A description of our work on artificial pancreas systems for treating type-1 diabetes.
Note: Work on this particular NSF project concluded in 2019. We will post details of our current work on artificial pancreas devices soon.
Introduction
This page describes collaborative research funded by the US National Science Foundation (NSF) under
awards CPS-1446900 and CPS-1446751.
All opinions expressed here are those of the authors and not necessarily of the US National Science Foundation (NSF).
This project is a collaboration between engineers, computer
scientists, mathematicians, and clinical researchers centered around
the mathematical modeling, simulation and formal analysis of the
artificial pancreas. The artificial pancreas is a concept that
involves the development of a series of increasingly sophisticated
/Cyber-Physical Systems/ that will automate the delivery of insulin to
people with Type-1 Diabetes.
Background on Type-1 Diabetes
What follows below is a simple (possibly simplistic) explanation that omits many key details. It is suitable as an overview
and not intended to be a clinical presentation. The interested reader should consult a book. I recommend the Pink Panther Book by Drs. Chase and Maahs.
What is Type-1 Diabetes?
Type-1 Diabetes (T1D) is a serious condition caused by the inability
or the decreased ability of the pancreas to secrete insulin. This can
happen due to numerous reasons: for example, as a result of an
autoimmune response that destroys the beta cells of the pancreas; or
as a result of certain surgical procedures. Insulin is an important
hormone that is responsible for regulating the uptake of glucose by
various cells in the body causing the lowering of blood glucose
levels. Lack of insulin leads to increased blood glucose
levels. However, the glucose in the blood is not taken up by the cells
that convert them to energy or store them. As a result, the cells are
starved of much needed energy and this results in their breakdown.
T1D Treatment
Currently, T1D is treated through the external administration of
artificial insulin analogs either through daily injections or an
insulin infusion pump. The latter option is increasingly popular with
patients, since it provides substantial freedom in their choice of
meals, exercise and other activities that affect their metabolism. The
infusion pumps are operated manually by the patient based on their
anticipated mealtimes, meal carbohydrate amounts and exercise
times/intensities. The infusion pumps currently available in the
market can deliver short acting insulin analogs at programmable doses,
all day and night. This typically involves a background “basal”
delivery of insulin and numerous boluses delivered at
mealtimes. Manual control of pump has two main drawbacks:
It imposes the burden of managing the blood glucose levels on the
patients. Typically this involves roughly 6-9 blood glucose
measurements, basal adjustments and bolus infusions throughout the
day. Failure to maintain blood glucose levels inside the euglycemic
range [70,170] mg/dl can have serious consequences. If too much
insulin is infused, it can drive the blood glucose levels dangerously
low, causing hypoglycemia. If too little is infused, it can cause high
blood gluocse levels with short term (ketacidosis) and long term
consequences (damage to kidneys, heart, nerves and eyes).
The Artificial Pancreas
The artificial pancreas (AP) is a series of
increasingly sophisticated devices that will increasingly automate
insulin delivery to the patient. It has been enabled by insulin
infusion pumps that can deliver insulin automatically and be
programmed externally by a software-based controller. On the other
hand, glucose sensor technology has developed sufficiently to provide
accurate, near realtime subcutaneous glucose measurement. Many stages
have been envisioned for the AP, and the technological feasibility of
each stage has been demonstrated in clinical trials. The original
stages of the AP were proposed by the JDRF as follows:
Pump shutoff: this stage simply shuts down the pump and alarms when the blood glucose is too low. This technology has already been useful at nighttimes to alert patients and their care providers when the blood glucose level drops.
Predictive Pump shutoff: this device forecasts the near term future glucose trend form the current observed glucose values. If the forecast calls for a low in the next 30-70 minutes, the pump is shutoff in anticipation of the low and turned back on as soon as the glucose levels rise.
Hybrid Closed Loop: This stage controls the basal insulin delivery, relying on daytime boluses by the user.
