designed routes, per-suite lock files, manifests, the f104 in-distribution list, the declared soil subset and the spread-headroom group list. | 2 | 43.7 MB |
|
| 413 |
+
| `gator_rigid_f104_drive_folders` | m4b | Drive folders of the Gator rigid f104 read-out in the README bar table: the Gator-trained planner (800/800) and the straight route at 6 m/s. | 3 | 68.7 MB |
|
| 414 |
+
| `polaris_unseen_drive_folders` | m4b | Drive folders behind the 3,000 cells of m4_polaris_unseen_soil.csv (2,618 distinct drives). | 3 | 183 MB |
|
| 415 |
+
| `vehicle_drive_task_lists` | m4b | Task lists of the Gator/HMMWV all-data drives, Polaris on f104, the gradient-refined Gator/HMMWV drives, Polaris on the unseen arenas and the smoke test. | 16 | 6.6 MB |
|
| 416 |
+
| `vehicle_eval_indexes_and_smoke_extract` | m4b | Per-drive indexes of the Gator all-data comparison, Polaris on f104 and Polaris on the unseen arenas, their task-mapping files, the per-route smoke and collection extract, and the definition of smoke sample A. | 11 | 54.0 MB |
|
| 417 |
+
| `vehicle_f104_drive_folders` | m4b | Drive folders behind the 9,600 cells of m4_vehicles_f104_soil.csv (9,243 distinct drives; Gator, HMMWV and Polaris variants). | 3 | 702 MB |
|
| 418 |
+
| `vehicle_route_picks` | m4b | Locked picks of the vehicle arms: Gator own models, HMMWV-trained model, gradient refinements, Polaris on f104 and on the 8 unseen arenas, and the Gator rigid planner. | 3 | 86.2 MB |
|
| 419 |
+
| `vehicle_smoke_test_drives` | m4b | Complete drive folders of the vehicle smoke test on sample A (144 f104 soil routes) for the 7 variants with their own folders (three Polaris driveline/wheel variants, stock and re-geared M113, Gator at 1 ms and 0.5 ms steps), plus the failure record of the one crashed drive. The polaris_stock and hmmwv_stored columns of m4_vehicle_smoke.csv come from drives released in the raw collections polaris_f104_soil_collection_runs and hmmwv_soil_f104_collection (not required: about 3.7 GB). | 3 | 148 MB |
|
| 420 |
+
|
| 421 |
+
#### Processed training files
|
| 422 |
+
|
| 423 |
+
| Item | Milestone | What it is | Files | Size |
|
| 424 |
+
|---|---|---|---|---|
|
| 425 |
+
| `nav_depth_corridors_arena_g203` | m1 | Navigation corridors on sibling training arena g203. 4,925 rows. | 1 | 366 MB |
|
| 426 |
+
| `nav_depth_corridors_arena_g216_heldout` | m1 | Navigation corridors on held-out arena g216 (read by the trainer for its held-out route-choice score, not fitted). 4,956 rows. | 1 | 368 MB |
|
| 427 |
+
| `nav_depth_corridors_arena_g217` | m1 | Navigation corridors on sibling training arena g217. 5,008 rows. | 1 | 372 MB |
|
| 428 |
+
| `nav_depth_corridors_arena_g228` | m1 | Navigation corridors on sibling training arena g228. 4,937 rows. | 1 | 367 MB |
|
| 429 |
+
| `nav_depth_corridors_arena_g231_heldout` | m1 | Navigation corridors on held-out arena g231 (read by the trainer for its held-out score, not fitted). 4,889 rows. | 1 | 363 MB |
|
| 430 |
+
| `nav_depth_corridors_f104_extra_drives` | m1 | Navigation corridors from f104 gen_v1 / sensor_v1 test drives (fitted; f104 is a training arena). 3,111 rows. | 1 | 231 MB |
|
| 431 |
+
| `nav_depth_corridors_f104_training_routes` | m1 | Direct-depth route-corridor tensors for the navigation risk ensemble (matched_Dabs, 3 seeds): all 36,199 f104 rigid training routes (labels reused from the merged station set via R/sensor_v1/labels_station_ds_all.npz, corridors re-sampled from the back-projected depth grid). 36,199 rows. | 1 | 2.7 GB |
|
| 432 |
+
| `rigid_specialist_training_rows` | m2 | Training rows of the earlier rigid-only f104 planner N2 (column specialist_rigid_standing, arm Srigid): merged night-1 + night-2 rigid route set, 36,199 routes / 2,700 groups (train 33,840, val 764, test 1,595; 31,851 fitted per N2_meta.json). | 1 | 1.1 GB |
|
| 433 |
+
