Learning to Control an Unstable System with One Minute of Data: Leveraging Gaussian Process Differentiation in Predictive Control
Learning to Control an Unstable System with One Minute of Data: Leveraging Gaussian Process Differentiation in Predictive Control
复制标题
DOI：
10.1109/iros51168.2021.9636786
复制
发表时间：
2021-03
期刊：
影响因子：
--
通讯作者：
I. D. Rodriguez;Ugo Rosolia;A. Ames;Yisong Yue
中科院分区：
文献类型：
--
作者：
I. D. Rodriguez;Ugo Rosolia;A. Ames;Yisong Yue
我们提出了一种简单有效的方法来控制不稳定的机器人系统使用估计动力学模型。具体来说，我们展示了如何利用高斯过程的可微性来创建一个与状态相关的真正连续动力学的线性化近似，该近似可以与模型预测控制集成。我们的方法与大多数用于系统识别的高斯过程方法兼容，并且可以使用适量的训练数据学习准确的模型。我们通过学习不稳定系统的动力学来验证我们的方法，例如具有7-D状态空间和2-D输入空间的赛格威（仅使用一分钟的数据），并且我们表明所得到的控制器对未建模的动态和干扰具有鲁棒性，而基于标称模型的最先进的控制方法在小扰动下可能会失败。代码在https://github.com/learning-and-control/core上开源。
We present a straightforward and efficient way to control unstable robotic systems using an estimated dynamics model. Specifically, we show how to exploit the differentiability of Gaussian Processes to create a state-dependent linearized approximation of the true continuous dynamics that can be integrated with model predictive control. Our approach is compatible with most Gaussian process approaches for system identification, and can learn an accurate model using modest amounts of training data. We validate our approach by learning the dynamics of an unstable system such as a segway with a 7-D state space and 2-D input space (using only one minute of data), and we show that the resulting controller is robust to unmodelled dynamics and disturbances, while state-of-the-art control methods based on nominal models can fail under small perturbations. Code is open sourced at https://github.com/learning-and-control/core.