Content uploaded by Smita Dikondkar
Author content
All content in this area was uploaded by Smita Dikondkar on Mar 29, 2017
Content may be subject to copyright.
Design and Development of Insulin
Delivery System Prototype
Prof. Smita R. Dikondwar
Department of Instrumentation and Control,
College of Engineering Pune, Shivaji Nagar
Pune , India.
E-mail:shmit_dikondwar@yahoo.com
Abstract— The manual method to insulin delivery for diabetes
mellitus require significant efforts from clinical stand point
and not guarantee of an optimal performance. The insulin
delivery system is a low cost portable, automated insulin
delivery pump which is used by diabetics to administer insulin
as and when they require it at regular intervals. An insulin
delivery system for diabetes patients, which is used for the
continuous subcutaneous insulin infusion (CSII) therapy. The
purpose of design & development of insulin delivery system is
to deliver precise and accurate insulin infusion rate
continuously to the patient using continuous glucose
monitoring system (CGMS) data & providing cost effective
solution along with flexible lifestyle.
The ergonomic design is developed considering Type 1
and Type 2 diabetes patient model. The proposed control
strategy is evaluated at different CGMS values, which
correspond to realistic meal profiles and workouts performed
by patient. The system is a safety-critical embedded system
which estimates and control the infusion rate of insulin based
upon relative proportional control law. Insulin delivery system
is an indigenous concept with highly reliable control of micro-
fluidic delivery.
Infusion delivery system is used to deliver extremely small
volume of liquid in predefined time duration, at a constant
flow rate, therefore the system find applications in research
and development related to biotechnology, bioengineering,
chemical laboratories and in analytical instruments.
Keywords— Embedded system control; Glucose meter; Insulin
pump and reservoir; Mini stepper motor; Insulin rate
regulation; Cannula and infusion set.
I. INTRODUCTION
Insulin-dependent diabetes mellitus is a chronic metabolic
disorder, which is characterized by the disability of the body
to maintain the blood glucose levels into physiological
ranges. Particularly, it is an autoimmune disease in which
the beta- cells of the pancreas are destroyed, resulting in the
absence of insulin secretion. Improperly managed diabetes
can lead to complications such as nerve damage, brain
damage, heart disease, stroke, vision loss, amputation, and
kidney disease and eventually death. Diabetes related
complications are a worldwide epidemic with high medical,
economic and social costs. Tight control of blood glucose
levels has also been shown to reduce the mortality of
diabetic, and non-diabetic, intensive care unit patients by up
to 50%. Diabetic individuals monitor food intake and daily
activity to maintain blood sugar levels at an adequate level.
Far ease of management subjects are encouraged to stick to
strict routines and diets and injections. However, this regime
can lead to severe limitation of maintaining a strict daily
regimen over several years. A typical day for a diabetic
individual might involve injecting long-acting insulin
approximately four times and injecting rapid acting insulin
before meals[28], to reduce the post- prandial blood glucose
spike. Type 1 and 2 diabetes is observed in all age groups
[26]. Insulin delivery system (CSII) can also have a
significant impact on the treatment of juveniles. With
accuracy and enabling more effective treatment.
However to cope up contemporary lifestyle with
diabetes mellitus, Insulin delivery system is best suited and
preferred. Developing country like India has over 40
millions of diabetic patients and 10-10.2% are type1
amongst the diabetics. Considering the cost of healthcare,
the type1 is very expensive disease. According to the
diabetes control and complications trial and the U.K.
prospective diabetes study (UKPDS), an intensive
glucaemic control has the ability to reduce the
aforementioned [23] complications. Advances in technology
have led to the development of continuous glucose
monitoring systems (CGMS) and insulin delivery pump
therapy. Consequently, the knowledge based analysis
implementation of an ―artificial electromechanical
pancreas‖[9], which incorporates a CGMS, an insulin pump
and the appropriate computational data for the automatic
adjustment of insulin infusion rates is[23][24] feasible.
