-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
456 lines (435 loc) · 18.3 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
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="Generalizable Long-Horizon Manipulations with Large Language Models">
<title>Generalizable Long-Horizon Manipulations with Large Language Models</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="./css/bulma.min.css">
<link rel="stylesheet" href="./css/bulma-carousel.min.css">
<link rel="stylesheet" href="./css/bulma-slider.min.css">
<link rel="stylesheet" href="./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="./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="./js/fontawesome.all.min.js"></script>
<script src="./js/bulma-carousel.min.js"></script>
<script src="./js/bulma-slider.min.js"></script>
<script src="./js/index.js"></script>
<link rel="stylesheet" href="index.css">
<style>
</style>
</head>
<body>
<section class="hero">
<div class="hero-body" style="background-color: white;">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title" style="width: 1200px; color: black;">Generalizable Long-Horizon Manipulations with
Large Language Models</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
Haoyu Zhou<sup>1</sup>,</span>
<span class="author-block">
Mingyu Ding<sup>2</sup>,</span>
<span class="author-block">
Weikun Peng<sup>1</sup>,</span>
</span>
<span class="author-block">
Masayoshi Tomizuka<sup>2</sup>,</span>
</span>
<span class="author-block">
Lin Shao<sup>1</sup>,</span>
</span>
<span class="author-block">
Chuang Gan<sup>3,4</sup></span>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>National University of Singapore;</span>
<span class="author-block"><sup>2</sup>UC Berkeley, USA;</span>
<span class="author-block"><sup>3</sup>University of Massachusetts Amherst, USA;</span>
<span class="author-block"><sup>4</sup>MIT-IBM Watson AI Lab, USA</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2310.02264"
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/2310.02264"
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://www.youtube.com/watch?v=TpCIRg9XuEg&ab_channel=object814"
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>
</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">Abstract</h2>
<div class="content has-text-justified">
<p>
This work introduces a framework harnessing the capabilities of Large Language Models (LLMs) to generate primitive task conditions for
generalizable long-horizon manipulations with novel objects and unseen tasks. These task conditions serve as guides for the
generation and adjustment of Dynamic Movement Primitives (DMP) trajectories for longhorizon task execution. We further create a challenging robotic manipulation task suite
based on Pybullet for long-horizon task evaluation.
Extensive experiments in both simulated and realworld environments demonstrate the effectiveness of our framework on both familiar tasks involving new objects and novel but related tasks,
highlighting the potential of LLMs in enhancing robotic system versatility and adaptability
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Framework Structure</h2>
<div class="content has-text-justified">
<p>
We leverage LLMs to generate and generalize primitive task conditions for both familiar tasks with novel objects and novel but related tasks.
Subsequently, the highlevel task conditions guide the generation and adjustment of lowlevel trajectories originally learned from demonstrations for longhorizon task execution.
</p>
</div>
<div class="columns is-vcentered">
<div class="column is-three-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/1.mp4" type="video/mp4">
</video>
</div>
<div class="column is-two-fifths has-text-centered">
<img src="compressed/2.png"/>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-three-fifths ">
<h2 class="title is-3">Task condition generation and generalization experiment</h2>
<p class="has-text-justified">
We evaluate the ability of our framework to generate and generalize task conditions on all 10 primitive tasks.
The LLM (GPT-3.5) is provided with condition examples. Comparison is made with task conditions generated from environments (methods explained in paper).
A successfully generated task condition should contain accurate and enough information to guide the execution of the primitive task.
</p>
</div>
<div class="column is-two-fifths has-text-centered">
<img src="compressed/3.png"/>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-three-fifths">
<h2 class="title is-3">Simulation experiment</h2>
<p class="has-text-justified">
To evaluate our framework's ability to execute long-horizon tasks using DMP trajectories generated and adjuested by task conditions, we design a challenging Robotic Manipulation Task Suite in Pybullet.
</p>
<p class="has-text-justified">
The environment consists of two 7 Dof robots Franka and Kinova with a kitchen scene including various interactive objects.
It contains 10 diverse primitive tasks (37 if considering different objects) and 4 long-horizon tasks in simulation.
</p>
</div>
<div class="column is-two-fifths has-text-centered">
<img src="compressed/4.png"/>
</div>
</div>
</div>
</div>
<br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-half has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/5.mp4" type="video/mp4">
</video>
<p class="has-text-centered">
Demonstration: Put the bowl into the bottom drawer
</p>
</div>
<div class="column is-half has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/6.mp4" type="video/mp4">
</video>
<p class="has-text-centered">
Generation and execution: Put the bowl into the upper drawer
</p>
</div>
</div>
</div>
</div>
<br><br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-half has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/7.mp4" type="video/mp4">
</video>
<p class="has-text-centered">
Demonstration: Put the bottle into the box
</p>
</div>
<div class="column is-half has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/8.mp4" type="video/mp4">
</video>
<p class="has-text-centered">
Generation and execution: Put the bowl into the box
</p>
</div>
</div>
</div>
</div>
<br><br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/9.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Move the bottle into the box w/o task condition <br><br></h3>
<p class="has-text-centered">
The move into primitive task is failed because initial goal position obtained from environment pointcloud is deviated.