Fully automatic closed loop: This stage controls both basal delivery and automatic meal boluses.
Multihormone closed loop: This stage uses multiple hormones instead of just insulin. Specifically, glucagon (the counterregulatory hormone to insulin) is also used to increase blood glucose levels and in some designs, amylin is added to control the digestion of food.
Recently, it has been noted that all stages have been implemented and
proven to be technologically viable in various stages of clinical
trials. A simpler classification has been proposed by the JDRF as (a)
daytime vs. nighttime-only control and (b) insulin-only
vs. multi-hormonal control.
Formal Verification
The formal verification project proposes to verify AP control systems to find possible issues with these systems before deployment.
What is formal verification?
Engineered systems are often prone to design faults that arise due to
numerous reasons, including design and implementation mistakes. These
are particularly problematic with software systems that tend to be
designed by a large team of people. Verification is the process by
which, these devices are tested and known defects that arise are
fixed. However, manual verification is expensive and prone to errors,
as well. Therefore, formal verification techniques try to automate the
verification process to provide more guarantees at the end of the
verification. Ideally, these are exhaustive techniques that either
provide a mathematical proof that the system is correct or expose a
bug/defect in the design.
However, there are fundamental reasons why fully exhaustive and
automated verification of software systems is a really hard problem,
if not downright impossible! Nevertheless, we have been trying many
partial solutions that have yielded varying degrees of success: these
solutions focus on solving the problem for special classes of systems
or abstracting the system model into a special form before reasoning
about them.
Why verify the Artificial Pancreas?
The artificial pancreas is safety critical. Ultimately, it takes
autonomy away from patients with T1D in return for a reduced burden of
diabetes management. It is essential that the control software be free
of common software bugs such a buffer overflows, division by zero, and
memory leaks. However, the AP is not just software: it includes
physical components like sensors, pumps and an entire human body in
the loop! As such, it has to deal with a large set of uncertainties in
its operation. Examples include unannounced meals, exercise, sensor
noise, erroneous sensor readings, pressure induced sensor attenuation,
set failures, and patient physiological changes. Besides software
bugs, we need to exhaustively run our designs through billions of
scenarios to see if the system behaves “reasonably” in all
scenarios. Worst of all, we need a good definition of what
“reasonable” is.
Aren’t Clinical Trials Good Enough?
Note: Clinical trials differ widely in what they study, and how they
study. The answer below is not truly representative of all possible
studies that have been carried out for AP controllers
In some sense, clinical trials will be “good enough” when the
technology is deployed to a large population of users and tested for a
long time under varying conditions. Otherwise, most trials are not
meant to test for the “worst case”. Worst case effects are by
definition rare and only seen when a product is deployed.
What stops us from verifying AP controllers?
As mentioned earlier, verification is a very hard problem. We have
billions, if not trillions, of scenarios to run through and test. It
is hard to go through each one by one. As a result, mathematical
models should be used by engineers to capture processes like the human
insulin-glucose response, the patterns of sensor noise, set failures
and other disturbances. There has been a lot of work on simulating AP
controllers. In fact, some models like the Dalla-Man et al. model have
been used with FDA’s encouragement as a stand in for animal trials of
AP controllers. However, these models are nonlinear and as such, many
verification tools that exist cannot handle these models or can be
very slow. However, we are making progress on tools like S-Taliro,
Flow* and dReal, that can perform restricted forms of automated
verification for nonlinear dynamics.
To conclude, a significant gap needs to be bridged between what formal
tools can do currently, and what is needed to automate an exhaustive
analysis of AP controllers.
Project Goals
The basic goal of this project is to provide a focused effort around
developing formal tools for verifying artificial pancreas
controllers. The project has multiple thrusts including:
. Developing disturbance models for meal, exercise, sensor noise, dropouts, set
failures and other disturbances.
. Developing simpler, perhaps even nondeterminstic models, of insulin-glucose response. We are looking
into delay differential models and nondeterministic “minimal models”
built using the available data.
