Research at NVIDIA: RMPflow - A Computational Graph for Automatic Motion Policy Generation
14 578
10.7
NVIDIA1.47 млн
Следующее
18.01.19 – 27 3723:14Популярные
43 дня – 162 5182:14
94 дня – 13 0253:03Опубликовано 17 января 2019, 19:50
In this work, researchers from NVIDIA, Georgia Institute of Technology, and the University of Washington, developed a new motion generation and control framework that enables globally stable controller design within intrinsically non-Euclidean spaces.
Non-Euclidean geometries are not often modeled explicitly in robotics, but are nonetheless common in the natural world. One important example is the apparent non-Euclidean behavior of obstacle avoidance. Obstacles become holes in this setting. As a result, straight lines are no longer a reasonable definition of shortest distance—geodesics must, therefore, naturally flow around them. This behavior implies a form of non-Euclidean geometry: the space is naturally curved by the presence of obstacles.
Learn more: arxiv.org/pdf/1811.07049.pdf
Non-Euclidean geometries are not often modeled explicitly in robotics, but are nonetheless common in the natural world. One important example is the apparent non-Euclidean behavior of obstacle avoidance. Obstacles become holes in this setting. As a result, straight lines are no longer a reasonable definition of shortest distance—geodesics must, therefore, naturally flow around them. This behavior implies a form of non-Euclidean geometry: the space is naturally curved by the presence of obstacles.
Learn more: arxiv.org/pdf/1811.07049.pdf
Свежие видео
2 дня – 107 24027:35
3 дня – 5651:28
3 дня – 4943:02
5 дней – 61 45010:19
8 дней – 158 4621:00
10 дней – 2 6904:37Случайные видео
53 дня – 2911:35
326 дней – 70 52410:26
352 дня – 8570:55
20.01.23 – 18 3339:58
03.06.22 – 159 9123:41
11.01.08 – 4 8070:26
1 день – 3 9339:59
1 день – 145 8458:01
2 дня – 3173:57
2 дня – 12 14211:42
4 дня – 2 2112:59
5 дней – 594 1100:47
2 дня – 6 70220:34