Detection of failure modes in therapeutic infusion sets

Abstract

An injection device for detecting failure modes including at least one of occlusions, air bubbles, leaks, and empty reservoir, based on analysis of an electric signal. The injection device includes a peristaltic fluidic pump including at least one chamber separated by a flexible membrane into a fluid compartment and an air compartment, at least one thermoelectric air compressor configured to compress air in the air compartment to deflect the flexible membrane and displace fluid in the fluid compartment toward its destination, an electric power supply configured to power the at least one thermoelectric air compressor, and means of reading and interpreting electric signals from an electric circuit including the air compressor.

Description

[0001] Detection of Failure Modes in Therapeutic Infusion Sets

[0002] Specification

[0003] Cross references to related applications

[0004] This application claims the benefit of priority under 35 U.S.C. §119(e) from copending, commonly owned U.S. provisional patent applications and PCT applications U.S. Ser. No. 63 / 662,031 filed June 20, 2024 and PCT / US25 / 34515, filed June 20, 2025, titled “Microfluidic Peristaltic Pump Assembly”, and U.S. Ser. No. 63 / 662,052 filed June 20, 2024 and PCT / US25 / 34525 filed June 20, 2025 titled “Bi-directional Thermoelectric Gas Compressor Amenable to Nanofabrication.” The entire content of these applications is incorporated herein by reference.

[0005] Technical Field

[0006] The present disclosure relates to the medical field of fluidic pumps and infusion sets which deliver therapeutics, such as insulin, to individuals through an injection to their subcutaneous tissue or other tissue, and to methods of detecting failure modes, such as occlusions, air bubbles, and leaks in a fluidic path between a storage reservoir of a therapeutic substance and the tissue of an individual. Background

[0007] Continuous subcutaneous insulin infusion sets (IIS) equipped with an external pump were introduced to individuals with type 1 diabetes (T1 D) in the 1970s, and tremendous progress has been made since that time in terms of device usability and safety. Recent advances include development of relatively compact patch pumps which, combined with new continuous glucose monitoring (CGM) systems, have been increasingly used to improve glycemic control by individuals with T1 D as well as insulinrequiring type 2 diabetes. The progress in technology was immediately followed by an increase in a number of individuals using the devices. When used properly, a treatment is considered to be safe and effective (see G. Zhang, 0. Cohen, and S. Chattaraj, Development of the Extended Infusion Set and Its Mechanism of Action, Journal of Diabetes Science and Technology, 2022, doi 10.1177 / 19322968221112120, and references therein). There are, however, major challenges encountered in the development of infusion sets, with many of them related to various modes of failure of IIS which can lead to hyperglycemia, diabetic ketoacidosis, and, in most troubling cases, to hypoglycemia. Among these health- and sometimes life-threatening failures, insulin occlusions, leaks, and air bubbles are the most frequent.

[0008] Insulin pump occlusions, i.e. blockages in the insulin path on its way from an insulin cartridge (or storage reservoir) to the subcutaneous tissue, imply that the insulin pump cannot maintain a set flow rate either because of a total or partial occlusion. If this happens, the pressure inside the insulin cartridge will increase, inducing an occlusion alarm (which very often fails to do so due to high thresholds set to prevent false positives related to a plunger sticking at the insulin cartridge surface), the flow rate of insulin delivery will decrease from the set flow rate down to zero, and if the occlusion is not relieved, then blood glucose levels will rise (D.C. Klonoff, G. Freckmann, and L. Heinemann, Insulin Pump Occlusions: For Patients Who Have Been Around the (Infusion) Block, Journal of Diabetes Science and Technology 2017, Vol. 11 (3) 451-454, doi 10.1177 / 1932296817700). Two main risks induced by a pump occlusion are hyperglycemia due to failure of insulin delivery, and hypoglycemia when an insulin infusion rate is increased to overcome the occlusion, and a large volume of insulin that was not infused during the occlusion is suddenly delivered.

