Fundamental Challenges in Deep Learning for Stiff Contact Dynamics
Fundamental Challenges in Deep Learning for Stiff Contact Dynamics
复制标题
DOI：
10.1109/iros51168.2021.9636383
复制
发表时间：
2021-03
期刊：
影响因子：
--
通讯作者：
Mihir Parmar-;Mathew Halm;Michael Posa
中科院分区：
文献类型：
--
作者：
Mihir Parmar-;Mathew Halm;Michael Posa
摩擦接触作为腿运动和操纵的核心基础行为已经被广泛研究，并且其几乎不连续的性质使得规划和控制困难，即使当机器人的精确模型可用时。在这里，我们提出的经验证据表明，首先学习一个准确的模型可能会受到接触的干扰，因为现代深度学习方法并不是为了捕捉这种非平滑性而设计的。我们隔离接触的非光滑的影响，通过改变顺应性接触模拟器的机械刚度。即使对于一个简单的系统，我们发现刚度本身也会显着降低训练过程，泛化和数据效率。我们的研究结果提出了严重的问题，模拟测试环境，不准确地反映刚性机器人硬件的刚度。显著的额外调查将是必要的，以充分了解和减轻这些影响，我们建议未来的研究几种途径。
Frictional contact has been extensively studied as the core underlying behavior of legged locomotion and manipulation, and its nearly-discontinuous nature makes planning and control difficult even when an accurate model of the robot is available. Here, we present empirical evidence that learning an accurate model in the first place can be confounded by contact, as modern deep learning approaches are not designed to capture this non-smoothness. We isolate the effects of contact’s non-smoothness by varying the mechanical stiffness of a compliant contact simulator. Even for a simple system, we find that stiffness alone dramatically degrades training processes, generalization, and data-efficiency. Our results raise serious questions about simulated testing environments which do not accurately reflect the stiffness of rigid robotic hardware. Significant additional investigation will be necessary to fully understand and mitigate these effects, and we suggest several avenues for future study.