| `shared_model_baseline_reanchored_rows` | m2 | Re-anchored rigid + soil rows with a 2 s history on the f104 route ids: 115,868 rows (105,193 fitted). Training file of the generalist ensembles H, P, T, Sp_crm, Sp_rigid (columns oracle_tag_*, shared_hist_standing/3s/1s/0p5s, pooled_*, shared_masked_3s, specialist_soil_3s/1s, specialist_rigid_3s). | 1 | 2.1 GB |
|
| 434 |
+
| `shared_model_final_anchor_k60_rows` | m2 | Final shared-model training file 3 of 3: decision rows at frame 60 (3 s), the k=60 slice of anchor_k40_60_80.npz (14,972 rigid + 14,727 soil). 29,699 rows, 26,936 fitted. | 1 | 541 MB |
|
| 435 |
+
| `shared_model_final_reanchored_plus_branch_rows` | m2 | Final shared-model training file 1 of 3: mixed_reanchor rows plus 4,614 moving-prefix branch rows of both worlds. 120,482 rows, 109,244 fitted. | 1 | 2.2 GB |
|
| 436 |
+
| `shared_model_final_short_anchor_rows` | m2 | Final shared-model training file 2 of 3: early decision rows at frames 10/20/30 of every rigid and soil f104 episode. 89,998 rows, 81,634 fitted. | 1 | 1.6 GB |
|
| 437 |
+
| `soil_specialist_training_rows` | m2 | Training rows of the earlier soil-only f104 planner CRM_N2 (column specialist_soil_standing, arm Scrm): one row per validated HMMWV soil route (15,235); 13,821 fitted per CRM_N2_deploy.json. | 1 | 471 MB |
|
| 438 |
+
| `tracker_dynamics_training_cache` | m3 | Per-episode 50 ms state/action/pose/power cache of 43,235 f104 episodes (rigid 26,500 = 24,000 production + 1,500 perturbed + 1,000 round-1 policy harvest; soil 16,735 = 15,235 collection + 1,500 perturbed) plus cache_manifest.json and build_report.json. | 3 | 2.0 GB |
|
| 439 |
+
| `tracker_group_split_file` | m3 | Night-2 soil twin dataset whose group/split arrays define the 1,200-group split (1,089/56/55) used by the cache builder and cross-checked by the NRD trainer. | 1 | 169 MB |
|
| 440 |
+
| `terrain_rigid_subset_f104_only` | m4a | HMMWV rigid, f104 only, 1,089 groups. 93,397 rows, 84,787 fitted. Trained M1a and M1b (columns f104_only_ens1/ens2 _fixed2mps/_speedfree); byte-identical to H_f104_hmmwv_rigid.npz (Gator-study H). | 1 | 1.7 GB |
|
| 441 |
+
| `terrain_rigid_subset_three_arenas_all` | m4a | HMMWV rigid, all data: f104 1,089 + g203 1,083 + g228 1,063 groups. 278,735 rows, 250,490 fitted. Trained A3 (column three_arenas_all_data). | 1 | 5.0 GB |
|
| 442 |
+
| `terrain_rigid_subset_three_arenas_same_total` | m4a | HMMWV rigid, f104 + g203 + g228, same total. 112,726 rows, 84,481 fitted. Trained M3a and M3b (columns three_arenas_same_total_ens1/ens2). | 1 | 2.0 GB |
|
| 443 |
+
| `terrain_rigid_subset_two_arenas` | m4a | HMMWV rigid, f104 + g203, same total. 101,916 rows, 84,310 fitted. Trained M2 (column two_arenas_same_total). | 1 | 1.8 GB |
|
| 444 |
+
| `terrain_soil_subset_f104_only` | m4a | HMMWV soil, f104 only, 1,089 training groups, tiers 0-6. 32,152 rows, 29,210 fitted. Trained M1a_soil and M1b_soil (columns f104_only_ens1/ens2, soil) and is byte-identical to H_f104_hmmwv_soil.npz (Gator-study H_soil). | 1 | 583 MB |
|
| 445 |
+
| `terrain_soil_subset_three_arenas_all` | m4a | HMMWV soil, all data: f104 1,089 + g203 559 + g228 520 groups. 64,400 rows, 57,898 fitted. Trained A3_soil (column three_arenas_all_data). | 1 | 1.2 GB |
|
| 446 |
+
| `terrain_soil_subset_three_arenas_same_total` | m4a | HMMWV soil, f104 + g203 + g228, 363 groups each. 35,573 rows, 29,071 fitted. Trained M3a_soil and M3b_soil (columns three_arenas_same_total_ens1/ens2). | 1 | 644 MB |
|
| 447 |
+
| `terrain_soil_subset_two_arenas` | m4a | HMMWV soil, f104 + g203, 545 groups each. 33,227 rows, 29,048 fitted. Trained M2_soil (column two_arenas_same_total). | 1 | 603 MB |
|
| 448 |
+
| `gator_rigid_subset_all` | m4b | Gator rigid on f104, all 24,000 route ids. 93,551 rows, 84,922 fitted. Trained G, the Gator rigid planner cited in the README bar table (800/800). | 1 | 1.7 GB |
|
| 449 |
+
