Robust Data-Driven Control for Safety-Critical Systems

该项目旨在开发在安全关键应用中控制机器人和网络物理系统的新方法。该项目希望在如何利用机器学习的力量进行机器人控制方面产生新的知识，同时始终保证安全性和稳定性。该项目的成果将是新的算法和对数据、学习和模型之间相互作用的更深入理解，以及在手术机器人和双足行走机器人上的实验验证。该项目将通过大幅增加应用范围来提供显著的好处，在这些应用中，机器学习的力量可以安全地应用于提高机器人的能力和吸收。

This project aims to develop new approaches to controlling robotic and cyber-physical systems in safety-critical applications. This project expects to generate new knowledge in how to harness the power of machine learning for robot control, while guaranteeing safety and stability at all times. The outcomes of this project will be new algorithms and a deeper understanding of the interplay of data, learning, and models, as well as experimental validation on a surgical robot and a bipedal walking robot. This project will provide significant benefits by dramatically increasing the range of applications in which the power of machine learning can be safely applied to advance the capabilities and uptake of robotics.