[0009] Another reason for IIS failure is air bubbles, which for the most part are formed in the insulin storage reservoir, and may lead to hyperglycemia, especially in children or patients requiring low infusion rates (L. Heinemann, Air Bubbles in Insulin Pumps: A Clinically Relevant Issue?, Journal of Diabetes Science and Technology 2022, Vol. 16(6) 1351-1355, doi doi.org / 10.1177 / 1932296822110188). There are three sources of air bubbles in IIS. The first source is related to filling an insulin infusion line after changing the insulin storage reservoir in the pump, or directly filling the storage reservoir. The second source is when cold insulin is taken from a refrigerator and placed into the pump. Warming of the liquid in the pump causes outgassing and the formation of air bubbles; an amount of outgassing is a direct function of a temperature difference between the insulin storage (e.g., in the refrigerator) and the ambient atmosphere. Insulin manufacturers train customers that using cold insulin causes air bubbles in the storage reservoir and tubing, but recommendations of the insulin manufacturers are not always followed. The third, often related to the second, and clearly most unpredictable, source of air bubbles is due to temperature differences between the insulin in the cartridge / storage reservoir and the external world during pump usage, which exposes the insulin to changes in ambient temperature over time, and depending on wear conditions, such as proximity to skin / body or within a belt clip, might influence the degree of the outgassing.

[0010] Insulin leaks are yet another problem that occurs relatively often in IIS. Those can be insulin reservoir leaks, which happen if for example the rubber seals around the plunger become weak or damaged (https: / / www.diabetes.co.uk / insulin-pumps / insulin- pump-problems.html), or cannula leaks, which may be caused by infusion sets coming loose when sweat is able to gather at an infusion site, or the rate of delivery exceeds a maximum rate that may be dissipated by the tissue at a delivery time. In any case, insulin leaks may result in hyperglycemia and diabetic ketoacidosis (DKA) due to the insulin being underdelivered.

[0011] In summary, IIS failures due to occlusions, air bubbles, and leaks, is a problem well recognized in the industry, which at this time has no satisfactory solutions, currently being mitigated through the education of individual users.

[0012] Summary

[0013] Before explaining at least one exemplary embodiment of the present disclosure in detail, it is to be understood that the present disclosure is not limited in its application to the details of construction and to the arrangements of the components set forth in the following description or illustrated in the drawings. The disclosed device and method are capable of other embodiments and of being practiced and carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein are for the purpose of the description and should not be regarded as limiting. The primary objective of the disclosed technology is timely, accurate, and reliable detection of failure modes, such as occlusions, air bubbles, and leaks, occurring in infusion sets and similar medical devices which deliver therapeutic medicines, such as insulins, by injecting them into subcutaneous tissue or other tissues of an individual, i.e. , therapeutic injection devices. In the disclosed embodiment, the detection is achieved through the recording and analysis of an electric potential and / or electric current in a thermoelectric air nanocompressor (US Provisional Application Ser. No. 63 / 662,052) which deflects the membrane of a peristaltic manifold assembly (US Provisional Application Ser. No. 63 / 662,031).

[0014] The operating sequence of the detection is as follows. First, the electric current is supplied through the thermoelectric nanocompressor. This can be done either by maintaining a constant bias of the electric potential (voltage) across the nanocompressor, or by adjusting the voltage to support a constant electric current. When the electric current passes through the thermoelectric elements of the nanocompressor, they become heated at one end due to the Peltier effect. The temperature gradients across thermoelectric elements result in temperature gradients in the gas / air adjacent to the thermoelectric elements. Each of the thermoelectric elements has an array of apertures built in them. Because an aperture diameter is on the order of the air mean free path, the gas temperature gradients cause gas transport through the apertures. The gas transport is driven in a direction from cold to hot by the gas kinetic effect of thermal diffusion. Such a motion of gas leads to an air pressure difference at opposite ends of the nanocompressor, where one end communicates with the ambient atmosphere, and the other end communicates with a sealed chamber being compressed, with the compression deflecting the membrane. The deflection of the air-modulated membrane by compression (or rarefaction) of air moves fluid, such as insulin, which is located behind the membrane.

[0015] As the thermoelectric elements of the nanocompressor become heated, and the adjacent gas / air is compressed, an increased air pressure starts to impact the temperature across the thermoelectric elements through air-surface heat conduction. The higher the air pressure, the smaller the temperature, as well as the temperature difference, of the thermoelectric elements. A decrease in the temperature of thermoelectric elements, and a smaller temperature difference, in turn, impact the properties of the electric circuit, primarily through changes in electrical resistivity and Seebeck coefficients, which affect both the Ohm current and the electromotive force generated by the Seebeck effect.

[0016] Such a reverse impact of the time-dependent air pressure inside the nanocompressor and in the compressed chamber on the properties of the electric circuit indicates that any observable change in that time dependence may be analyzed and interpreted. For occlusions, the deflection of the membrane becomes more difficult, and the air pressure buildup in the compressed chamber will accelerate. For air bubbles, as well as fluid leaks, air pressure buildup will be slower.