The paper presents the design and development of low cost
insulin delivery system based upon a semi-close loop
control which makes use of knowledge base data (CGMS)
for the simulation of glucose-insulin metabolism of diabetes
patients. The system controls infusion rate of insulin based
upon blood glucose levels In this, the work is aimed at
accuracies of insulin rate and cost effectiveness [17][21] and
patient flexibility with the use of CSII.
II. GLUCOSE METABOLIC PROCESS
Insulin, a hormone secreted by the pancreatic beta cells of
the liver,is critical in assisting body cells to assimilate the
glucose present in the blood. In normal conditions, the liver
continuously monitors the current blood glucose level and
determines the amount of insulin to be secreted to maintain
tight physiological constraints, known as
normoglycemia(e.g.Typical Blood glucose range 60-
110mg/dl).[13] Type-I diabetes is chronic disease, the body
cannot utilize the glucose in the blood due to the precisely
destruction of the beta pancreatic cell of the liver and the
resulting inability of the pancreas to secrete insulin. If
untreated, this would result in large perturbations of the
plasma glucose level and lead to hyper- or hypoglycemia,
respectively), which in turn may cause severe damage to the
eyes, kidneys, nerves, heart and blood vessels of the
patients. Currently, the treatment of Type-I diabetes
involves delivering insulin externally to regulate the blood
glucose level to maintain normoglycemia, refer Fig 1.
Fig. 1 Glucose concentration and insulin secretion profile
for the day of a normal person.
Insulin can be delivered through two different regimes..The
first one is known as the conventional insulin regime
consists of oral insulin intake and inject of slow-acting
insulin at pre-determined times (typically,4 times daily)The
second regime is the more flexible insulin pump therapy
where patients can monitor their blood glucose periodically
and input the proper insulin amount as needed. The
conventional insulin regime is far from ideal for the
treatment of Type-I diabetes because the regulation [38]of
the insulin is basically an open-loop process. In other words,
patients need to determine the amount of insulin to be taken
based solely on their personalized prescription and an
estimate of the carbohydrate content of the food they take. It
provides real-time monitoring of blood glucose level
(CGM) and allows patients to control the insulin pump to
deliver [14]insulin, the effect of which can be seen almost
immediately. A typical basal and bolus insulin-glucose
profile shown in Fig 1. Invariably insulin reacts over these
glucose levels to maintain it as in typical [1]glucose values
i.e.60-110mg/dl.Fig 2(a)(b).better illustrates this balance of
the glucose levels by using insulin delivery system.
Fig. 2(a) The glucose levels shooting up and down before insulin delivery
Fig. 2(b) The controlled glucose levels after infusions by CSII system.
III. INSULIN DELIVERY SYSTEM STRUCTURE
The insulin delivery system mainly consists of four major
components: Mini-stepper motor with linear actuator and an
embedded controller system, reciprocating piston pump and
infusion set , an invasive glucose meter, refer Fig .3.
Fig 3 Block Schematic of Insulin Delivery (CSII) System
The proposed insulin delivery system is about the size of a
pager, comprises a reservoir of insulin and embedded
system control for insulin to be delivered. The pump
infusion rate is typically in the range of 0.0209U/min-
50U/min. The insulin is given at predetermined
times(Typically,4 times daily) [25]via a cannula,soft tube
that rests into subcutaneous tissue. Cannula can be used up
to 72 hours after insertion, which is done painlessly using an
automatic inserter.
A. Methodology
In the development of the design the stages followed with
some basic constraints [10]considered significantly while
selection of hardware and software of the system. Such as
glucose sensor deployment, [7] ergonomic design aspects,
portability and power management.
1) Blood glucose monitoring unit selection: The survey of
Type I and TypeII [20][23][25] patient model renders to
the selection of ‗Accuchek Active‘ model by Roche
diagnostics. A reliably used glucose meter by the
patients and doctors as, It shows the better accuracy and
performance for CGMS.
