An official website of the United States government
Abstract This paper concerns the design and rigorous in silico evaluation of a closed-loop hemorrhage resuscitation algorithm with blood pressure (BP) as controlled variable. A lumped-parameter control design model relating volume resuscitation input to blood volume (BV) and BP responses was developed and experimentally validated. Then, three alternative adaptive control algorithms were developed using the control design model: (i) model reference adaptive control (MRAC) with BP feedback, (ii) composite adaptive control (CAC) with BP feedback, and (iii) CAC with BV and BP feedback. To the best of our knowledge, this is the first work to demonstrate model-based control design for hemorrhage resuscitation with readily available BP as feedback. The efficacy of these closed-loop control algorithms was comparatively evaluated as well as compared with an empiric expert knowledge-based algorithm based on 100 realistic virtual patients created using a well-established physiological model of cardiovascular (CV) hemodynamics. The in silico evaluation results suggested that the adaptive control algorithms outperformed the knowledge-based algorithm in terms of both accuracy and robustness in BP set point tracking: the average median performance error (MDPE) and median absolute performance error (MDAPE) were significantly smaller by >99% and >91%, and as well, their interindividual variability was significantly smaller by >88% and >94%. Pending in vivo evaluation, model-based control design may advance the medical autonomy in closed-loop hemorrhage resuscitation.
more »
« less
A Generative Approach to Testing the Performance of Physiological Control Algorithms
Abstract Physiological closed-loop control algorithms play an important role in the development of autonomous medical care systems, a promising area of research that has the potential to deliver healthcare therapies meeting each patient's specific needs. Computational approaches can support the evaluation of physiological closed-loop control algorithms considering various sources of patient variability that they may be presented with. In this article, we present a generative approach to testing the performance of physiological closed-loop control algorithms. This approach exploits a generative physiological model (which consists of stochastic and dynamic components that represent diverse physiological behaviors across a patient population) to generate a select group of virtual subjects. By testing a physiological closed-loop control algorithm against this select group, the approach estimates the distribution of relevant performance metrics in the represented population. We illustrate the promise of this approach by applying it to a practical case study on testing a closed-loop fluid resuscitation control algorithm designed for hemodynamic management. In this context, we show that the proposed approach can test the algorithm against virtual subjects equipped with a wide range of plausible physiological characteristics and behavior and that the test results can be used to estimate the distribution of relevant performance metrics in the represented population. In sum, the generative testing approach may offer a practical, efficient solution for conducting preclinical tests on physiological closed-loop control algorithms.
more »
« less
- Award ID(s):
- 1760817
- PAR ID:
- 10575884
- Publisher / Repository:
- ASME
- Date Published:
- Journal Name:
- ASME Letters in Dynamic Systems and Control
- Volume:
- 4
- Issue:
- 3
- ISSN:
- 2689-6117
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This paper addresses the problem of learning the optimal control policy for a nonlinear stochastic dynam- ical. This problem is subject to the ‘curse of dimension- ality’ associated with the dynamic programming method. This paper proposes a novel decoupled data-based con- trol (D2C) algorithm that addresses this problem using a decoupled, ‘open-loop - closed-loop’, approach. First, an open-loop deterministic trajectory optimization problem is solved using a black-box simulation model of the dynamical system. Then, closed-loop control is developed around this open-loop trajectory by linearization of the dynamics about this nominal trajectory. By virtue of linearization, a linear quadratic regulator based algorithm can be used for this closed-loop control. We show that the performance of D2C algorithm is approximately optimal. Moreover, simulation performance suggests a significant reduction in training time compared to other state of the art algorithms.more » « less
-