| `gator_soil_subset_all` | m4b | Gator soil on f104, all 15,235 validated ids (tiers 0-12). 55,826 rows, 50,822 fitted. Trained G_full_soil (columns gator_own_model_sampling, gator_own_model_sampling_grad). | 1 | 1.0 GB |
|
| 450 |
+
| `gator_soil_subset_tiers0to6` | m4b | Gator soil on f104, tiers 0-6 (8,399 drives). 30,827 rows, 28,057 fitted. Trained G_soil (column gator_own_model_tiers0to6_sampling). | 1 | 557 MB |
|
| 451 |
+
| `hmmwv_soil_subset_gator_matched_all` | m4b | HMMWV soil on f104, exactly the route ids the Gator validated (all tiers). 58,268 rows, 52,923 fitted. Trained H_full_soil (columns gator_hmmwv_model_sampling, hmmwv_own_model_sampling, hmmwv_own_model_sampling_grad). | 1 | 1.1 GB |
|
| 452 |
+
| `polaris_soil_subset_all` | m4b | Polaris soil on f104, all 15,229 valid ids (tiers 0-12). 57,228 rows, 52,021 fitted. Trained polaris_full_soil (all Polaris planner columns on f104 and the unseen arenas). | 2 | 1.0 GB |
|
| 453 |
+
|
| 454 |
+
#### Raw collections
|
| 455 |
+
|
| 456 |
+
| Item | Milestone | What it is | Files | Size |
|
| 457 |
+
|---|---|---|---|---|
|
| 458 |
+
| `hmmwv_rigid_f104_early_waves` | m1, m2 | Seven small early HMMWV rigid f104 waves whose drives are in the navigation model's 36,199-route f104 label file: prospective_v1 309, fixed_speed_v1 292, speed_c4 206, wide_v1 146, speed_c6 90, band_v1 73, showcase_v1 14. The last six make up the 821 rows the label file tags as MPPI arms. | 4 | 84.9 MB |
|
| 459 |
+
| `hmmwv_rigid_f104_first_overnight_collection` | m1, m2 | Earlier HMMWV rigid f104 production collection (1,500 groups, the first overnight wave). It is not part of the 24,000-route pool. It supplies 11,412 of the navigation model's 36,199 f104 routes and part of the rigid N2 specialist's training rows. | 4 | 681 MB |
|
| 460 |
+
| `hmmwv_rigid_f104_pool_designed` | m1, m2, m3, m4a | HMMWV rigid-ground pool on f104, designed routes: 1,200 start/goal groups x 12 routes. This is the first half of the 24,000-route rigid pool. Every later rigid f104 model and the tracker cache were built from this pool, and the Gator and Polaris re-drove its route ids. | 4 | 853 MB |
|
| 461 |
+
| `hmmwv_rigid_f104_pool_onpolicy` | m1, m2, m3, m4a | HMMWV rigid-ground pool on f104, planner-style (on-policy) routes: 1,200 groups x 8 routes. This is the second half of the 24,000-route rigid pool. | 5 | 1.1 GB |
|
| 462 |
+
| `hmmwv_rigid_multi_arena_test_drives` | m1 | HMMWV rigid test drives on f104, g203, g217 and g228 from the multi-arena study (planner, rule-based, fixed-speed and straight-route arms; 1,103 / 1,098 / 1,132 / 1,111). The navigation trainer (sensor_train_v2.py) fits every row from a training arena whatever its source, so these are training rows. | 4 | 250 MB |
|
| 463 |
+
| `hmmwv_rigid_sensor_study_test_drives` | m1 | HMMWV rigid drives from the sensor study's two test campaigns (sensor_v1/test 3,614 and test2 4,523) on the training arenas f104, g203, g217 and g228. They enter the navigation training rows through ds_v2_gen_<arena>.npz. | 4 | 408 MB |
|
| 464 |
+
| `hmmwv_rigid_sibling_arenas_designed_routes` | m1 | HMMWV rigid designed-route drives on the navigation model's three sibling training arenas g203, g217 and g228 (1,800 each, 5,400 total). The g216 and g231 drives in the same folder are held-out evaluation only and are excluded. | 4 | 330 MB |
|
| 465 |
+
| `hmmwv_rigid_f104_branch_anchor_records` | m2 | Per-anchor records of the HMMWV rigid f104 branch drives (<episode>\_\_a4, 800 anchor states): the automatic branch choice (branch_auto.json, 789) or the skip record (skipped.json, 11) and the completion marker. The drives themselves are hmmwv_rigid_f104_branch_continuations. | 3 | 519 kB |
|
| 466 |
+
| `hmmwv_rigid_f104_branch_continuations` | m2 | HMMWV rigid drives that branch from a moving state on f104: 2,367 continuation drives (<anchor>\_\_a4\_\_c<j>; 2,358 produce rows, 9 were skipped) from 800 anchor states. The per-anchor records are hmmwv_rigid_f104_branch_anchor_records. | 4 | 251 MB |
|
