forked from nerfies/nerfies.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
400 lines (329 loc) · 16.5 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="FMB: A Functional Manipulation Benchmark for Generalizable Robotic Learning">
<meta name="keywords" content="FMB, Robotic">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>FMB: A Functional Manipulation Benchmark for Generalizable Robotic Learning</title>
<!-- <!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-PYVRSFMDRL"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'G-PYVRSFMDRL');
</script> -->
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/favicon.svg">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<nav class="navbar" role="navigation" aria-label="main navigation">
<div class="navbar-brand">
<a role="button" class="navbar-burger" aria-label="menu" aria-expanded="false">
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
</a>
</div>
<div class="navbar-menu">
<div class="navbar-start" style="flex-grow: 1; justify-content: center;">
<div class="navbar-item has-dropdown is-hoverable">
<a class="navbar-link">
More Research
</a>
<div class="navbar-dropdown">
<a class="navbar-item" href="https://hypernerf.github.io">
HyperNeRF
</a>
<a class="navbar-item" href="https://nerfies.github.io">
Nerfies
</a>
<a class="navbar-item" href="https://latentfusion.github.io">
LatentFusion
</a>
<a class="navbar-item" href="https://photoshape.github.io">
PhotoShape
</a>
</div>
</div>
</div>
</div>
</nav>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">FMB: A Functional Manipulation Benchmark for Generalizable Robotic Learning</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://people.eecs.berkeley.edu/~jianlanluo/">Jianlan Luo</a><sup>1</sup>,</span>
<span class="author-block">
<a href="https://charlesxu0124.github.io/">Charles Xu</a><sup>1</sup>,</span>
<span class="author-block">
<a href="https://people.eecs.berkeley.edu/~svlevine/">Sergey Levine</a><sup>1</sup>,
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>University of California Berkeley,</span>
<span class="author-block"><sup>2</sup></span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2011.12948"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<span class="link-block">
<a href="https://arxiv.org/abs/2011.12948"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
<!-- Video Link. -->
<span class="link-block">
<a href="https://youtu.be/gt3wxt-rXfo"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-youtube"></i>
</span>
<span>Video</span>
</a>
</span>
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/manipulationdataset/ManipulationDataset"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
<!-- Dataset Link. -->
<span class="link-block">
<a href="https://github.com/google/nerfies/releases/tag/0.1"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="far fa-images"></i>
</span>
<span>Data</span>
</a>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">A Functional Manipulation Benchmark for Generalizable Robotic Learning</h2>
<div class="content has-text-justified">
Our benchmark for studying robotic learning for functional manipulation consists of a variety of easily reproducible 3D printed objects, each one requiring a sequence of grasping, reorientation, and assembly behaviors. Generalization can be evaluated on test objects and varied positions, as well as more complex multi-stage assembly tasks. We also provide an imitation learning system that provides a basic set of policies for each skill, allowing researchers to use our tasks as a toolkit for studying any portion of the pipeline.
</div>
</div>
</div>
<!--/ Abstract. -->
<!-- Paper video. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Video</h2>
<div class="publication-video">
<iframe src="https://youtu.be/gt3wxt-rXfo"
frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
</div>
</div>
</div>
<!--/ Paper video. -->
</div>
</section>
<!-- TODO: embed google survey -->
<section class="section" id="Code">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Training Code</h2>
<div class="content has-text-justified"
<p>
<a href="https://github.com/manipulationdataset/ManipulationDataset">This repository </a> contains the basic training code used to obtain our baseline results. Keep in mind that you will have to modify the bash script when you would like to change the training settings.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="Files">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Material and CAD Files</h2>
<div class="content has-text-justified"
<p>
The equipment used in our project is entirely available online or 3D printed. We have included the links to purchase the equipment and 3D printing filament as well as the CAD files below.
<a href="https://docs.google.com/spreadsheets/d/1FNBmwlPkXbzScR-CIzNcIH1uaLNXOD9VvLRkpvJ98Z4/edit?usp=sharing">This master spreadsheet </a> contains all the equipment needed and how to obtain them. Each section of the spreadsheet is detailed below.
</p>
<h3 class="title is-4">Purchased Equipment</h3>
<a href="https://docs.google.com/spreadsheets/d/1FNBmwlPkXbzScR-CIzNcIH1uaLNXOD9VvLRkpvJ98Z4/edit#gid=445569452&range=A1"> Links to purchase the required equipment can be found here. </a>
<h3 class="title is-4"> Basic Task Pieces</h3>
We designed a series of basic assembly task boards with 9 slots for different shaped pieces and different sizes as well as 54 pegs of different shapes, sizes, lengths, and colors. You can choose to print all or some of these suitable for your task. In addition, we designed an environment fixture for reorienting the pieces before insertion if they are grasped horizontally.
Refer to <a https://docs.google.com/document/d/14jp37YcES9fKLRLuuB4xPy7HWZb-7hi8xv4HEAamd-E/edit?usp=sharing>color/shape name and number reference sheet </a> for the name and number assigned to each shape and color throughout our project.
<h3 class="title is-4"> Novel Objects for Grasping (Optional)</h3>
In addition to the original 54 insertion pieces, we designed 5 novel objects of different sizes, shapes, and colors to study the generalization abilities of the grasp policy.
