arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.03787cs.ROcs.AIcs.LG
收藏

约束流形上的神经操作规划

Neural Manipulation Planning on Constraint Manifolds

University of California San Diego(加州大学圣迭戈分校)

Ahmed H. Qureshi, Jiangeng Dong, Austin Choe, Michael C. Yip

摘要:

任务约束的存在对运动规划提出了重大挑战。尽管所有最近的进步,现有的算法仍然是计算昂贵的大多数规划问题。在本文中,我们提出了约束运动规划网络(CoMPNet),多模态运动约束的第一个神经规划器。我们的方法包括以下组件:i)约束和环境感知编码器; ii)神经机器人配置生成器,其输出约束流形上/附近的配置;以及iii)双向规划算法,其采用所生成的配置来创建可行的机器人运动轨迹。我们表明,CoMPNet解决了实际的运动规划任务,涉及无约束和约束的问题。此外,它推广到对象的新的不可见位置,即,在特定的环境中,成功率很高,这在训练中是看不到的。当与最先进的约束运动规划算法相比时,CoMPNet在计算速度上的数量级改进优于显著更低的方差。

英文摘要:

The presence of task constraints imposes a significant challenge to motion planning. Despite all recent advancements, existing algorithms are still computationally expensive for most planning problems. In this paper, we present Constrained Motion Planning Networks (CoMPNet), the first neural planner for multimodal kinematic constraints. Our approach comprises the following components: i) constraint and environment perception encoders; ii) neural robot configuration generator that outputs configurations on/near the constraint manifold(s), and iii) a bidirectional planning algorithm that takes the generated configurations to create a feasible robot motion trajectory. We show that CoMPNet solves practical motion planning tasks involving both unconstrained and constrained problems. Furthermore, it generalizes to new unseen locations of the objects, i.e., not seen during training, in the given environments with high success rates. When compared to the state-of-the-art constrained motion planning algorithms, CoMPNet outperforms by order of magnitude improvement in computational speed with a significantly lower variance.

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。

↑