2) Selection of Insulin Piston Syringe pump: The infusion
pump with disposable reciprocating piston and barrel or
reservoir with volume of 300units(3ml) is considered for
72 hours after survey. The pump material is checked for
its biocompatibility with insulin.
3) Selection of Mini Stepper Motor with Linear Actuator:A
bipolar stepper motor [33] with captive linear actuator,
linear travel when extended - 13.84mm.Linear travel
0.013mm per step with 0.001mm resolution. Plastic nut
is fitted at the ip of actuator of size- (lxb) 3mm x10mm.
4) Selection of Embedded System Components: The entire
system is a battery operated due to the life and
portability aspects. The components of entire system are
selected considering [19][24] the system constraints.
The controller with suitable current and voltage ratings
and data-program memory with peripheral compatibility
[15] is considered.
Thus low cost design with expandable features
Microcontroller-AVR ATmega series is suitable for
proposed embedded control.High-performance with up
to 16 MIPS[27] throughput at 16 MHz, microcontroller
,Lcd and keypad with indicators ,alarms, and alkaline
battery supply, refer Fig.5.
5) Insulin Delivery System Architecture: The desirable
system must be robust and respond accurately to the
variability of the patient‘s metabolic profile. Automated
insulin delivery therapy‘s ultimate success: A linear
actuator that would compress the syringe to inject the
insulin. Essentially, [6][10] automation of delivery
pump replaces the human element – the finger that
would push down on the syringe‘s plunger with a
actuator that would push down on the plunger which
plays an integral role with precise controls [2].refer
Fig.4. System architecture with hardware components
such as mini motor-linear actuator, syringe pump, Lcd,
keypad, and enclosure.
Fig.4 Insulin delivery system architecture
Fig.5 Block diag.of embedded controller for insulin delivery system
IV. PRINCIPLE OF WORKING
There are many complex influences between glucose and
insulin concentration for any person, normal or diabetic.
However, the steady state glucose concentration[20][23] in
the body is finally determined by how much insulin is
present. In order to lower glucose concentration in the
blood, insulin needs to be infused. Hence, the controller
defines the [16] insulin infusion rate u(t), based on the
measured BG. The goal is to compensate excess BG, and
infuse precise insulin doses. Referring to CGMS data,
relative proportional control and sliding scale controls are
the control strategies to be followed.
One of the better-known, more effective diabetes control
systems is a Relative proportional control law(RPC). This is
widely used form of law is based on the idea of strictly
limiting the absolute blood sugar level by applying resisting
insulin in weighted proportion to the magnitude of the
excursion from the desired (basal) level.
The control is based on a relative proportional control
(G/Gь) with a constant term note that when G is blood sugar
at the desired level and the insulin infusion rate is u(t) the
basal infusion rate necessary to maintain blood glucose at a
constant level.
Stepper motor control is decided by controller (RPC) as per
insulin entered by user
[ u(t)= uơ(1+G/Gь)]
The daily dose ,patient can obtain it from the equation as
follows,
The Daily Dose :
Q(U) = [A(mg/dl)-100]x10xB(Kg)x0.6÷1000÷2]
Where, A=BG, B=Weight of patient
However RPC law need the BG values and total insulin
dose.Therefore rate of motor-actuator steps per minute can
be decided by Sliding scale control(SSC), which is based on
case analysis by expert and patient operations. Thus
controller delivers rate as it is entered by patient.
In case of RPC law controller need daily dose and current
BG as inputs and controller decides what may the rate of
delivery.
B. Other Calculations
1) The Volume of insulin syringe (VI)- to be
delivered decides total linear displacement of
piston VI=0.013mm. Linear travel by motor /step
2) Insulin piston displacement per step decides the
insulin delivery rate as per the range of the system:
Delivery /Step = 0.2µl/min to 200µl/min
(0.029U/min-20U/min)
3) Time of delivery, t(min)- Calculated by controller
as per entered insulin rate and so the actuator speed
(Sp) is adjusted.
4) The brown out detection, as the vdd is less than
2.25V it internally enable the reset that shows
battery is low and the power save mode is enabled.