| 467 |
+
| `hmmwv_soil_f104_branch_continuations` | m2 | HMMWV soil drives that branch from a moving state on f104: 752 anchor states x 3 = 2,256 continuation drives (<episode>\_\_c<j>). | 4 | 246 MB |
|
| 468 |
+
| `hmmwv_soil_f104_collection` | m2, m3, m4a, m4b | HMMWV soil collection on f104 (Chrono's particle-based deformable soil, 0.24 m deep over a rigid floor): 15,235 validated route ids in 1,200 groups, tiers 0-12 (9,168 designed + 6,067 planner-style). The Gator and Polaris later re-drove the same ids. | 5 | 1.4 GB |
|
| 469 |
+
| `hmmwv_rigid_tracker_held_control_drives` | m3 | HMMWV rigid drives with perturbed controls on f104 training routes: controls held for a true 50 ms with brake taps added (1,500 drives, <route>\_\_b3). | 4 | 100 MB |
|
| 470 |
+
| `hmmwv_rigid_tracker_policy_harvest` | m3 | HMMWV rigid drives by the round-1 learned tracker on 1,000 training routes (<route>\_\_harvest). They were added to the dynamics cache. | 4 | 68.9 MB |
|
| 471 |
+
| `hmmwv_soil_tracker_held_control_drives` | m3 | HMMWV soil drives with the same held-control perturbations on f104 training routes (1,500 drives, <route>\_\_b3). | 4 | 130 MB |
|
| 472 |
+
| `arena_gator_collection_task_files` | m4a, m4b | Task files passed to the collection builds of the arena and vehicle studies: the Gator rigid pool (rigid_v2), the Gator soil builds (soil_v3 for tiers 0-12, soil_v2 for tiers 0-6, which also defines the HMMWV g203/g228 soil tiers) and the HMMWV g203/g228 rigid pools (rigid_hmmwv_v1). | 4 | 103 MB |
|
| 473 |
+
| `hmmwv_rigid_g203_pool` | m4a | HMMWV rigid 24,000-route pool on the extra training arena g203 (1,200 groups x 20 routes; 14,400 route_ ids + 9,600 op_ ids). ag_build_ds.py selected all 24,000 (92,225 rows). | 6 | 2.0 GB |
|
| 474 |
+
| `hmmwv_rigid_g228_pool` | m4a | HMMWV rigid 24,000-route pool on the extra training arena g228 (93,113 rows). | 6 | 2.0 GB |
|
| 475 |
+
| `hmmwv_soil_g203_tiers0to6` | m4a | HMMWV soil drives on g203, tiers 0-6 (the first 7 routes of each of 606 start/goal groups): 4,242 driven, of which 4,240 pass QA (2 rejected as crm_qa:unstalled_break). | 4 | 385 MB |
|
| 476 |
+
| `hmmwv_soil_g228_tiers0to6` | m4a | HMMWV soil drives on g228, tiers 0-6: 4,242 driven, all validated. | 4 | 377 MB |
|
| 477 |
+
| `gator_collection_qa_records` | m4b | Recorded QA of the Gator collections and the validated-id lists passed to the subset builder: soil all tiers, soil tiers 0-6 and rigid. | 8 | 9.1 MB |
|
| 478 |
+
| `gator_f104_rigid_collection_runs` | m4b | The Gator driving the 24,000-route rigid f104 pool (1,200 groups x 20: 14,400 designed routes and 9,600 planner-style routes), 165.7 simulated hours. Chain: ci_f104_gator_rigid.npz (93,551 rows), then subset G_f104_gator_rigid.npz, then the G_rigid ensemble (e5/deploy/G). | 6 | 2.0 GB |
|
| 479 |
+
| `gator_f104_soil_collection_runs` | m4b | The Gator driving the 15,235 HMMWV f104 soil route ids (tiers 0-12: 9,168 designed routes and 6,067 planner proposals), 141.1 simulated hours. These are the source of ci_f104_gator_crm.npz (tiers 0-12, 55,826 rows), which feeds the G_full_soil ensemble. They are also the source of the stage-1 file (tiers 0-6, 8,399 runs, 30,827 rows), which feeds the G_soil ensemble. | 6 | 2.3 GB |
|
| 480 |
+
| `polaris_collection_build_record` | m4b | The Polaris stage-2 build record (15,235 candidates, 6 rejected, 15,229 selected) and the validated id list passed to the subset builder. | 2 | 552 kB |
|
| 481 |
+
| `polaris_collection_task_file` | m4b | soil_v5_polaris.json: the task file passed as --tasks-crm to the Polaris stage-2 build. Its tier 0-12 rows are the 15,235 collection ids. | 1 | 15.9 MB |
|
| 482 |
+
| `polaris_f104_soil_collection_runs` | m4b | The Polaris (stock driveline) driving the same 15,235 f104 soil ids: 15,229 valid, 93.0 simulated hours. Chain: stage-2 ci_f104_polaris_crm.npz (tiers 0-12, 57,228 rows), then polaris_full_f104_soil.npz, then the polaris_full_soil ensemble. The 144 sample-A rows of the smoke arm 'polaris' are part of this collection. | 5 | 1.7 GB |
