CAREER: Hypersampled Model Predictive Control: Reconciling Analog Systems with Digital Controllers

职业：超采样模型预测控制：协调模拟系统与数字控制器

• 批准号: 2046212
• 负责人: Marco Nicotra
• 金额: $ 56.74万
• 依托单位: University of Colorado at Boulder
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-08-15 至 2026-07-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2046212&HistoricalAwards=false
• 关键词: CAREER; Hypersampled; Model; Predictive; Control

This Faculty Early Career Development Program (CAREER) grant will support research related to constrained control theory, promoting its dissemination in engineering science, practice, and education. Constrained control theory is a branch of mathematics aimed at steering a dynamical system to its desired target, while also enforcing a set of safety/quality restrictions. This enabling technology allows engineering systems to live up to their full potential by operating at the limit of their specifications. Model Predictive Control (MPC) achieves this objective by solving an optimal control problem at every sampling instant. However, doing so requires a large number of computations in a relatively short amount of time, making the approach challenging to implement on high frequency engineering applications. This award supports fundamental research for the development of a new constrained control paradigm that combines systems theory, numerical methods, and algorithmic optimization to reduce the computational effort and hardware requirements of MPC. Results from this research will enable the widespread adoption of MPC in the aerospace, automotive, energy, and robotic industries, thus benefitting the U.S. economy and society. Hypersampled MPC can reduce the computational burden of traditional MPC by decoupling the prediction model, used to formulate the optimal control problem, from the control model, used to prove that the closed-loop system is asymptotically stable and satisfies constraints. This reformulation requires an extensive revision of existing MPC theory, where the two models are implicitly assumed the same. This research project will lay the foundational groundwork for Hypersampled MPC and investigate its implications on how to handle warm-starting, preconditioning, implicit recursions, parametric uncertainties, and disturbance rejection. Experimental validation will be performed on standardized educational platforms and will be released as an open source plug-and-play toolbox. The supporting documentation will then be expanded into a series of self-contained teaching modules for a hands-on laboratory class in constrained control, thereby promoting access to this transformational technology.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.

这项学院早期职业发展计划(Career)拨款将支持与约束控制理论相关的研究，促进其在工程科学、实践和教育中的传播。约束控制理论是数学的一个分支，旨在引导动力系统达到其期望的目标，同时也强制实施一组安全/质量限制。这一使能技术允许工程系统通过在其规范的限制下运行来发挥其全部潜力。模型预测控制(MPC)通过在每个采样时刻求解一个最优控制问题来实现这一目标。然而，这样做需要在相对较短的时间内进行大量的计算，这使得该方法在高频工程应用中的实现具有挑战性。该奖项支持开发一种新的约束控制范例的基础研究，该范例结合了系统理论、数值方法和算法优化，以减少MPC的计算工作量和硬件要求。这项研究的结果将使MPC在航空航天、汽车、能源和机器人行业得到广泛采用，从而使美国经济和社会受益。超采样预测控制通过将预测模型与控制模型解耦，减少了传统预测控制的计算负担，预测模型用于描述最优控制问题，用于证明闭环系统渐近稳定并满足约束条件。这种重新表述需要对现有的MPC理论进行广泛的修订，在现有的MPC理论中，两个模型隐含地假设相同。这一研究项目将为超采样预测控制奠定基础，并研究其对如何处理热启动、预条件、隐递归、参数不确定性和干扰抑制等问题的影响。实验验证将在标准化的教育平台上进行，并将作为开源即插即用工具箱发布。然后，支持文件将扩展为一系列独立的教学模块，用于受约束控制的动手实验室课程，从而促进对这项变革性技术的访问。该奖项反映了NSF的法定使命，并已通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。