<h3 class="title is-4"> Complex Assembly Task (Optional) </h3>
In addition to the three basic insertion task boards, we introduce two more complex assembly boards where interlocking pieces need to be assembled in a specific order.
<h3 class="title is-4"> 3D Printed Pieces </h3>
Camera Mounts
We designed two types of camera mounts for the RealSense D405 cameras. One is used to mount the camera in a fix position pointing at the workspace, while the other is mounted between the flange and the gripper of the Franka Panda arm for a eye-in-hand view.
</div>
</div>
</div>
</div>
</section>
<section class="section" id="Files">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Workspace setup</h2>
<div class="content has-text-justified">
Once all the necessary materials have been obtained, you can set up the workspace for the robot according to the <strong> instructions and dimensions in the diagram below. </strong>
<ol>
<li>Fix the bin 12 cm in front of the robot base.</li>
<li>Fix the reorientation fixture onto the bin according to the dimensions in the diagram.</li>
<li>Mount side camera posts outside the bin according to the dimensions in the diagram.</li>
<li>Fix the 3D-printed side camera mount onto the interior side of the post with the bottom edge of the mount 15cm above the bottom of the bin. </li>
<li> Mount the wrist camera between the flange and gripper of the Franka arm. </li>
</ol>
</div>
<!-- TODO: insert image -->
</div>
</div>
</div>
</section>
<section class="section" id="Files">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Evaluation</h2>
<h3 class="title is-4">Evaluation procedure</h3>
<p> <strong> The randomization box is 32 x 32 cm. The randomization angle is ±10 degrees from the normal orientation. </strong></p>
<p> <strong> The peg randomization box is 20x25 cm. </strong></p>
<p> <strong> Between each grasping trial, randomize the location and angle of the peg. </strong></p>
<p> <strong> Between each insertion trial, randomize the location of the board within the randomization box. Feel free to also rotate it, but generally keep it in the same orientation. </strong></p>
Experiment 1: Comparing depth for grasping
<div class="content has-text-justified">
<ol>
<ul>
<li><strong>Training Phase:</strong> Train a policy on all grasping data with (1) RGB only data, and (2) RGB + Depth data.</li>
<li><strong>Testing Phase 1:</strong> Roll out each policy on 10 randomly selected pegs for 5 trials each. Maintain a 50:50 ratio between horizontal and vertical grasps.</li>
<li><strong>Testing Phase 2:</strong> Roll out each policy on 5 randomly selected novel objects for 5 trials each, again aiming for a 50:50 ratio between horizontal and vertical grasps.</li>
</ul>
</ol>
</div>
Experiment 2: Comparing oval grasping generalization
<div class="content has-text-justified">
<ol>
<ul>
<li> Training phase: Train a policy </li>
<li><strong>Testing Phase:</strong> Roll out each policy on all 6 ovals for 5 trials each, with approximately half of the trials as horizontal grasps and the other half as vertical grasps.</li>
</ul>
</ol>
</div>
Experiment 3: Comparing depth for insertion
<div class="content has-text-justified">
<ol>
<ul>
<li><strong>Training Phase:</strong> Train an insertion policy only on insertion data for (1) RGB round data, (2) RGB hexagon data, and (3) RGB double square data. Train three additional policies in the same manner but include RGB + Depth data.</li>
<li><strong>Testing Phase:</strong> Evaluate each policy on the peg that it was trained with. For each shape, evaluate on all 6 pegs of that shape, with 5 trials for each peg.</li>
</ul>
</ol>
</div>
Experiment 4: Insertion generalization between different shapes
<div class="content has-text-justified">
<ol>
<ul>
<li> Training Phase: Train an insertion policy on RGB Data with (1) all round data, and (2) all insertion data.</li>
<li> Testing Phase 1: Evaluate the policy trained only on round data on all 6 round pegs, with 5 trials each peg.</li>
<li> Testing Phase 2: Evaluate the policy trained only on round data on all 6 star pegs, with 5 trials each peg. </li>
<li> Testing Phase 3: Evaluate the policy trained on all data on all 6 round pegs, with 5 trials each peg.</li>
<li> Testing Phase 4: Evaluate the policy trained on all data on all 6 star pegs, with 5 trials each peg.</li>
</ul>
</ol>
</div>
Experiment 5: Long horizon
<div class="content has-text-justified">
<ol>
<ul>
<li> Training Phase: Using RGB only data, train on all data with the same network.</li>
<li> Testing Phase: Evaluate on 5 random pegs, conducting 2 trials each for a total of 10 trials.</li>
</ul>
</ol>
</div>
Evaluation Results from our baseline
</div>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link"
href="./static/videos/nerfies_paper.pdf">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/keunhong" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
If you have any questions about this project, please contact us at [email protected]
</p>
<p>
All data is provided under the <a href="https://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution 4.0 International License </a>.
</p>
<p>
This website is licensed under a <a rel="license"
href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
<p>
This means you are free to borrow the <a
href="https://github.com/nerfies/nerfies.github.io">source code</a> of this website,
we just ask that you link back to this page in the footer.
Please remember to remove the analytics code included in the header of the website which
you do not want on your website.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>