arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1909.07876cs.ROcs.AIcs.LG
收藏

学习使用固定状态表示操作对象集合

Learning to Manipulate Object Collections Using Grounded State Representations

University of Utah(犹他大学);NVIDIA(英伟达)

Matthew Wilson, Tucker Hermans

摘要:

我们提出了一种方法,模拟到真实的机器人学习,利用模拟器的状态信息的方式,可扩展到许多对象。我们首先训练一对编码器网络来捕获潜在空间中的多对象状态信息。其中一个编码器是CNN,它使我们的系统能够在现实世界中对RGB图像进行操作;另一个是图神经网络(GNN)状态编码器,它直接消耗一组原始对象姿势,并实现更准确的奖励计算和价值估计。一旦经过训练,我们就在强化学习算法中使用这些编码器来训练可以操纵许多对象的基于图像的策略。我们评估我们的方法上的任务,将一组对象推到所需的桌面区域。与仅依赖于图像或使用固定长度状态编码的方法相比,我们的方法实现了更高的成功率,在现实世界中表现良好,无需微调,并且可以推广到训练过程中看不到的不同数量和类型的对象。

英文摘要:

We propose a method for sim-to-real robot learning which exploits simulator state information in a way that scales to many objects. We first train a pair of encoder networks to capture multi-object state information in a latent space. One of these encoders is a CNN, which enables our system to operate on RGB images in the real world; the other is a graph neural network (GNN) state encoder, which directly consumes a set of raw object poses and enables more accurate reward calculation and value estimation. Once trained, we use these encoders in a reinforcement learning algorithm to train image-based policies that can manipulate many objects. We evaluate our method on the task of pushing a collection of objects to desired tabletop regions. Compared to methods which rely only on images or use fixed-length state encodings, our method achieves higher success rates, performs well in the real world without fine tuning, and generalizes to different numbers and types of objects not seen during training.

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。

↑