arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.04351cs.ROcs.AIcs.CVcs.LG
收藏

在杂乱和动态环境中学习状态表示和导航

Learning a State Representation and Navigation in Cluttered and Dynamic Environments

Robotic Systems Lab, ETH Zürich(苏黎世联邦理工学院机器人系统实验室);NVIDIA(英伟达公司)

David Hoeller, Lorenz Wellhausen, Farbod Farshidian, Marco Hutter

摘要:

在这项工作中,我们提出了一个基于学习的管道,实现本地导航与四足机器人在杂乱的环境中静态和动态的障碍。给定高级导航命令,机器人能够基于来自深度相机的帧安全地移动到目标位置,而无需对环境进行任何显式映射。首先,使用状态表示学习将图像序列和相机的当前轨迹融合以形成世界的模型。然后,这个轻量级模块的输出被直接输入到用强化学习训练的目标达成和避障策略中。我们表明,将管道解耦到这些组件中会导致一个样本有效的策略学习阶段,可以在短短十几分钟内在模拟中进行全面训练。关键部分是状态表示,它经过训练,不仅可以以无监督的方式估计世界的隐藏状态,还可以帮助弥合现实差距,实现成功的仿真到现实迁移。在我们对四足机器人ANYmal的模拟和现实实验中,我们表明我们的系统可以处理嘈杂的深度图像,避免训练过程中看不到的动态障碍物,并具有局部空间意识。

英文摘要:

In this work, we present a learning-based pipeline to realise local navigation with a quadrupedal robot in cluttered environments with static and dynamic obstacles. Given high-level navigation commands, the robot is able to safely locomote to a target location based on frames from a depth camera without any explicit mapping of the environment. First, the sequence of images and the current trajectory of the camera are fused to form a model of the world using state representation learning. The output of this lightweight module is then directly fed into a target-reaching and obstacle-avoiding policy trained with reinforcement learning. We show that decoupling the pipeline into these components results in a sample efficient policy learning stage that can be fully trained in simulation in just a dozen minutes. The key part is the state representation, which is trained to not only estimate the hidden state of the world in an unsupervised fashion, but also helps bridging the reality gap, enabling successful sim-to-real transfer. In our experiments with the quadrupedal robot ANYmal in simulation and in reality, we show that our system can handle noisy depth images, avoid dynamic obstacles unseen during training, and is endowed with local spatial awareness.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