-
Notifications
You must be signed in to change notification settings - Fork 1
/
index.html
306 lines (274 loc) · 12.4 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="">
<meta name="keywords" content="PIP-Loco, Planning, Legged Robots, Reinforcement Learning, Quadruped Robot">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title></title>
<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/stochlab.png">
<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>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<img src="./static/images/final_logo.png" width="20%" alt="Stoch Lab Logo"/>
<!-- <img src="./static/images/IISc_Logo.jpg" width="10%" alt="IISc Logo" align="right"/> -->
<img src="./static/images/iisc_logo.png" width="11%" alt="IISc Logo" align="right"/>
<div class="columns is-centered"></div>
</div>
</div>
</section>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="column has-text-centered">
<h1 class="title is-2 publication-title">PIP-Loco: A Proprioceptive Infinite Horizon Planning Framework for Quadrupedal Robot Locomotion</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://aditya-shirwatkar.github.io/">Aditya Shirwatkar</a>,
</span>
<span class="author-block">
<a href="">Naman Saxena</a>,
</span>
<span class="author-block">
<a href="">Kishore Chandra</a>,
</span>
<span class="author-block">
<a href="https://www.shishirny.com/">Shishir Kolathaya</a>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"> Indian Institute of Science (IISc), Bangalore </span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<!-- <span class="link-block">
<a href=""
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/2409.09441"
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/_mPUU_fAuQY"
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://www.github.com/StochLab/PIP-Loco"
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>
</div>
</div>
<!-- Under Review Text -->
<div class="is-size-6 publication-authors">
<span class="author-block"> (Preprint under review) </span>
</div>
<div class="column has-text-centered">
<div class="columns is-centered has-text-centered">
<div class="column">
<img src="images/results/gravel.gif" width="48%">
<img src="images/results/stairs.gif" width="48%">
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column">
<h2 class="title is-3">Abstract</h2>
<div class="content has-text-justified">
<p>
A core strength of Model Predictive Control (MPC) for quadrupedal locomotion has been its ability to enforce constraints and provide interpretability of the sequence of commands over the horizon. However, despite being able to plan, MPC struggles to scale with task complexity, often failing to achieve robust behavior on rapidly changing surfaces. On the other hand, model-free Reinforcement Learning (RL) methods have outperformed MPC on multiple terrains, showing emergent motions but inherently lack any ability to handle constraints or perform planning. To address these limitations, we propose a framework that integrates proprioceptive planning with RL, allowing for agile and safe locomotion behaviors through the horizon. Inspired by MPC, we incorporate an internal model that includes a velocity estimator and a Dreamer module. During training, the framework learns an expert policy and an internal model that are co-dependent, facilitating exploration for improved locomotion behaviors. During deployment, the Dreamer module solves an infinite-horizon MPC problem, adapting actions and velocity commands to respect the constraints. We validate the robustness of our training framework through ablation studies on internal model components and demonstrate improved robustness to training noise. Finally, we evaluate our approach across multi-terrain scenarios in both simulation and hardware.
</p>
<br>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column">
<h2 class="title is-3"> Comparison to Related Work </h2>
</div>
</div>
<div class="columns is-centered">
<div class="column is-four-fifths">
<div class="content has-text-justified" style="padding-top: 3%">
<p>
Here, an adaptation module models environment hidden parameters (like friction, contact forces, etc).
An internal model predicts the robot base velocities, latent states for terrain characteristics/disturbance responses, or future observations up to <i>H</i> steps.
NLM, PLM, and FLM denote the No-Latent, Partially-Latent, and Fully-Latent Dreamer Modules, respectively (see the paper for details).
</p>
</div>
<img src="images/Comparison.png" width="100%">
</div>
</div>
</div>
</section>
<section class="hero is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-bottom: 3%">
<h2 class="title is-3"> Training Architecture </h2>
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<img src="images/Training.png" alt="PIP-Loco Training Architecture" width="80%">
<div class="content has-text-justified" style="padding-top: 5%">
<p>
The internal model (comprising of a velocity estimator and Dreamer module) learns in a co-dependent way with the Asymmetric Actor-Critic. The Dreamer module facilitates temporal reasoning by dreaming about future observations and latent states, enhancing exploration for improved locomotion behaviors.
</p>
</div>
</div>
</div>
</div>
</section>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-bottom: 3%">
<h2 class="title is-3"> Deployment </h2>
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<!-- Add some padding between images -->
<img src="images/Main.png" alt="PIP-Loco Main" width="40%" style="padding-right: 5%">
<img src="images/Deployment.png" alt="PIP-Loco Deployment" width="35%" style="padding-left: 5%">
<img src="images/Alg1.png" alt="PIP-Loco Alg1" width="75%" style="padding-top: 5%">
<div class="content has-text-justified" style="padding-top: 5%">
<p>
The Dreamer module solves an infinite-horizon MPC problem to generate actions for the robot, ensuring robust constraint handling and adaptive locomotion across terrains.
See the paper for more details on the algorithm.
</p>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-bottom: 3%">
<h2 class="title is-3"> Results </h2>
</div>
</div>
<div class="columns is-centered">
<div class="column is-four-fifths">
<div class="content has-text-justified">
<p>
Our approach demonstrates robust locomotion behaviors across multiple terrains, including gravel surfaces, slopes, stairs, and steep drops of upto 45cm.
</p>
</div>
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-top: 5%">
<h2 class="title is-5"> PIP-Loco with NLM </h2>
</div>
</div>
<div class="columns is-centered">
<img src="images/results/combined.gif" width="39%" style="padding-right: 2%">
<img src="images/results/drop.gif" width="39%" style="padding-left: 2%">
</div>
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-top: 5%">
<h2 class="title is-5"> PIP-Loco with PLM & FLM </h2>
</div>
</div>
<div class="columns is-centered">
<img src="images/results/plm.gif" width="39%">
</div>
<div class="columns is-centered has-text-centered">
<div class="column" style="padding-top: 5%">
<h2 class="title is-5"> Importance of Planning </h2>
</div>
</div>
<div class="columns is-centered">
<img src="images/results/comparison.gif" width="49%" style="padding-right: 2%">
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title is-5">Cite</h2>
<pre><code>
@misc{shirwatkar2024piplocoproprioceptiveinfinitehorizon,
title={PIP-Loco: A Proprioceptive Infinite Horizon Planning Framework for Quadrupedal Robot Locomotion},
author={Aditya Shirwatkar, Naman Saxena, Kishore Chandra and Shishir Kolathaya},
year={2024},
eprint={2409.09441},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2409.09441},
}
</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link" href="https://github.com/StochLab" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
</div>
</footer>
</body>
</html>