. Developing formal verification tools
for nonlinear dynamics so that they can reason about controller
implementations and nonlinear dynamics models.
. Developing a set of
properties that need to be verified. Case-studies on real-life
systems that have been deployed in clinical trials.
. Last but not least, Develop ways to
present verification results to engineers and clinicians to better
understand the root cause of defects and judge their real-life
feasibility.
Project Team
The project involves a mutidisciplinary team of computer scientists, chemical engineers, mathematicians and clinical researchers from multiple institutions.
Principal Investigators
Faye Cameron (joint PI), Chemical Engg. RPI.
Sriram Sankaranarayanan (joint PI), Computer Science, CU Boulder.
David Maahs (joint PI), Pediatric Endocrinology, Barbara Davis Center for Childhood Diabetes, UC Denver Medical School.
B. Wayne Bequette (senior person), Chemical Engg., RPI.
David Bortz (co-PI), Applied Mathematics, CU Boulder.
Shalom Ruben (co-PI), Mechanical Engg., CU Boulder.
Students
Taisa Kushner (PhD, IQ Biology), CU Boulder.
Alumni
Suhas Akshar Kumar
Ram Das Diwakaran
Alexandra Okeson
Dr. Xin Chen
Collaborators
Greg Forlenza, MD, Pediatric Endocrinology, Barbara Davis Center for Childhood Diabetes, UC Denver Medical School.
Laurel Messer, PhD, School of Nursing and Barbara Davis Center for Childhood Diabetes, UC Denver Medical School.
Publications
Weighted Transducers for Robustness Verification
Emmanuel Filiot,
Nicolas Mazzocchi,
Jean-Francois Raskin,
Sriram Sankaranarayanan,
and Ashutosh Trivedi
In Intl. Conference on Concurrency Theory (CONCUR),
pp. 17:1–17:21,
2020.
Automata theory provides us with fundamental notions such as
languages, membership, emptiness and inclusion that in turn allow us
to specify and verify properties of reactive systems in a useful
manner. However, these notions all yield "yes"/"no" answers that
sometimes fall short of being satisfactory answers when the models
being analyzed are imperfect, and the observations made are prone to
errors. As a result, a common engineering approach is not just to
verify that a system satisfies a property, but it does so
robustly. In this paper, we present notions of robustness
that place a metric on words, thus providing a natural notion of
distance between words. Such a metric naturally leads to a
topological neighborhood of words and languages, leading to
quantitative and robust versions of the membership, emptiness and
inclusion problems. The main contribution of this work is to study
such problems in the context of weighted transducers. We provide
algorithms for solving the robust and quantitative versions of the
membership and inclusion problems while providing useful motivating
case studies including a modeling an error-prone human operator and
approximate pattern matching in a large type-1 diabetes dataset.
Multi-hour Blood Glucose Prediction in T1D: A Patient Specific Approach using Shallow Neural Network Models
Taisa Kushner,
Marc D. Breton,
and Sriram Sankaranarayanan
Diabetes Technology and Therapeutics,
22
(12)
pp. 883-891,
2020.
