Collaborative Research: CPS: Medium: Automating Complex Therapeutic Loops with Conflicts in Medical Cyber-Physical Systems

危重患者可以接受多种治疗，临床医生可以自动进行治疗，以便专注于高级别的决策任务，同时减少繁琐的监测和滴定任务。然而，多个重症监护治疗的自动化带来了独特的挑战。给病人的每一种治疗都会引起多种生理变化。因此，当对患者进行多个单独的自动化治疗时，它们可能会发生冲突，并将患者推向不安全的生理状态。无论如何，以前的大多数工作都忽略了这种冲突。为了应对这一挑战，我们将开发创新的数学算法，以实现安全的自动化患者监控、治疗指导和潜在冲突的医疗协调。我们将通过利用和发展控制、推理和优化理论来实现这一目标，从而在这些领域创造新的知识。该项目的成功完成将带来复杂的交互自动化系统的协调和调解的新方法，包括医疗保健和医疗系统。该项目将数据科学和控制理论的(网络)学科结合在一起，支持患者治疗学(物理)学科中的安全和冲突感知的自动化治疗。我们建议开发一个变革性的通用框架，可以调解冲突，并在医疗CPS中复杂的危重护理循环的自动化中确保患者安全。我们将开发(I)通过顺序过滤实现的生理监测，以使用有限的测量来跟踪患者的内部状态的安全性；(Ii)通过多变量非线性输出反馈控制以及屏障证书实现的复杂治疗循环的调解；以及(Iii)通过人口信息适应实现的治疗设置点的动态更新，以确定可安全到达的针对患者的治疗设置点。我们将使用一个复杂的重症监护复苏应用程序作为一个引人注目的案例研究来演示建议框架的有效性。该项目的成功将带来一套可部署到各种医疗保健环境中的创新技术，这些技术可以极大地提高下一代医疗CP的自主性。该奖项反映了NSF的法定使命，并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。

Critically ill patients receive multiple treatments, which can be automated for clinicians to enable focus on high-level decision-making tasks, while reducing effort on tedious monitoring and titration tasks. However, automation of multiple critical care treatments presents unique challenges. Each treatment given to a patient elicits multiple physiological changes. Thus, when multiple isolated automated treatments are given to a patient, they can conflict and drive the patient to unsafe physiological states. Regardless, most previous work has neglected such conflicts. To address this challenge, we will develop innovative mathematical algorithms to enable safe automated patient monitoring, treatment guidance, and reconciliation of potentially conflicting medical treatments. We will achieve this goal by leveraging and advancing theory of control, inference, and optimization, thereby creating new knowledge in these fields. Successful completion of this project will lead to new approaches to the coordination and mediation of complex interacting automation systems, including healthcare and medical systems.This project brings together the (cyber) disciplines of data science and control theory, in support of safe and conflict-aware automated treatments in the (physical) discipline of patient therapeutics. We propose to develop a transformative universal framework that can mediate conflicts and assure patient safety in the automation of complex critical care loops in medical CPS. We will develop (i) physiological monitoring enabled by sequential filtering to track the safety of a patient’s internal state using limited measurements; (ii) mediation of complex therapeutic loops enabled by multivariable nonlinear output-feedback control augmented by barrier certificates; and (iii) dynamic updating of treatment setpoints enabled by population-informed adaptation to determine safely reachable patient-specific treatment setpoints. We will demonstrate the efficacy of the proposed framework using a complex critical care resuscitation application as a compelling case study. The success of this project will lead to a suite of innovative technologies deployable to various medical care settings, which can drastically advance autonomy in next-generation medical CPS.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.