[0017] Brief Description of the Drawings

[0018] Figure 1 schematically illustrates a three-chamber microfluidic manifold, where each chamber is driven by its own dedicated thermoelectric nanocompressor.

[0019] Figure 2 schematically illustrates a thermoelectric nanocompressor.

[0020] Figure 3 schematically illustrates a pumping chamber of a microfluidic manifold. Figure 4 shows an exemplary embodiment of the detection circuit for reading and interpreting signals.

[0021] Figure 5 shows qualitative impact of occlusions, air bubbles, and fluid leaks, on time dependent electric potential across a nanocompressor at a constant electric current.

[0022] Figure 6 shows qualitative impact of occlusions and air bubbles on time dependent potential across consecutive stages of a 20-stage thermoelectric nanocompressor, as well as a resultant electric current through the nanocompressor.

[0023] Figure 7 shows qualitative impact of occlusions, air bubbles, and fluid leaks, on time dependent voltage differential between two halves of a nanocompressor at a constant voltage across it.

[0024] Figure 8 shows qualitative impact of different levels of occlusion on time dependent voltage differential between two halves of a nanocompressor at a constant voltage across it.

[0025] Figure 9 shows qualitative impact of air bubble volume on time dependent voltage differential between two halves of a nanocompressor at a constant voltage across it.

[0026] Figure 10 shows qualitative impact of two failure modes, an air bubble and a fluid leak, which occur at the same time, on time dependent voltage differential between two halves of a nanocompressor at a constant voltage across it.

[0027] Detailed Description

[0028] Figure 1 shows schematics of an exemplary embodiment used as a hardware platform for detecting failure modes such as occlusions, air bubbles, and leaks, in the fluidic path of an IIS. The embodiment represents a manifold assembly of a peristaltic pump, which is a type of therapeutic injection device as a non-limiting example. It is discussed in detail in US Provisional Application 63 / 662,031 , and here it is explained schematically to the degree related to the disclosed device and method.

[0029] There are three major components that comprise the manifold assembly. First, there is flexible membrane 101 , which separates the manifold assembly into upper and lower parts. Second, the upper part of the assembly is a fluidic manifold, which has a fluidic inlet 108 that supplies fluid such as insulin, and is connected with a fluid reservoir, a fluidic outlet 109 that is connected with a cannula, and a fluidic channel 105 which connects the inlet 108 through a pumping chamber 103 with the outlet 109. Third, the lower part of the assembly is a pneumatic manifold driven by air. When the manifold assembly depicted in Fig. 1 is used to deliver fluid such as insulin, there is insulin above the membrane 101 , and air below it. Locking mechanisms 106 prevent insulin leaks.

[0030] In the pneumatic manifold (i.e. , lower part of the assembly), an inlet air chamber 102, an air filled part of the pumping chamber 103, and an outlet chamber 104 are all controlled by their own dedicated thermoelectric air nanocompressors 107 (i.e., three dedicated nanocompressors). Each of these three nanocompressors operates independently and deflects the membrane 101 by subsequently increasing and decreasing air pressure in the chamber they respectively control such that . A peristaltic sequence of membrane deflections in the three chambers (i.e., inlet air chamber 102, pumping chamber 103, and outlet chamber 104) moves fluid from the fluidic inlet 108 to the fluidic outlet 109.

[0031] Three nanocompressors may, in a general case, have different geometries, but they all use thermoelectric effect to heat air and thermal diffusion to move it. Their exemplary configurations are discussed in detail in US Patent US 11 ,885,320 and US provisional application 63 / 662,052.

[0032] By this operation, a method of detecting failure modes including at least one of occlusions, air bubbles, leaks, and empty reservoir, in a combination of peristaltic fluidic pumps such as therapeutic injection devices, based on analysis of an electric signal, where a therapeutic injection device comprises: a peristaltic fluidic pump including at least one chamber separated by a flexible membrane into a fluid compartment and an air compartment, at least one thermoelectric air compressor configured to compress air in the air compartment to deflect the flexible membrane and displace fluid in the fluid compartment toward its destination, an electric power supply configured to power the at least one thermoelectric air compressor, and means of reading and interpreting signals from an electric circuit including the air compressor.