V. HARDWARE AND SOFTWARE IMPLEMENTATION
The installation of Stepper motor - Syringe pump and
infusion set, LCD, keypad and peripheral component
interfaced with programmed microcontroller along with
battery forms complete hardware implemented insulin
delivery system. Refer Fig.6 shows setup for Micro-liquid
delivery tests [2]performed with microcontroller output for
liquid delivery per step with single stepping mode of
motor.[16]. Refer Fig. 7.
The Insulin Delivery system logic algorithm is
implemented and executed via BASCOM and AVR Studio4
software using SPI interfacing[27] and verified with system
hardware.
Fig.6 Setup for micro-liquid delivery tests
Fig. 7 System in troubleshoot
VI. OBSERVATIONS AND RESULTS
The tests are performed for liquid delivery for various flow-
rates, the delivered liquid is measured by micro-liter syringe
for various number of cycles the observations are noted.
Thus, the test results are:
-The minimum value of delivered insulin observed in
tests is 0.209µl = 0.0209Units
-The Resolution is 0.025µl = 0.0025Units
- Accuracy ±0.01%
-The Repeatability of the test is very good for (alternate
or continuous) step cycles with different time delays.
-The system has wide span of operation which is 0.209
µl- 209µl(0.0209U-20.9U), Hence comparatively this
design offers very high span.
-During experimentation of liquid delivery tests: The
Linear Actuator of mini stepper motor performance is
exemplarily with no vibrations and noise due to rotor
movement is observed. It shows smooth operation for
entire tests.
Fig. 8 Insulin Delivery System
VII. CONCLUSION AND FUTURE SCOPE
A. Conclusion
The Insulin delivery system is indigenous, low cost, precise
(Nanoliter & Microliter) -liquid delivery design prototype.
The insulin delivery system accuracy is about ± 0.01 %
accuracy, resolution- 0.025µl(0.0025Unit) and good
repeatability. The system has large span of operation, The
range of system is 0.2µl/min to 200µl/min (0.02U/min-
20U/min), The system gives five times fine liquid delivery
where the market device offers a range of 1µl/hr to 250µl/hr
(0.1U/hr- 25U/hr). The system can use knowledge base data
(CGMS) as assistance for insulin dose decisions.
B. Future Scope
Developed system can be deployed with biosensor for
advanced control strategies like Model based predictive
control to improve accuracy and complete automation of
system. Telemetry can be the best suited for wireless
advanced control of delivery system.
This system can also be used in biomedical,
bioengineering, chemical laboratory and analytical
instruments for several research and applications.
REFERENCES
[1] William Goh, Michel Pasuqier,“Adaptive control of infusion pump
for type I diabetes control using a Self tuning regulator” 10th
Intl.Conf. on control Automation,Robotics & Vision
Hanoi,Vietnam,17-20 Dec 2008.
[2] D. Henderson. Presented at NSTI ,Nanotech2007.―Novel piezo motor
enables positive displacement microfluidic pump‖.
[3] Zhi Xu,Sheng Liu,Zlyin Gan,Bin Ma,Guojun Liu,Xinxia ghai
Zhang,―An integrated intelligent insulin pump‖ 2006 IEEE.7th
International Conf. on Electronics Packaging Technology.
[4] Ruoting Yang¹, Mingjun Zhang², and Tyzh-Jong Tarn ―Dynamic
modeling and control of a micro-needle integrated piezoelectric
micro-pump for diabetes care‖ 2006 IEEE.
[5] R.S. Parker, F.J. Doyle 111.“Model predictive control for infusion
pump insulin delivery” 18th Annual International IEEE Conf. on
Engineering in Medicine & Biology Society, 1996.
[6] R. Dudde, ltzehoe &T. Vering ,Disetronic Medical Systems,
“Advanced insulin infusion using a control loop concept and
realization of a control-loop application for automated delivery of
insulin” June16, 2009,IEEE.
