Towards Safe and Robust Closed-Loop Artificial Pancreas Using Improved PID-Based Control Strategies

Alshalalfah, Abdel-Latif; Hamad, Ghaith Bany; Mohamed, Otmane Aït

doi:10.1109/tcsi.2021.3058355

Cited by 15 publications

(6 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These control methods use mathematical models of glucose dynamics, and a small number of samples is required for data efficiency. PID control is also comparable [64] to model-based control in terms of performance, and PID with adaptive weights [65] has been proposed recently. In [66], linear parameter-varying control was incorporated with insulin, food intake, and metabolism subsystems.…”

Section: Classical Control For Glucose Regulationmentioning

confidence: 99%

A Blood Glucose Control Framework Based on Reinforcement Learning With Safety and Interpretability: In Silico Validation

et al. 2021

View full text Add to dashboard Cite

Controlling blood glucose levels in diabetic patients is important for managing their health and quality of life. Several algorithms based on model predictive control and reinforcement learning (RL) have been proposed so far, most of which use prior knowledge of physiological systems, the mathematical structure of blood glucose dynamics, and many episodes including failures for training the policy network in RL. To be smoothly adopted in clinical settings, we propose a fast online learning method underlining safety and interpretability. A random forest regressor and a dual attention network were exploited for glucose prediction and extension of state variables. The soft actor-critic network to determine insulin dosing was guided by proportional-integral-derivative (PID) control in the early phase, and an adaptive safe actor with suspension and additional insulin dosing was incorporated. The performance of the models was validated using an FDA-approved type 1 diabetes simulator. The results showed comparable outcomes with PID control. Using this system, glucose dynamics could be captured despite minimal prior knowledge. The extended state variables were correlated with basic states such as glucose, insulin, and meal intake, their derivatives, and their integrals, which can be fundamental elements of mathematical modeling of physiological responses. Attention scores and attribution scores in the prediction and control models represented the focused features and the internal operation of the models with interpretability. We expect this study to provide some insights on how RL can be practically adopted in clinical environments and how interpretability can provide hints of machines' thoughts for clinical applications.INDEX TERMS blood glucose control, reinforcement learning, safe and interpretable control, in silico validation, simulation for clinical application

show abstract

Section: Classical Control For Glucose Regulationmentioning

confidence: 99%

A Blood Glucose Control Framework Based on Reinforcement Learning With Safety and Interpretability: In Silico Validation

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Existing control algorithms can be broadly discerned to be reactive and predictive. Reactive control action includes control algorithms such as proportional-integral and derivative (PID), fuzzy, and other approaches that compute insulin based on current measurements and lack predictive capabilities [ 7 , 8 , 9 ]. Some additional capabilities are available in the reactive controllers as well.…”

Section: Introductionmentioning

confidence: 99%

Human Digital Twin for Personalized Elderly Type 2 Diabetes Management

Padmapritha

Srinivasan

Kesavadev

et al. 2023

JCM

View full text Add to dashboard Cite

Managing Elderly type 2 diabetes (E-T2D) is challenging due to geriatric conditions (e.g., co-morbidity, multiple drug intake, etc.), and personalization becomes paramount for precision medicine. This paper presents a human digital twin (HDT) framework to manage E-T2D that exploits various patient-specific data and builds a suite of models exploiting the data for prediction and management to personalize diabetes treatment in E-T2D patients. These models include mathematical and deep-learning ones that capture different patient aspects. Consequently, the HDT virtualizes the patient from different viewpoints using an HDT that mimics the patient and has interfaces to update the virtual models simultaneously from measurements. Using these models the HDT obtains deeper insights about the patient. Further, an adaptive patient model fusing this information and a learning-based model predictive control (LB-MPC) algorithm are proposed. The geriatric conditions are captured as model parameters and constraints while solving the LB-MPC to personalize the insulin infusion for E-T2D management. The HDT is deployed on and illustrated with 15 patients using clinical trials and simulations. Our results show that HDT helps improve the time-in-range from 3–75% to 86–97% and reduces insulin infusion by 14–29%.

show abstract

“…In addition, cascade control structures are introduced in [10], [11], [12], [13], and [14]. Cascade control structure is realized by sensors feedback.…”

Section: Introductionmentioning

confidence: 99%

“…This point is proved mathematically in this paper. Secondly, compared with the traditional methods [10], [11], [12], [13], [14], the proposed method uses only the outermost loop feedback which can reduce the number of feedback loops and sensors and without sacrificing control performance. In addition, direct current motor speed/position systems are given to verify the theoretical analysis and also show that the control performance of the generalized PID control with FFC is satisfactory.…”

Section: Introductionmentioning

confidence: 99%

Equivalence Analysis of Cascade Control for a Class of Cascade Integral Systems

2023

View full text Add to dashboard Cite

A class of generalized proportional-integral-derivative (PID) control with feedforward compensation (FFC) is obtained for a class of cascade integral systems by equivalence analysis of the cascade control and such the analysis is obtained by using the information of the model and outermost loop feedback. Firstly, a new type of error related to the traditional error such as proportional (P) or proportional integral (PI) is given. Secondly, by analyzing cascade control for a class of ideal cascade integral systems, the generalized PID control with FFC is presented based on the proposed new type of error. Then, the generalized PID control with FFC is extended to a class of non-ideal cascade integral systems to reduce the number of feedback loop and sensors. Finally, the simulation results of the speed/position servo system of direct current motor are given to verify the theoretical analysis results.INDEX TERMS Equivalence analysis of cascade control, generalized PID control, class of cascade integral systems, new definition of the traditional error.

show abstract

Towards Safe and Robust Closed-Loop Artificial Pancreas Using Improved PID-Based Control Strategies

Cited by 15 publications

References 38 publications

A Blood Glucose Control Framework Based on Reinforcement Learning With Safety and Interpretability: In Silico Validation

A Blood Glucose Control Framework Based on Reinforcement Learning With Safety and Interpretability: In Silico Validation

Human Digital Twin for Personalized Elderly Type 2 Diabetes Management

Equivalence Analysis of Cascade Control for a Class of Cascade Integral Systems

Contact Info

Product

Resources

About