[0033] A schematic illustration of a compressor, which may be nano-scale compressor or nanocompressor, is shown in Fig. 2. Here, nanocompressor 201 is mounted at an entrance of an air channel 204 which leads to an inlet air chamber (e.g., chamber 102 in Fig. 1 ). In the shown example, the nanocompressor is powered by a power supply 205, and the bulk air flow comes through nanocompressor entrance 202, and leaves through nanocompressor exit 203. There are three aperture stages in the schematic design, but the actual number of apertures and stages should be determined by a design optimization. Due to gas compression, air pressure near the exit stage will be higher than at the entrance stage. Because air pressure quickly adjusts to changes, air pressure at entrance 202 will be close to the atmospheric pressure, and air pressure at exit 203 will be close to the inlet air chamber pressure. Nanocompressors that control pumping chamber 103 and the outlet chamber 104 are similar or identical to the nanocompressor illustrated in Fig 2. In the exemplary embodiment, the nanocompressor that controls the pumping chamber 103 is used for the occlusion, air bubble, and leak detection, although the inlet and outlet chamber nanocompressors may also be used for that purpose, separately or all of them providing a voltage and / or electric current signal for a detection algorithm. The configuration of the exemplary embodiment is shown in Fig. 3. In this figure, the nanocompressor is schematically illustrated as block 301. When the detection process starts, membrane 303 is fully depressed, keeping air channel 302 closed (the membrane deflection is maintained by a small negative, with respect to the ambient atmosphere, pressure in air channel 302, maintained by nanocompressor 301 ). Pumping chamber 304, at that time, is therefore fully filled with fluid, such as insulin.

[0034] The process of transporting insulin from fluidic inlet to fluidic outlet is performed by applying electric current to nanocompressor 301 in the direction that promotes the air flow from atmosphere to air channel 302, thus increasing air pressure in air channel 302. Increasing pressure in air channel 302 will move membrane 303 all the way up until it is stopped by an upper wall of pumping chamber 304. When this happens, the pumping chamber 304 will be fully filled with elevated-pressure air as an air compartment. It is important to note that to avoid possible undesirable leaks of fluid upstream and downstream, it is necessary to use a pre-seal process. In this process, the compression of air channel 302 starts with both inlet and outlet closed. Then, after some level of pressure is achieved in the air channel 302, an outlet valve is opened. Under normal operation, pre-seal is conducted with membrane 303 in a downward position due to the incompressibility of liquids to act as a fluid compartment.

[0035] The detection algorithm must take into account the fact that the voltage and the electric current that power the nanocompressor will change as functions of the rate of membrane deflection. The functional dependence of these electric properties depends on material properties of thermoelectric elements used by the nanocompressor, primarily on temperature dependence of Seebeck coefficient and electric conductivity of thermoelectric materials, as well as the magnitude of electric properties themselves. Most of the time, both Seebeck coefficient and electric conductivity increase with temperature. Temperatures of both cold and hot sides of thermoelectric elements in turn depend on heat conduction between surfaces and the surrounding air. The higher air pressure, the higher thermal conduction between the air and the surfaces. Radiation also plays a role, albeit a relatively minor one.

[0036] When there is an occlusion downstream from the manifold, be it in the cannula on at the infusion site, a higher air pressure needs to be applied to move membrane 303 upward. That means that high pressure in the air chamber will be reached faster than when there is no occlusion. When there is an air bubble in pumping chamber 304, such that the chamber above the membrane 302 is filled with air and not fluid, higher air pressure below membrane 302 will move membrane 303 even during the pre-seal stage, because air above the membrane 302 may easily be compressed. When there is a leak downstream of the manifold, the membrane 302 will stay in the downward position during the pre-seal stage but then will move much more rapidly after that because there is lower pressure downstream as compared to regular pumping. In all three cases of device failure (i.e., occlusion, air bubble, downstream leak), as well as in case of emptying the fluid reservoir, the deviation of the membrane deflection rate from its regular (expected) profile immediately affects the electric properties of the nanocompressor that controls the membrane 302, primarily voltage and current, and thus may be detected.

