BriefGPT.xyz
Mar, 2022
首先不要摔倒：学习如何利用一堵墙和一个损坏的机器人
First do not fall: learning to exploit the environment with a damaged humanoid robot
HTML
PDF
Timothée Anne, Eloïse Dalin, Ivan Bergonzani, Serena Ivaldi, Jean-Baptiste Mouret
TL;DR
文章介绍了一种叫D-Reflex的方法，通过学习神经网络选择与墙接触的位置，用于机器人姿势的稳定控制，可以避免超过75%的可避免摔倒，适用于1.75米、100公斤，30自由度的TALOS机器人。
Abstract
humanoid robots
could replace humans in hazardous situations but most of such situations are equally dangerous for them, which means that they have a high chance of being damaged and fall. We hypothesize that
humanoid r
→