Considering current insulin action profiles and the nature of glycemic responses to insulin, there is an acute need for longer-term, accurate, blood glucose predictions to inform insulin dosing schedules and enable effective decision support for the treatment of type 1 diabetes (T1D). However, current methods achieve acceptable accuracy only for prediction horizons of up to one hour, whereas typical post-prandial excursions and insulin action profiles last 4-6 hours. In this work, we present models for prediction horizons of 60-240min developed by leveraging “shallow” neural networks, allowing for significantly lower complexity when compared to related approaches. Methods: Patient-specific neural network-based predictive models are developed and tested on previously collected data from a cohort of twenty-four subjects with T1D. Models are designed to avoid serious pitfalls through incorporating essential physiological knowledge into model structure. Patient-specific models were generated to predict glucose 60, 90, 120, 180, and 240 minutes ahead, and a "transfer learning" approach to improve accuracy for patients where data is limited. Finally, we determined subgroup characteristics which result in higher model accuracy overall. Results: RMSE was 28+/- 4 mg/dL, 33+/-4 mg/dL, 38±6 mg/dL, 40+/-8 mg/dL, and 43+/-12mg/dL for 60, 90, 120, 180, and 240minutes respectively. For all prediction horizons, at least 93% of predictions were clinically acceptable by the Clarke Error Grid. Variance of historic continuous glucose monitor (CGM) values was a strong predictor for the need of transfer learning approaches. Conclusions: A shallow neural network, using features extracted from past CGM data and insulin logs is able to achieve multi-hour glucose predictions with satisfactory accuracy. Models are patient-specific, learned on readily-available data without the need for additional tests, and improve accuracy while lowering complexity when compared to related approaches, paving the way for new advisory and closed loop algorithms able to encompass most of the insulin action time-frame.
Conformance verification for neural network models of glucose-insulin dynamics
Taisa Kushner,
Sriram Sankaranarayanan,
and Marc Breton
In Hybrid Systems: Computation and Control (HSCC),
pp. 13:1–13:12,
2020.
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.
Models, Devices, Properties and Verification of Artificial Pancreas Systems
Taisa Kushner,
B. Wayne Bequette,
Faye Cameron,
Gregory Forlenza,
David Maahs,
and Sriram Sankaranarayanan
In Automated Reasoning for Systems Biology And Medicine (Edited by Paulo Zuliani and Pietro Lio),
pp. 93-131,
2019.
In this chapter, we present the interplay between models of human
physiology, closed loop medical devices, correctness specifications
and verification algorithms in the context of the artificial
pancreas. The artificial pancreas refers to a series of increasingly
sophisticated closed loop medical devices that automate the delivery
of insulin to people with type-1 diabetes. On one hand, they hold
the promise of easing the everyday burden of managing type-1
diabetes. On the other, they expose the patient to potentially
deadly consequences of incorrect insulin delivery that could lead to
coma or even death in the short term, or damage to critical organs
such as the eyes, kidneys and the heart in the longer
term. Verifying the correctness of these devices involves a careful
modeling of human physiology, the medical device, and the surrounding
disturbances at the right level of abstraction. We first provide a
brief overview of insulin glucose regulation and the spectrum of
associated mathematical models. At one end are physiological models
that try to capture the transport, metabolism, uptake and
interactions of insulin and glucose. On the end are data driven
models which include time series models and neural networks. The
first part of the chapter examines some of these models in detail in
order to provide a basis for verifying medical devices. Next, we
present some of the devices which are commonly used in blood glucose
control, followed by a specification of key correctness properties
and performance measures. Finally, we examine the application of
some of the state-of-the-art approaches to verification and
falsification of these properties to the models and devices
considered. We conclude with a brief presentation on future
directions for next generation artificial pancreas, and the
challenges involved in reasoning about them.
Factory-Calibrated Continuous Glucose Monitoring: How and Why It Works, and the Dangers of Reuse Beyond Approved Duration of Wear
Gregory Forlenza,
Taisa Kushner,
Laurel Messer,
R. Paul Wadwa,
and Sriram Sankaranarayanan
Diabetes: Technology and Therapeutics,
21
(4)
pp. 222-229,
2019.
Continuous glucose monitors (CGM) display real-time glucose values enabling greater
glycemic awareness with reduced management burden. Factory-calibrated CGM systems
allow for glycemic assessment without the pain and inconvenience of fingerstick
glucose testing. Advances in sensor chemistry and CGM algorithms have enabled
factory-calibrated systems to have greater accuracy than previous generations of
CGM technology. Despite these advances many patients and providers are hesitant
about the idea of removing fingerstick testing from their diabetes care. In this
commentary, we aim to review the clinical trials on factory-calibrated CGM systems,
present the algorithms which facilitate factory-calibrated CGMs to improve accuracy,
discuss clinical use of factory-calibrated CGMs, and finally present two cases
demonstrating the dangers of utilizing exploits in commercial systems to
prolong sensor life.
