bfz111 commited on
Commit
5bfb078
·
verified ·
1 Parent(s): 213db09

Add dataset card and split manifests

Browse files
LICENSE ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MIT License
2
+
3
+ Copyright (c) 2026 Anonymous Authors
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
README.md CHANGED
@@ -1,3 +1,166 @@
1
  ---
2
  license: mit
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: mit
3
+ language:
4
+ - en
5
+ pretty_name: STDP Dataset
6
+ tags:
7
+ - robotics
8
+ - visual-navigation
9
+ - trajectory-planning
10
+ - unmanned-aerial-vehicles
11
+ - signal-temporal-logic
12
  ---
13
+
14
+ # STDP Dataset
15
+
16
+ This repository provides the Stage 1 data used to train, validate, and evaluate
17
+ STDP, a policy for signal temporal logic-guided UAV visual navigation. The
18
+ release contains no author or affiliation information and is prepared for
19
+ double-blind review.
20
+
21
+ ## Dataset Contents
22
+
23
+ Each trajectory sample combines a temporal-logic navigation instruction,
24
+ multi-view visual observations, and a reference motion trajectory.
25
+
26
+ | Component | Contents | Intended use |
27
+ |---|---|---|
28
+ | Semantic STL specification | Object-level temporal and logical constraints | Defines the navigation task presented to the policy |
29
+ | Coordinate-grounded STL specification | The same specification with object references mapped to XYZ coordinates | Supports supervision, analysis, and geometry-aware evaluation |
30
+ | Multi-view RGB observations | Synchronized front, down, left, and right camera images | Provides local visual context without a task-level global semantic map |
31
+ | Reference trajectory | CSV records containing absolute XYZ positions, camera yaw, frame indices, and image filenames | Supplies waypoint supervision and synchronizes motion with observations |
32
+ | Split manifests | JSONL records linking instructions, scenes, trajectories, and image directories | Defines the released training, validation, and evaluation subsets |
33
+ | Trajectory archives | One tar archive per trajectory | Enables selective downloading while preserving the original directory layout |
34
+
35
+ ## Scenes and Task Types
36
+
37
+ The data covers four furnished indoor scenes. `world_v1` is the primary scene
38
+ used by the training, validation, and same-scene evaluation splits.
39
+ `world_v2`, `world_v3`, and `world_v4` provide different indoor layouts for
40
+ cross-layout evaluation.
41
+
42
+ The released trajectories cover four task families:
43
+
44
+ | Task type | Objective |
45
+ |---|---|
46
+ | Reach-Avoid | Reach a specified target while avoiding constrained regions |
47
+ | Multi-Target | Visit multiple targets under temporal and logical constraints |
48
+ | Either-Or | Satisfy one of multiple valid target alternatives |
49
+ | Door Puzzle | Obey ordered reachability and avoidance requirements |
50
+
51
+ ## Data Splits
52
+
53
+ | Manifest | Purpose |
54
+ |---|---|
55
+ | `splits/train.jsonl` | Stage 1 policy training in the primary scene |
56
+ | `splits/validation.jsonl` | Model selection and validation in the primary scene |
57
+ | `splits/test_same.jsonl` | Evaluation on held-out trajectories from the primary scene |
58
+ | `splits/test_cross.jsonl` | Evaluation in the three cross-layout scenes |
59
+
60
+ Only the test trajectories used in the reported evaluation are included.
61
+
62
+ Each line in a split manifest is a JSON object with the following main fields:
63
+
64
+ | Field | Description |
65
+ |---|---|
66
+ | `task_type` | Navigation task family |
67
+ | `stl` | Semantic STL specification |
68
+ | `stl_xyz` | Coordinate-grounded STL specification |
69
+ | `world` | Scene identifier |
70
+ | `trajectory_name` | Unique trajectory name within a scene |
71
+ | `trajectory_path` | Repository-relative path to the reference CSV after extraction |
72
+ | `images` | Repository-relative paths to the four camera directories after extraction |
73
+
74
+ ## Repository Layout
75
+
76
+ ```text
77
+ README.md
78
+ LICENSE
79
+ splits/
80
+ train.jsonl
81
+ validation.jsonl
82
+ test_same.jsonl
83
+ test_cross.jsonl
84
+ archives/
85
+ world_v1/<trajectory_name>.tar
86
+ world_v2/<trajectory_name>.tar
87
+ world_v3/<trajectory_name>.tar
88
+ world_v4/<trajectory_name>.tar
89
+ ```
90
+
91
+ Each archive contains the trajectory CSV and all synchronized images using the
92
+ original repository-relative paths. For example:
93
+
94
+ ```text
95
+ data/world_v1/trajectory/trajectory_0127_01.csv
96
+ data/world_v1/images/trajectory_0127_01/images_front/
97
+ data/world_v1/images/trajectory_0127_01/images_down/
98
+ data/world_v1/images/trajectory_0127_01/images_left/
99
+ data/world_v1/images/trajectory_0127_01/images_right/
100
+ ```
101
+
102
+ ## Usage
103
+
104
+ Install the Hugging Face Hub client:
105
+
106
+ ```bash
107
+ pip install -U huggingface_hub
108
+ ```
109
+
110
+ Read the desired JSONL manifest, then use its `world` and `trajectory_name`
111
+ fields to download the corresponding archive:
112
+
113
+ ```python
114
+ import json
115
+ from huggingface_hub import hf_hub_download
116
+
117
+ repo_id = "bfz111/STDP-Dataset"
118
+
119
+ manifest = hf_hub_download(
120
+ repo_id=repo_id,
121
+ repo_type="dataset",
122
+ filename="splits/test_same.jsonl",
123
+ )
124
+
125
+ with open(manifest, "r", encoding="utf-8") as stream:
126
+ sample = json.loads(next(stream))
127
+
128
+ archive = hf_hub_download(
129
+ repo_id=repo_id,
130
+ repo_type="dataset",
131
+ filename=(
132
+ f"archives/{sample['world']}/"
133
+ f"{sample['trajectory_name']}.tar"
134
+ ),
135
+ )
136
+ print(archive)
137
+ ```
138
+
139
+ Extract the archive from the root of a local dataset directory:
140
+
141
+ ```bash
142
+ tar -xf /path/to/trajectory_0127_01.tar -C /path/to/local_dataset
143
+ ```
144
+
145
+ After extraction, `trajectory_path` and the four entries in `images` resolve
146
+ directly against `/path/to/local_dataset`.
147
+
148
+ To download the complete release, use `snapshot_download`:
149
+
150
+ ```python
151
+ from huggingface_hub import snapshot_download
152
+
153
+ snapshot_download(
154
+ repo_id="bfz111/STDP-Dataset",
155
+ repo_type="dataset",
156
+ local_dir="STDP-Dataset",
157
+ )
158
+ ```
159
+
160
+ ## License
161
+
162
+ The dataset is released under the MIT License. See `LICENSE` for details.
163
+
164
+ ## Citation
165
+
166
+ Coming soon.
splits/test_cross.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
splits/test_same.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
splits/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
splits/validation.jsonl ADDED
The diff for this file is too large to render. See raw diff