|
| 483 |
+
|
| 484 |
+
#### Assets
|
| 485 |
+
|
| 486 |
+
| Item | Milestone | What it is | Files | Size |
|
| 487 |
+
|---|---|---|---|---|
|
| 488 |
+
| `arena_heightmaps` | m1, m2, m3, m4a, m4b | Heightmap and metadata of every arena driven in a milestone: f104 and 17 generated arenas, plus the arena family description. | 37 | 4.8 MB |
|
| 489 |
+
| `crm_soil_config_and_vehicle_variants` | m2, m3, m4a, m4b | The single deformable-soil setting (crm_main.json), the M113 smoke-test soil setting (crm_m113.json) and the Polaris and M113 vehicle definition variants. | 14 | 42.0 kB |
|
| 490 |
+
| `f104_terrain_grid` | m3 | Metric terrain grid of f104 used for the terrain crops of the vehicle-dynamics model and the tracker training environment. | 1 | 3.3 MB |
|
| 491 |
+
|
| 492 |
+
<!-- END fill_card.py: items -->
|
| 493 |
+
|
| 494 |
+
### What one drive folder contains
|
| 495 |
+
|
| 496 |
+
Raw collections and evaluation drives share one folder layout: one folder per drive, named by its route id, with a
|
| 497 |
+
suffix for drive variants (`__b3` drives with perturbed held controls, `__a4__cN` on rigid ground and `__cN` on soil for
|
| 498 |
+
continuations branching from a moving state, `__harvest` drives by the first learned tracker). Raw collections release only the files the dataset builders
|
| 499 |
+
read:
|
| 500 |
+
|
| 501 |
+
| File | Present in | Contents |
|
| 502 |
+
|---|---|---|
|
| 503 |
+
| `trajectory.npz` | every drive | the drive, one row per 50 ms interval (below) |
|
| 504 |
+
| `command_reference.npz` | every drive | the route the follower was given and the speed it was asked for (below) |
|
| 505 |
+
| `anchor_state.npz` | every drive | the vehicle at the start of recording, with a short history array (below) |
|
| 506 |
+
| `outcome.json` | every drive | how the drive ended: `status` (for example `goal_reached`, `timeout`, `rollover`, `prolonged_blockage_terminated`, `soil_breakthrough_terminated`), `goal_reached`, `elapsed_s`, goal distances and progress (m), path length (m), positive engine work (kJ), frame count, the follower's settings; on soil a `crm` block (particle counts, sinkage and slip summaries); for the Gator and Polaris a `vehicle` block |
|
| 507 |
+
| `case.json` | every drive | the start/goal task: `id`, `arena` (the arena's asset path), start pose (in `layout`), `goal_xy`, `goal_radius_m`, `horizon_s`, the route parameters of its group (sideways offsets, speeds, speed profiles), and the train/val/test `split` of its group |
|
| 508 |
+
| `episode_complete.json` | every drive | completion marker: elapsed time and the SHA256 of every file the collector wrote, including files not released |
|
| 509 |
+
| `initial_state_validation.json` | when recorded | check of the settled start: horizontal speed, roll, pitch, heading and position error against their limits |
|
| 510 |
+
| `crm_extra.npz` | soil drives | per 50 ms: chassis height `pos_z_m` and heightmap ground under it `bmp_ground_z_m` (m), chassis quaternion `quat` (e0..e3), wheel-centre heights `spindle_z_m` (m), `slip_ratio` and longitudinal soil force `fsi_force_wheel_fx_n` (N) per wheel (`wheel_order` fl, fr, rl, rr), and `tire_radius_m` |
|
| 511 |
+
| `collection_request.json` | soil drives | the collector's inputs: case and route paths with their SHA256, soil settings, seed, stop rules, and the vehicle setup for the Gator and Polaris |
|
| 512 |
+
| `vehicle_extra.npz` | Gator and Polaris | per 50 ms (`frame`): `belly_clearance_min_m`, the lowest of the points sampled on the chassis underside minus the undisturbed ground below it (m, negative = inside the soil or ground), `belly_points_below_surface`, `belly_argmin_point`, chassis vertical speed `chassis_vz_mps`; and those points in the chassis frame, `belly_points_chassis_m` |