[7] Harald Kirchsteiger, Luigi del Re, Eric Renard and Margot
Mayrhofer American Control Conf., ―Robustness properties of
optimal insulin bolus administrations for type I diabetes‖ June10-12,
2009
[8] Pasquale Palumboz Pierdomenico Simona Panunzi Andrea De
Gaetano, 47th IEEE Conf.on Decision & Control,“Robust closed-
loop control of plasma glycemia: discrete-delay model
approach”Dec. 2008.
[9] Enrique J.G´omez,M.Elena Hernando P´erez,Thomas,“The INCA
system: a further step towards telemedical artificial pancreas”IEEE
Transactions on Information Technology in Biomedicine,VOL.12,
2008.
[10] Insulin Pump website,for “Insulin pump system design”.Online
available at www.lancs.ac.uk/ug/mccarthb/insulinpump.zip.
[11] O.Vega-Hern´andez1, D.U.Campos-Delgado2, Espinoza-Trejo3
.“Actuator fault tolerant control for an artificial pancreas” IEEE
2005.
[12] Farid Amirouche,E Yu Zhou Tom Johnson.―Current micropump
technologies and their biomedical applications‖ Springer-Verlag,Feb
2009
[13] Paul Martin, Member, IEEE, Saul Genuth.“Normalization of insulin
delivery to diabetics by pulsed insulin infusion” IEEE Transactions
on Biomedical Engineering, Vol.2 BME-24,Mar 1977.
[14] Md.Shafiqul Islam, James Leech, Charles C.Y.Lin and Chrostowski
“Peak blood glucose prediction algorithm for a meal intake” IEEE
2007.
[15] W. J. SPENCER “A review of programmed insulin delivery systems”
IEEE Transactions on Biomedical Engineering, Vol.3.BME-28, Mar.
1981.
[16] J. Geoffrey Chase,Z-H Lam,K-S Hwang, 7th Int. Conf. on Control,
“Active insulin infusion control of blood glucose derivative” Dec
2002.
[17] Leah Tolosa, Ignacy Gryczynski, Lisa Eichhorn, Felix Castellano,
Govind Rao,―Glucose sensor for low cost lifetime-based sensing
using genetically engineered protein‖ Idealibrary, 2-17July1998.
[18] Louis Shenkman, Marina Koukaki, Spyridoula Karamanou,
“Innovative approaches continuous glucose monitoring”.Proceedings
of the 29th International Conf. IEEE, EMBS, Aug23-26, 2007.
[19] Bipin Nair Harish Kumar K. Guruvayurappan. “Dual
Microcontroller-Based Liquid System”.
[20] Saeedeh Lotfi Md. Abad ,Keivan Maghooli.“Different approaches
for detecting glucose”, IEEE 2008.
[21] Cynthia M. Furse.―Biomedical telemetry: today‘s opportunities and
challenges‖, IEEE 2009.
[22] “Deltec cozmo pumps” by Kerk motion Products, Inc". Application
Case study No.12, WWW.KERKMOTION.COM
[23] (2006) GlucoSim website .Online available on http://216.47.139.198/
glucosim(retr.may2006) and Insulin sliding scale. mht-link
[24] International Electrotechnical Commission,Safety functional safety
related systems –part3,Edi.1.0,1998.
[25] Medtronic Website.“Real-time continuous glucose monitoring
system”,2008 Online available at http://www.medtronic.com(retr.may
[26] Medical guidance for diabetes model and practices: Consultant in
Endocrinology, Pune & TypeI patient model.
[27] Atmel Corporation, Atmel®.2009, ―2466S–AVR–05/09 datasheet,‖
& ―Oct 199, L293E Datasheet,‖ 1994 SGS-thomson
Microelectronics.
[28] S. D. Teddy, C. Quek, E. M.-K. Lai, and A. Cinar, “PSECMAC
intelligent insulin schedule for diabetic blood glucose management
under nonmeal announcement” IEEE Transactions on Neural
Networks, Vol. 21, NO. 3, Mar.2010.