Abstract This paper presents a virtual patient generator (VPG) intended to be used for preclinical in silico evaluation of autonomous vasopressor administration algorithms in the setting of experimentally induced vasoplegia. Our VPG consists of two main components: (i) a mathematical model that replicates physiological responses to experimental vasoplegia (induced by sodium nitroprusside (SNP)) and vasopressor resuscitation via phenylephrine (PHP) and (ii) a parameter vector sample generator in the form of a multidimensional probability density function (PDF) using which the parameters characterizing the mathematical model can be sampled. We developed and validated a mathematical model capable of predicting physiological responses to the administration of SNP and PHP. Then, we developed a parameter vector sample generator using a collective variational inference method. In a blind testing, the VPG developed by combining the two could generate a large number of realistic virtual patients (VPs), which could simulate physiological responses observed in all the experiments: on the average, 98.1% and 74.3% of the randomly generated VPs were physiologically legitimate and adequately replicated the test subjects, respectively, and 92.4% of the experimentally observed responses could be covered by the envelope formed by the subject-replicating VPs. In sum, the VPG developed in this paper may be useful for preclinical in silico evaluation of autonomous vasopressor administration algorithms.more » « less
-
null (Ed.)Deep neural network (DNN) has become increasingly popular and DNN testing is very critical to guarantee the correctness of DNN, i.e., the accuracy of DNN in this work. However, DNN testing suffers from a serious efficiency problem, i.e., it is costly to label each test input to know the DNN accuracy for the testing set, since labeling each test input involves multiple persons (even with domain-specific knowledge) in a manual way and the testing set is large-scale. To relieve this problem, we propose a novel and practical approach, called PACE (which is short for P ractical AC curacy E stimation), which selects a small set of test inputs that can precisely estimate the accuracy of the whole testing set. In this way, the labeling costs can be largely reduced by just labeling this small set of selected test inputs. Besides achieving a precise accuracy estimation, to make PACE more practical it is also required that it is interpretable, deterministic, and as efficient as possible. Therefore, PACE first incorporates clustering to interpretably divide test inputs with different testing capabilities (i.e., testing different functionalities of a DNN model) into different groups. Then, PACE utilizes the MMD-critic algorithm, a state-of-the-art example-based explanation algorithm, to select prototypes (i.e., the most representative test inputs) from each group, according to the group sizes, which can reduce the impact of noise due to clustering. Meanwhile, PACE also borrows the idea of adaptive random testing to select test inputs from the minority space (i.e., the test inputs that are not clustered into any group) to achieve great diversity under the required number of test inputs. The two parallel selection processes (i.e., selection from both groups and the minority space) compose the final small set of selected test inputs. We conducted an extensive study to evaluate the performance of PACE based on a comprehensive benchmark (i.e., 24 pairs of DNN models and testing sets) by considering different types of models (i.e., classification and regression models, high-accuracy and low-accuracy models, and CNN and RNN models) and different types of test inputs (i.e., original, mutated, and automatically generated test inputs). The results demonstrate that PACE is able to precisely estimate the accuracy of the whole testing set with only 1.181%∼2.302% deviations, on average, significantly outperforming the state-of-the-art approaches.more » « less
-
Neural networks present a useful framework for learning complex dynamics, and are increasingly being considered as components to closed loop predictive control algorithms. However, if they are to be utilized in such safety-critical advisory settings, they must be provably "conformant" to the governing scientific (biological, chemical, physical) laws which underlie the modeled process. Unfortunately, this is not easily guaranteed as neural network models are prone to learn patterns which are artifacts of the conditions under which the training data is collected, which may not necessarily conform to underlying physiological laws. In this work, we utilize a formal range-propagation based approach for checking whether neural network models for predicting future blood glucose levels of individuals with type-1 diabetes are monotonic in terms of their insulin inputs. These networks are increasingly part of closed loop predictive control algorithms for "artificial pancreas" devices which automate control of insulin delivery for individuals with type-1 diabetes. Our approach considers a key property that blood glucose levels must be monotonically decreasing with increasing insulin inputs to the model. Multiple representative neural network models for blood glucose prediction are trained and tested on real patient data, and conformance is tested through our verification approach. We observe that standard approaches to training networks result in models which violate the core relationship between insulin inputs and glucose levels, despite having high prediction accuracy. We propose an approach that can learn conformant models without much loss in accuracy.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1115/1.4065934