|
| 513 |
+
| `native_height_check.json` | Gator rigid drives | Chrono's terrain height against the heightmap at 50 points (m) |
|
| 514 |
+
| `branch_route.json`, `branch_reference.json` | branch continuations | the continuation route from the moving state (waypoints, speeds, stations, headings) |
|
| 515 |
+
| `branch_auto.json`, `skipped.json` | rigid branch anchor records | the automatic branch choice at each anchor state, or why the anchor was skipped |
|
| 516 |
+
|
| 517 |
+
Evaluation drive folders are released with every recorded file except logs, and hold at least `trajectory.npz`, `outcome.json` and
|
| 518 |
+
`episode_complete.json`, usually `case.json`; navigation missions instead hold `decisions.json`, `routes.json` and
|
| 519 |
+
`trajectory.npz`.
|
| 520 |
+
|
| 521 |
+
`outcome.json` also has a `safe_goal_reached` flag. It is the collector's own flag (no obstacle contact, no 2 s
|
| 522 |
+
effortful stop, no rollover), not the study's label. The study's label counts rolling back on a climb and is
|
| 523 |
+
recomputed from `trajectory.npz`: see the outcome codes in
|
| 524 |
+
[`traversing/results/README.md`](https://github.com/uwsbel/NeDM/blob/main/traversing/results/README.md).
|
| 525 |
+
|
| 526 |
+
**`trajectory.npz`.** `T` rows, one per 50 ms interval; row `i` is measured at the start of interval `i` and the
|
| 527 |
+
action is the one applied over it.
|
| 528 |
+
|
| 529 |
+
| Key | Shape, type | Contents |
|
| 530 |
+
|---|---|---|
|
| 531 |
+
| `state` | (T, 17) float32 | the 17-number vehicle state below |
|
| 532 |
+
| `action` | (T, 3) float32 | applied steering in [-1, 1], throttle in [0, 1], braking in [0, 1] |
|
| 533 |
+
| `pose` | (T, 3) float64 | world x (m), y (m), yaw (rad) |
|
| 534 |
+
| `power_kw` | (T,) float64 | engine output torque times transmission shaft speed (kW) |
|
| 535 |
+
| `positive_work_kj_per_interval` | (T,) float64 | positive engine work in the interval (kJ) |
|
| 536 |
+
| `contact_n` | (T,) float64 | largest contact force with placed obstacles (N); the arenas have none, so it is 0 |
|
| 537 |
+
| `parked` | (T,) bool | the vehicle is at the end of the route and asked to stop |
|
| 538 |
+
| `terminal_state`, `terminal_pose`, `terminal_parked` | (17,), (3,), () | the same quantities after the last interval |
|
| 539 |
+
| `state_fields` | (17,) str | the names of the state columns |
|
| 540 |
+
| `dt_s` | () float32 | 0.05 |
|
| 541 |
+
|
| 542 |
+
The 17-number state, in this order (`STATE_FIELDS` in `src/nedm/traverse/fdm_data.py`, the preset
|
| 543 |
+
`tire_normal_force_omega_pt` of `src/nedm/training/constants.py`, at the experiment commit
|
| 544 |
+
[`901d6c9`](https://github.com/uwsbel/NeDM/blob/901d6c9423a16c0fafc3d60056065415d5a725f2/src/nedm/training/constants.py)):
|
| 545 |
+
|
| 546 |
+
| # | Field | Unit |
|
| 547 |
+
|---|---|---|
|
| 548 |
+
| 0-1 | `vel_body_x_mps`, `vel_body_y_mps`: chassis velocity in the body frame (x forward) | m/s |
|
| 549 |
+
| 2-3 | `roll_rad`, `pitch_rad` | rad |
|
| 550 |
+
| 4-6 | `roll_rate_radps`, `ang_vel_body_y_radps` (pitch rate), `yaw_rate_radps` | rad/s |
|
| 551 |
+
| 7-10 | `tire_{fl,fr,rl,rr}_force_wheel_fz_n`: normal force on each wheel | N |
|
| 552 |
+
| 11-14 | `tire_{fl,fr,rl,rr}_spindle_omega_radps`: wheel spin rate | rad/s |
|
| 553 |
+
| 15 | `engine_motor_speed_radps` | rad/s |
|
| 554 |
+
| 16 | `engine_motorshaft_torque_nm` | N m |
|
| 555 |
+
|
| 556 |
+
**`command_reference.npz`.** `interval_start_s` and `desired_speed_mps` (T,): the time of each interval (s) and the
|
| 557 |
+
speed the follower was asked for (m/s). `reference_waypoints` (N, 2, m), `reference_stations` (N, m along the route),
|
| 558 |
+
`reference_speeds` (N, m/s) and `reference_headings` (N, rad): the route. Rigid branch continuations add `branch_frame` and
|
| 559 |
+
`branch_waypoints`, `branch_stations`, `branch_speeds`, `branch_headings`.