And a collision between bottle and box happened.
</p>
</div>
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/10.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Move the bottle into the box w/ task condition adjusting trajectory</h3>
<p class="has-text-centered">
As the collision between bottle and box is against task condition collision,
the manipulator backwards a few steps and adjust goal position to generate a new DMP trajectory,
eventually complete the task.
</p>
</div>
<div class="column is-5">
<h3 class="title is-6">Task condition by LLM </h3>
<p>
<strong>Task name:</strong> move bottle into box<br>
<strong>Relevant objects:</strong><br>
bottle, box, gripper<br>
<strong>Pre-conditions:</strong><br>
gripper grasping bottle<br>
box open<br>
<strong>Post-conditions:</strong><br>
gripper grasping bottle<br>
bottle inside box<br>
box open<br>
<strong>Collisions:</strong><br>
gripper and bottle
</p>
</div>
</div>
</div>
</div>
<br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/11.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Grasp the bottle w/o task condition <br><br></h3>
<p class="has-text-centered">
The move into primitive task is failed because initial goal position obtained from environment pointcloud is deviated.
And a collision between bottle and box happened.
</p>
</div>
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/12.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Grasp the bottle w/ task condition <br> generating trajectory</h3>
<p class="has-text-centered">
As the collision between bottle and box is against task condition collision,
the manipulator backwards a few steps and adjust goal position to generate a new DMP trajectory,
eventually complete the task.
</p>
</div>
<div class="column is-5">
<h3 class="title is-6">Task condition by LLM </h3>
<p>
<strong>Task name:</strong> grasp bottle<br>
<strong>Relevant objects:</strong><br>
bottle, gripper<br>
<strong>Pre-conditions:</strong><br>
gripper not grasping<br>
<strong>Post-conditions:</strong><br>
gripper grasping bottle<br>
<strong>Collisions:</strong><br>
gripper and bottle
</p>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Real-world experiment</h2>
<div class="content has-text-justified">
<p>
To demonstrate the practicality of our to perform longhorizon tasks in real scenarios, we set up a real-world long-horizon task,
featuring a dual-arm MOVO robot which has two 7 DoF manipulators and a Kinect RGB-D camera overhead.
We use Segment Anything to obtain segmented point clouds of the surroundings and objects.
Position control is performed as we use rangeIK for solving inverse kinematics.
</p>
</div>
<div class="columns is-vcentered">
<div class="column is-three-fifths">
<video controls autoplay muted loop>
<source src="compressed/13.mp4" type="video/mp4">
</video>
</div>
<div class="column is-two-fifths">
<p>
■ grasp the bottle;<br>
■ open the microwave;<br>
■ move the bottle in front of the microwave;<br>
■ move the bottle into the microwave:<br>
■ release the bottle;<br>
■ move out of the microwave;<br>
■ close the microwave.<br>
</p>
</div>
</div>
</div>
</div>
</div>
<br><br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/14.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Move the bottle into the microwave w/o task condition <br><br></h3>
</div>
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/15.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Move the bottle into the microwave w/ task condition adjusting trajectory</h3>
</div>
<div class="column is-5">
<h3 class="title is-6">Task condition by LLM </h3>
<p>
<strong>Task name:</strong> move bottle into microwave<br>
<strong>Relevant objects:</strong><br>
bottle, microwave, gripper<br>
<strong>Pre-conditions:</strong><br>
gripper grasping bottle<br>
microwave open<br>
<strong>Post-conditions:</strong><br>
gripper grasping bottle<br>
bottle inside microwave<br>
microwave open<br>
<strong>Collisions:</strong><br>
gripper and bottle
</p>
</div>
</div>
</div>
</div>
<br><br>
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="columns is-full-width">
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/16.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Grasp the bottle w/o task condition <br><br></h3>
</div>
<div class="column is-two-fifths has-text-centered">
<video controls autoplay muted loop>
<source src="compressed/17.mp4" type="video/mp4">
</video>
<h3 class="title is-6">Grasp the bottle w/ task condition <br> generating trajectory</h3>
</div>
<div class="column is-5">
<h3 class="title is-6">Task condition by LLM </h3>
<p>
<strong>Task name:</strong> right grasp bottle<br>
<strong>Relevant objects:</strong><br>
bottle, right gripper<br>
<strong>Pre-conditions:</strong><br>
right gripper not grasping<br>
<strong>Post-conditions:</strong><br>
right gripper grasping bottle<br>
<strong>Collisions:</strong><br>
right gripper and bottle
</p>
</div>
</div>
</div>
</div>
</section>
<footer class="footer">
<div class="container">
<p>
ICRA2025 submittion: Generalizable Long-Horizon Manipulations with Large Language Models
</p>
</div>
</footer>
</body>
</html>