by Jonas Frey, David Hoeller, Shehryar Khattak, Marco Hutter
Paper Link:
Accepted for IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS, 2022)
@inproceedings{frey2022traversability,
author={Frey, Jonas and Hoeller, David and Khattak, Shehryar and Marco, Hutter},
journal={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS, 2022)},
title={Locomotion Policy Guided Traversability Learning using Volumetric Representations of Complex Environments},
year={2022}
}
Video by Jonas Frey
Deployed Locomotion Policy from “Learning robust perceptive locomotion for quadrupedal robots in the wild“ by Takahiro Miki, Joonho Lee, Jemin Hwangbo, Lorenz Wellhausen, Vladlen Koltun, Marco Hutter
Paper Link:
Youtube Video (14s-23s):
Abstract:
Despite the progress in legged robotic locomotion, autonomous navigation in unknown environments remains an open problem. Ideally, the navigation system utilizes the full potential of the robots’ locomotion capabilities while operating within safety limits under uncertainty. The robot must sense and analyze the traversability of the surrounding terrain, which depends on the hardware, locomotion control, and terrain properties. It may contain information about the risk, energy, or time consumption needed to traverse the terrain. To avoid hand-crafted traversability cost functions we propose to collect traversability information about the robot and locomotion policy by simulating the traversal over randomly generated terrains using a physics simulator. Thousand of robots are simulated in parallel controlled by the same locomotion policy used in reality to acquire 57 years of real-world locomotion experience equivalent. For deployment on the real robot, a sparse convolutional network is trained to predict the simulated traversability cost, which is tailored to the deployed locomotion policy, from an entirely geometric representation of the environment in the form of a 3D voxel-occupancy map. This representation avoids the need for commonly used elevation maps, which are error-prone in the presence of overhanging obstacles and multi-floor or low-ceiling scenarios. The effectiveness of the proposed traversability prediction network is demonstrated for path planning for the legged robot ANYmal in various indoor and natural environments.
1 view
0
0
5 months ago 00:00:00 1
Нина Хрущёва и Лев Шлосберг / Пропаганда как локомотив истории / Люди мира
6 months ago 00:02:52 1
Learning Robust Autonomous Navigation and Locomotion for Wheeled-legged Robots
6 months ago 00:04:52 1
ANYmal Parkour: Learning Agile Navigation for Quadrupedal Robots
6 months ago 00:02:58 1
Learning to walk in confined spaces using 3D representation
8 months ago 00:17:16 1
“WHY ASBESTOS?” 1970 ASBESTOS INFORMATION COMMITTEE PROPAGANDA FILM XD52424
8 months ago 00:01:00 1
Thanks to NVIDIA Robotics and the Robotic System Labs (RSL), #ANYmal conquers challenging scenarios
9 months ago 00:06:46 1
DTC: Deep Tracking Control
9 months ago 00:04:09 1
Reinforcement Learning for Versatile, Dynamic, and Robust Bipedal Locomotion Control
1 year ago 00:22:50 1
The New Silk Road Becomes the World Land Bridge, A Tour
1 year ago 00:06:00 1
Lobsters, Wine, Weapons & A Helicopter To Escape: All Part Of Kim Jong Un’s Armoured Train To Russia
1 year ago 00:18:49 1
Railway Electrification Systems Worldwide: Explained
1 year ago 00:04:40 1
RL + Model-based Control: Using On-demand Optimal Control to Learn Versatile Legged Locomotion
2 years ago 00:02:33 1
Learning to Walk via Deep Reinforcement Learning
2 years ago 00:02:54 1
Reward-Adaptive Reinforcement Learning: Dynamic Policy Gradient Optimization for Bipedal Locomotion
2 years ago 00:02:16 1
Learning agile and dynamic motor skills for legged robots
2 years ago 00:08:18 2
DeepGait: Planning and Control of Quadrupedal Gaits using Deep Reinforcement Learning (Presentation)
2 years ago 00:03:16 4
Creating a Dynamic Quadrupedal Robotic Goalkeeper with Reinforcement Learning
2 years ago 00:02:56 1
Advanced Skills by Learning Locomotion and Local Navigation End-to-End
2 years ago 00:03:27 6
GenLoco: Generalized Locomotion Controllers for Quadrupedal Robots