|
| 560 |
+
|
| 561 |
+
**`anchor_state.npz`.** `state` (17,) and `pose` (3,) at the start of recording, `goal_xy` (2, m), `goal_radius_m`,
|
| 562 |
+
and `history` (16, 24): 16 steps of 50 ms (0.8 s), each with the 17 state numbers, the 3 previously applied
|
| 563 |
+
controls, the position relative to the current pose in its heading frame (x, y in m) and the sine and cosine of the
|
| 564 |
+
relative heading (`HISTORY_FIELDS` in `src/nedm/traverse/fdm_data.py`). At a standing start the history repeats the
|
| 565 |
+
first state with a braking command.
|
| 566 |
+
|
| 567 |
+
### `episodes.csv.gz`
|
| 568 |
+
|
| 569 |
+
One row per drive of a raw collection, next to its tar shards at `traversing/raw/<item>/episodes.csv.gz`:
|
| 570 |
+
|
| 571 |
+
| Column | Meaning |
|
| 572 |
+
|---|---|
|
| 573 |
+
| `run_id` | drive folder name |
|
| 574 |
+
| `collection` | the collection folder: the folder holding the drive's `runs/` folder, else the drive's parent folder |
|
| 575 |
+
| `vehicle`, `world` | `hmmwv`, `gator` or `polaris`; `rigid` or `soil` (from `outcome.json`, else the item's defaults) |
|
| 576 |
+
| `arena` | the arena asset path from `case.json`, for example `assets/traverse/arena_f104_50h_v1`; the item `arena_heightmaps` restores that folder (heightmap `arena_000.bmp` and `arena_meta.json`) |
|
| 577 |
+
| `status`, `elapsed_s` | end status and recorded time (s) from `outcome.json` |
|
| 578 |
+
| `n_files`, `bytes` | released files of the drive and their total size |
|
| 579 |
+
| `trajectory_sha256`, `outcome_sha256` | SHA256 of the drive's `trajectory.npz` and `outcome.json` |
|
| 580 |
+
| `shard` | the tar shard holding the drive |
|
| 581 |
+
|
| 582 |
+
These tables have no `load_dataset` configs of their own: the `datasets` library reads every config of a repository
|
| 583 |
+
with one file format, here the paper's Parquet. Load a table with the CSV reader and its Hub path instead:
|
| 584 |
+
|
| 585 |
+
```python
|
| 586 |
+
from datasets import load_dataset
|
| 587 |
+
eps = load_dataset("csv", split="train", data_files="hf://datasets/harryzhang1018/NeDM@6620faead5225ac9aa5ae8ab19bc2ef2db38a863/"
|
| 588 |
+
"traversing/raw/hmmwv_soil_f104_collection/episodes.csv.gz")
|
| 589 |
+
```
|
| 590 |
+
|
| 591 |
+
### Download and restore
|
| 592 |
+
|
| 593 |
+
The helper in the code repository downloads items at the pinned revision, checks every file against the manifest's
|
| 594 |
+
SHA256, and restores each item to its path in a checkout (for example `artifacts/traverse/crm_f104_v1/collect_v1/runs/`
|
| 595 |
+
for the HMMWV soil collection), which is where the study's scripts expect it. It needs `huggingface_hub`:
|
| 596 |
+
|
| 597 |
+
```bash
|
| 598 |
+
python traversing/scripts/release/download_traversing_data.py --list --all # items, sizes, restore paths
|
| 599 |
+
python traversing/scripts/release/download_traversing_data.py --milestone m2 --bundle models
|
| 600 |
+
python traversing/scripts/release/download_traversing_data.py --items hmmwv_soil_f104_collection --verify-members
|
| 601 |
+
python traversing/scripts/release/verify_release.py --local artifacts/hf_release/download # re-check downloaded files
|
| 602 |
+
```
|
| 603 |
+
|
| 604 |
+
The default revision is `6620faead5225ac9aa5ae8ab19bc2ef2db38a863`, the Hub commit pinned in
|
| 605 |
+
`traversing/manifests/hf_release_manifest.json`; `--revision` selects another one (with a warning). Items named in an
|
| 606 |
+
item's `requires` are added to the selection. Existing files with other content are kept unless `--overwrite` is
|
| 607 |
+
given. Without the helper, fetch a folder with
|
| 608 |
+
`huggingface_hub.snapshot_download("harryzhang1018/NeDM", repo_type="dataset", revision="6620faead5225ac9aa5ae8ab19bc2ef2db38a863", allow_patterns="traversing/models/m2_shared_history_model/*")`
|
| 609 |
+
and extract tar shards with `tar -xzf part-00000.tar.gz` at the root of a checkout.