Robust Data-Driven Control of Artificial Pancreas Systems using Neural Networks
Souradeep Dutta,
Taisa Kushner,
and Sriram Sankaranarayanan
In Computational Methods in Systems Biology,
Lecture Notes In Computer Science
Vol. 11095,
pp. 183-202,
2018.
In this paper, we provide an approach to data-driven
control for artificial pancreas system by learning
neural network models of human insulin-glucose
physiology from available patient data and using a
mixed integer optimization approach to control blood
glucose levels in real-time using the inferred
models. First, our approach learns neural networks
to predict the future blood glucose values from
given data on insulin infusion and their resulting
effects on blood glucose levels. However, to provide
guarantees on the resulting model, we use quantile
regression to fit multiple neural networks that
predict upper and lower quantiles of the future
blood glucose levels, in addition to the mean.
Using the inferred set of neural networks, we
formulate a model-predictive control scheme that
adjusts both basal and bolus insulin delivery to
ensure that the risk of harmful hypoglycemia and
hyperglycemia are bounded using the quantile models
while the mean prediction stays as close as possible
to the desired target. We discuss how this scheme
can handle disturbances from large unannounced meals
as well as infeasibilities that result from
situations where the uncertainties in future glucose
predictions are too high. We experimentally
evaluate this approach on data obtained from a set
of 17 patients over a course of 40 nights per
patient. Furthermore, we also test our approach
using neural networks obtained from virtual patient
models available through the UVA-Padova simulator
for type-1 diabetes.
A Data-Driven Approach to Artificial Pancreas Verification and Synthesis.
Taisa Kushner,
David Bortz,
David Maahs,
and Sriram Sankaranarayanan
In Intl. Conference on Cyber-Physical Systems (ICCPS’18),
pp. 242-252,
2018.
This paper presents a case study of a data driven approach to
verification and parameter synthesis for artificial pancreas control
systems which deliver insulin to patients with type-1 diabetes
(T1D). We present a new approach to tuning parameters using
non-deterministic data-driven models for human insulin-glucose
regulation, which are inferred from patient data using multiple time
scales. Taking these equations as constraints, we model the behavior
of the entire closed loop system over a five-hour time horizon cast
as an optimization problem. Next, we demonstrate this approach using
patient data gathered from a previously conducted outpatient
clinical study involving insulin and glucose data collected from
50 patients with T1D and 40 nights per patient. We use the
resulting data-driven models to predict how the patients would
perform under a PID-based closed loop system which forms the basis
for the first commercially available hybrid closed loop
device. Futhermore, we provide a re-tuning methodology which can
potentially improve control for 82% of patients, based on the
results of an exhaustive reachability analysis. Our results
demonstrate that simple nondeterministic models allow us to
efficiently tune key controller parameters, thus paving the way for
interesting clinical translational applications.
Compositional Relational Abstraction for Nonlinear Systems
Xin Chen,
Sergio Mover,
and Sriram Sankaranarayanan
ACM Transactions on Embedded Computing Systems (Special Issue for EMSOFT 2017),
16(5s)
pp. 187,
2017.
We propose techniques to construct abstractions for nonlinear
dynamics in terms of relations expressed in linear arithmetic. Such
relations are useful for translating the closed loop verification
problem of control software with continuous-time, nonlinear plant
models into discrete and linear models that can be handled by
efficient software verification approaches for discrete-time
systems. We construct relations using Taylor model based flowpipe
construction and the systematic composition of relational
abstractions for smaller components. We focus on developing
efficient schemes for the special case of composing abstractions for
linear and nonlinear components. We implement our ideas using a
relational abstraction system, using the resulting abstraction
inside the verification tool NuXMV, which implements numerous
SAT/SMT solver-based verification techniques for discrete
systems. Finally, we evaluate the application of relational
abstractions for verifying properties of time triggered controllers,
comparing with the Flow* tool. We conclude that relational
abstractions are a promising approach towards nonlinear hybrid
system verification, capable of proving properties that are beyond
the reach of tools such as Flow*. At the same time, we highlight
the need for improvements to existing linear arithmetic SAT/SMT
solvers to better support reasoning with large relational
abstractions.
