CAREER: A Hybrid Filtering and Robust Control Framework for Legged Robot Locomotion on Dynamic Rigid Surfaces

这个教师早期职业发展（Career）项目将专注于创建新的方法来建模，估计和控制有腿机器人的运动，以实现在动态刚性表面（DRS）（即移动且不变形的表面）上的稳定运动。虽然今天的腿式机器人系统在穿越静止表面（如楼梯、沙子和草地）方面表现出了非凡的能力，但DRS上的腿式运动（如船舶、飞机和火车）是一种新的机器人功能，尚未得到解决。这种新功能将使腿式机器人能够适应复杂、动态的人类环境（这对轮式或履带式机器人来说是非常具有挑战性的），从而使它们能够帮助许多关键的高风险应用，例如船上消防和灭火，以及公共交通工具的清洁/消毒，以遏制传染病的传播。实现这些功能需要可靠的机器人估计和控制，由于相关机器人行为的高度复杂性，这些行为是混合的（包括连续的腿部摆动运动和离散的脚着地事件），并且受时变DRS运动的影响，因此具有很大的挑战性。CAREER研究计划旨在解决这些基本问题，并为能够在非静止表面上自主导航的下一代有腿机器人系统的开发奠定基础。CAREER教育项目将加强马萨诸塞大学洛厄尔分校的机器人课程，同时吸引不同群体参与机器人教育和研究，包括代表性不足的本科生和研究生、K-12学生和普通公众。本课题的研究目标是利用动态建模、状态估计、反馈控制和混合系统理论来推进腿式机器人的控制理论，以实现DRS上可证明的腿式稳定运动。为了实现研究目标，将追求四个主要目标：(i)制定一个基于物理的模型，该模型可以捕获与DRS上的腿运动相关的混合时变机器人动力学；（ii）创建设计状态估计器的新方法，通过将可证明的不变滤波方法从连续系统扩展到混合动态系统，包括在DRS上移动的有腿机器人，从而实现具有收敛保证的实时状态估计；（iii）推导了一种基于李雅普诺夫的控制器设计方法，通过处理在连续阶段和离散事件中存在不确定性的混合时变机器人动力学，在DRS上产生稳定的运动；（iv）将建模、状态估计和控制器设计集成到一个基于模型的框架中，该框架可证明在DRS上维持腿式运动。该项目将支持PI解决超越当前技术水平的主要机器人挑战，并帮助建立机器人和控制方面的长期职业生涯。该项目由跨部门机器人基础研究项目支持，由工程（ENG）和计算机与信息科学与工程（CISE）联合管理和资助。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This Faculty Early Career Development (CAREER) project will focus on creating new methods to model, estimate, and control the movement of legged robots for enabling stable locomotion on dynamic rigid surfaces (DRS) (i.e., surfaces that move and do not deform). While today’s legged robot systems have demonstrated remarkable capabilities in traversing stationary surfaces (e.g., stairs, sand, and grass), legged locomotion on DRS (e.g., ships, aircraft, and trains) is a new robot functionality that has not been addressed. This new functionality will empower legged robots to negotiate complex, dynamic human environments (that are prohibitively challenging for wheeled or tracked robots) to allow them to aid in numerous critical high-risk applications, such as shipboard firefighting and fire suppression and cleaning/disinfection of public transportation vehicles to contain the spread of infectious diseases. Enabling such functionality demands reliable robot estimation and control, which are substantially challenging due to the high complexity of the associated robot behaviors that are hybrid (involving continuous leg-swinging motions and discrete foot-landing events) and subject to the time-varying DRS movement. The CAREER research program seeks to solve these fundamental problems and lay a foundation for the development of next-generation legged robot systems capable of autonomous navigation on nonstationary surfaces. The CAREER education program will enhance the robotics curriculum at the University of Massachusetts Lowell while engaging diverse groups, including underrepresented undergraduate and graduate students, K-12 students, and the general public, in robotics education and research.The research goal of the project is to draw upon dynamic modeling, state estimation, feedback control, and theory of hybrid systems to advance the control theory of legged robots in order to realize provably stable legged locomotion on a DRS. To achieve the research goal, four main objectives will be pursued: (i) formulation of a physics-based model that captures the hybrid, time-varying robot dynamics associated with legged locomotion on a DRS; (ii) creation of new methods of designing state estimators that achieve real-time state estimation with convergence guarantees by provably expanding an invariant filtering methodology from continuous systems to hybrid dynamical systems that include legged robots moving on a DRS; (iii) derivation of a Lyapunov-based controller design methodology to produce stable locomotion on a DRS by handling the hybrid, time-varying robot dynamics under uncertainties that reside in both continuous phases and discrete events; and (iv) integration of the modeling, state estimation, and controller design into a model-based framework that provably sustains legged locomotion on a DRS. The project will support the PI to solve major robotics challenges beyond the capabilities of the state of the art, and help establish a long-term career in robotics and control.This project is supported by the cross-directorate Foundational Research in Robotics program, jointly managed and funded by the Directorates for Engineering (ENG) and Computer and Information Science and Engineering (CISE).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.