|
| 610 |
+
|
| 611 |
+
### Loading notes
|
| 612 |
+
|
| 613 |
+
- **`.npz` files** contain object arrays in places: load them with `numpy.load(path, allow_pickle=True)`.
|
| 614 |
+
- **Models** are PyTorch checkpoints: a dict holding a state dict (under `state` for the risk networks, `model` for the
|
| 615 |
+
NRD, `model_state_dict` for the PPO checkpoint) plus the architecture settings. Load with
|
| 616 |
+
`torch.load(path, map_location="cpu", weights_only=True)`, allowing the NumPy arrays and paths some of them store
|
| 617 |
+
(`load_checkpoint` in `traversing/scripts/release/verify_release.py` shows how). The JSON next to each ensemble is
|
| 618 |
+
its training record. The network classes are in the experiment code at commit
|
| 619 |
+
[`901d6c9`](https://github.com/uwsbel/NeDM/tree/901d6c9423a16c0fafc3d60056065415d5a725f2) and are not on main yet.
|
| 620 |
+
- **The tracker actor** (`m3_tracker_round2_numpy_actor/actor.npz`) is plain NumPy: observation normalisation
|
| 621 |
+
(`obs_mean`, `obs_var`, `obs_eps`), the layer weights `W0..W3`, `b0..b3` of a 158 -> 512 -> 256 -> 128 -> 3
|
| 622 |
+
network, and the action scaling (`action_center`, `action_scale`, `action_low`, `action_high`). It runs without
|
| 623 |
+
torch.
|
| 624 |
+
- **Units** are in the key names (`_m`, `_mps`, `_rad`, `_radps`, `_n`, `_nm`, `_kw`, `_kj`, `_s`).
|
| 625 |
+
|
| 626 |
+
### Known limitations
|
| 627 |
+
|
| 628 |
+
- **Simulation only.** Every drive is a Chrono simulation; there is no real-vehicle data.
|
| 629 |
+
- **Whole-arena overhead camera.** Every planner map comes from a fixed overhead camera that sees the whole arena,
|
| 630 |
+
with ideal depth, and the planners use the simulator's exact vehicle pose.
|
| 631 |
+
- **One soil setting and one terrain generator.** All soil drives use the same 0.24 m particle layer, and only the
|
| 632 |
+
wheels touch the soil: the HMMWV's rigid tyre mesh, and plain cylinders for the Gator and Polaris calibrated to the
|
| 633 |
+
HMMWV's sinkage. All arenas come from one hill-and-crater generator.
|
| 634 |
+
- **The Polaris's engine values are not physical.** Chrono's stock Polaris driveline gives the wheels about 16 times
|
| 635 |
+
the engine's power, so in Polaris drives the engine speed and torque (state numbers 15 and 16), `power_kw` and the
|
| 636 |
+
work values do not describe a real engine.
|
| 637 |
+
- **Vehicle-specific models.** Each risk model was trained on one vehicle's drives and is meant for that vehicle; the
|
| 638 |
+
labels describe drives by Chrono's stock PID follower. The predicted failure probabilities rank routes but are not
|
| 639 |
+
calibrated.
|
| 640 |
+
- **Re-drives are not always bit-identical.** Chrono runs on the cluster repeat exactly on one node but not across
|
| 641 |
+
nodes. Re-driving the PID follower on soil reproduced 369 of 423 trajectories exactly and 419 of 423 end states.
|
| 642 |
+
|
| 643 |
+
### License and citation
|
| 644 |
+
|
| 645 |
+
BSD-3-Clause, the same as Part A and the code. The traversing study has no publication yet: please cite the NeDM paper
|
| 646 |
+
(Part A) and name the dataset revision you used (`6620faead5225ac9aa5ae8ab19bc2ef2db38a863`).
|