Formal Verification of a Multi-Basal Insulin Infusion Control Model.
Xin Chen,
Souradeep Dutta,
and Sriram Sankaranarayanan
In Workshop on Applied Verification of Hybrid Systems (ARCH),
pp. 16,
2017.
The artificial pancreas concept automates the delivery of
insulin to patients with type-1 diabetes, sensing the blood glucose
levels through a continuous glucose monitor (CGM) and using an
insulin infusion pump to deliver insulin. Formally verifying control
algorithms against physiological models of the patient is an
important challenge. In this paper, we present a case study of a
simple hybrid multi-basal control system that switches to different
preset insulin delivery rates over various ranges of blood glucose
levels. We use the Dalla-Man model for modeling the physiology of
the patient and a hybrid automaton model of the controller. First,
we reduce the problem state space and replace nonpolynomial terms by
approximations with very small errors in order to simplify the
model. Nevertheless, the model still remains nonlinear with up to 9
state variables.
Reachability analysis on this hybrid model is used to verify that
the blood glucose levels remain within a safe range overnight. This
poses challenges, including (a) the model exhibits many discrete
jumps in a relatively small time interval, and (b) the entire time
horizon corresponding to a full night is 720 minutes, wherein the
controller time period is 5 minutes. To overcome these difficulties,
we propose methods to effectively handle time-triggered jumps and
merge flowpipes over the same time interval. The evaluation shows
that the performance can be improved with the new techniques.
Analyzing Neighborhoods of Falsifying Traces in Cyber-Physical Systems
Ram Das Diwakaran,
Sriram Sankaranarayanan,
and Ashutosh Trivedi
In Intl. Conference on Cyber-Physical Systems (ICCPS),
pp. 109-119,
2017.
We study the problem of analyzing falsifying traces of
cyber-physical systems.
Specifically, given a system model and an input which is a counterexample
to a property of interest, we wish to understand which parts of the inputs
are “responsible” for the counterexample as a whole.
Whereas this problem is well known to be hard to solve precisely, we provide an
approach based on learning from repeated simulations of the system under test.
Our approach generalizes the classic concept of “one-at-a-time” sensitivity
analysis used in the risk and decision analysis community to understand how
inputs to a system influence a property in question.
Specifically, we pose the problem as one of finding a neighborhood of inputs
that contains the falsifying counterexample in question, such that each point
in this neighborhood corresponds to a falsifying input with a high probability.
We use ideas from statistical hypothesis testing to infer and validate such
neighborhoods from repeated simulations of the system under test.
This approach not only helps to understand the sensitivity of
these counterexamples to various parts of the inputs, but also
generalizes or widens the given counterexample by returning a
neighborhood of counterexamples around it.
We demonstrate our approach on a series of increasingly complex examples
from automotive and closed loop medical device domains.
We also compare our
approach against related techniques based on regression and machine learning.
Model-Based Falsification of an Artificial Pancreas Control System.
Sriram Sankaranarayanan,
Suhas Akshar Kumar,
Faye Cameron,
B. Wayne Bequette,
Georgios Fainekos,
and David M. Maahs
ACM SIGBED Review (Special Issue on Medical Cyber Physical Systems) ,
14
(2)
pp. 24-33,
2016.
We present a model-based falsification scheme for artificial
pancreas controllers. Our approach performs a closed-loop simulation
of the control software using models of the human insulin-glucose
regulatory system. Our work focuses on testing properties of an
overnight control system for hypoglycemia/hyperglycemia minimization
in patients with type-1 diabetes. This control system is currently
the subject of extensive phase II clinical trials.
We describe how the overall closed loop simulator is constructed, and
formulate properties to be tested. Significantly, the closed loop
simulation incorporates the control software, as is, without any
abstractions. Next, we demonstrate the use of a simulation-based
falsification approach to find potential property violations in the
resulting control system. We formulate a series of properties about
the controller behavior and examine the violations obtained. Using
these violations, we propose modifications to the controller software
to improve its performance under these adverse (corner-case)
scenarios. We also illustrate the effectiveness of robustness as a
metric for identifying interesting property violations. Finally, we
identify important open problems for future work.
Refining the Closed Loop in the Data Age: Research-to-Practice Transitions in Diabetes Technology (Editorial)
Gregory P. Forlenza,
Sriram Sankaranarayanan,
and David M. Maahs
Diabetes Technology and Therapeutics,
17
(5)
pp. 304-306,
2015.
The editorial analyzes a recent clinical study reporting on
the impact of insulin pump shutoff on patients with
type-1 diabetes. In a broader context, it examines
the need for enhanced formal verification techniques
applied to more sophisticated artificial pancreas
controllers that remain under development.
Towards a Verified Artificial Pancreas: Challenges and Solutions for Runtime Verification
Fraser Cameron,
Georgios Fainekos,
David M. Maahs,
and Sriram Sankaranarayanan
In Proceedings of Runtime Verification (RV’15),
Lecture Notes in Computer Science
Vol. 9333,
pp. 3-17,
2015.
In this paper, we briefly examine the recent developments in artificial pancreas controllers, that automate the delivery of insulin to patients with type-1 diabetes. We argue the need for offline and online runtime verification for these devices, and discuss challenges that make verification hard. Next, we examine a promising simulation-based falsification approach based on robustness semantics of temporal logics. These ideas are implemented in the tool S-Taliro that automatically searches for violations of metric temporal logic (MTL) requirements for Simulink(tm)/Stateflow(tm) models. We illustrate the use of S-Taliro for finding interesting property violations in a PID-based hybrid closed loop control system.
Determination of personalized diabetes treatment plans using a two-delay model
Stephen M. Kissler,
Cody Cichowitz,
Sriram Sankaranarayanan,
and David M. Bortz
J. Theoretical Biology,
359
(Oct)
pp. 101-111,
2014.
Diabetes cases worldwide have risen steadily over the past decades, lending urgency to the search for more efficient, effective, and personalized ways to treat the disease. Current treatment strategies, however, may fail to maintain ultradian oscillations in blood glucose concentration, an important element of a healthy alimentary system. Building upon recent successes in mathematical modeling of the human glucose-insulin system, we show that both food intake and insulin therapy likely demand increasingly precise control over insulin sensitivity if oscillations at a healthy average glucose concentration are to be maintained. We then suggest guidelines and personalized treatment options for diabetic patients that maintain these oscillations. We show that for a type II diabetic, both blood glucose levels can be controlled and healthy oscillations maintained when the patient gets an hour of daily exercise and is placed on a combination of Metformin and sulfonylurea drugs. We note that insulin therapy and an additional hour of exercise will reduce the patient’s need for sulfonylureas. Results of a modeling analysis suggest that a typical type I diabetic’s blood glucose levels can be properly controlled with a constant insulin infusion between 0.45 and 0.7 microU/(ml*min). Lastly, we note that all suggested strategies rely on existing clinical techniques and established treatment measures, and so could potentially be of immediate use in the design of an artificial pancreas.
Other Resources
The Flow* tool and S-Taliro tool have been used so far to verify the artificial pancreas control algorithms.
This page was authored and maintained by Sriram Sankaranarayanan, who is responsible for its content. Please direct all queries to Sriram at his email address.