[0037] An exemplary embodiment of the detection circuit for reading and interpreting signals is schematically illustrated in Fig 4. Here, eight rounded blocks 401 represent eight stages of a thermoelectric nanocompressor connected in series. The nanocompressor is powered by a dedicated power supply 405. The power supply 405 also provides electric power to detection block 407. The detection block 407 includes three major components: (i) at least one differential amplifier 402 (which may for example be an operational amplifier), which reads the voltage between in and out terminals of the nanocompressor (or, with appropriate circuit modification, the electric current into it), (ii) at least one analog-to-digital converter 403 (ADC) which receives the signal from the amplifier 402, and (iii) a central processing unit 404 which receives and interprets signals from the ADC 403, as well as controls the operation of the nanocompressor and the response to various failure modes. The response may be based on the detection of the entire nanocompressor signal, as well as its separate parts. The latter is schematically illustrated by electrical connector 406 which may be added if necessary to analyze separately the left half (stages one to four) and the right half (stages five to eight) of the exemplary eight-stage nanocompressor. If separate parts of the nanocompressor are tested, a dedicated amplifier and ADC are needed for each part, which then will be used in the failure detection analysis. The actual number of nanocompressor stages, and the stages subject to differential testing and analysis, may generally be a subject of multi- parametric optimization, aimed at finding the most robust and reliable way to detect failures in each specific application.

[0038] A sample detection algorithm may therefore proceed as follows:

[0039] Supply electric current to nanocompressor that controls the pumping chamber

[0040] Read electric potential at in and out terminals of nanocompressor as function of time

[0041] Amplify differential voltage signal read at the terminals

[0042] Digitize amplified differential voltage signal

[0043] Compare digitized signal with expected signal for regular pumping

[0044] Analyze differences between actual and expected signals

[0045] If large differences found (at or above a threshold of 10, 20, 30 or more percent deviation depending on design parameters and testing), run failure mode detection and user notification algorithm

[0046] If negligible differences (less than the threshold), move to the next pumping cycle To provide verification of the disclosed device and method of detecting occlusions, air bubbles, and leaks, numerical simulations are conducted that approximate the response of electric potential and electric current to various obstacles in the fluidic path. These computations were conducted in two stages. First, accurate kinetic modeling was conducted with the direct simulation Monte Carlo (DSMC) method to compute air flow through a nanocompressor. The DSMC modeling was conducted with multi-parametric sweeps over nanocompressor geometries, temperatures of thermoelectric elements, and air conditions. That modeling established air conduction heat fluxes, pressure gradients, and volumetric air flow through a nanocompressor. After that, the obtained DSMC information was used to solve one-dimensional energy conservation / heat conduction equations, coupled with thermoelectric circuit equations, which accounted for the air pressure change due to the operation of the nanocompressor and the deflection of the membrane. Note that for selected cases, one-dimensional simulations have been verified through comparison with full 3D self-consistent finite element computations of electric current and heat propagation through actual nanocompressor geometries.

[0047] An example of the impact of various failure modes, such as impediments in the fluidic path, on the electric potential across a 20-stage nanocompressor as a function of time for a constant current of 20 mA is given in Fig. 5. The membrane deflection starts at time 0. During the pre-seal stage, the membrane is not visibly deflected for regular (no impediments) pumping, as well when there is an occlusion or air leak, because the inlet and outlet valves are closed, and the fluid is incompressible. When there is an air bubble, however, the membrane immediately starts to deflect, and there is only a small change in the voltage slope versus time after the pre-seal stage. After the pre-seal, the occlusion impedes the upward motion of the membrane, and the voltage quickly increases. This is because in this case air pressure is higher in the air channel, which decreases the thermoelectric temperature and temperature difference, and increases the electric potential necessary to support a given electric current. For regular (no-failure) pumping, the membrane starts moving upward relatively quickly, which changes the slope of the voltage curve. If there is a leak, the membrane movement is even more rapid, and the voltage increase is more gradual than for the regular pumping. The numerical modeling indicates that the considered failure modes change the voltage curve of regular pumping by over 0.1 mV, which should be easily detectable with a 24-bit ADC. Another example which illustrates how to detect flow impediments in the fluidic path of IIS is shown in Fig. 6. This example pertains to the case of a constant voltage across a nanocompressor. Here, for a 20-stage nanocompressor, the voltage across it is kept constant at 1.4 V, and the electric potential across each stage is recorded at a time 0.05 s after the start of air compression. The numerical modeling shows that one can expect differences on the order of 10 / iV between the unimpeded and impeded fluid flow if only a voltage across a single stage is considered. The total current through a nanocompressor will also differ, and in the considered case it will decrease by several A if there is an occlusion downstream and increase by several ^A if there is an air bubble in the fluid flow.

[0048] One can also analyze differential signal measurements between the first and the last half of the nanocompressor as a function of time for a constant voltage across the nanocompressor (in the presented case it is the voltage difference between 1 st and 10th stage, and 11th and 20th stage). The resultant dependence for regular pumping, and for different failure modes, is plotted in Fig. 7. Here again, there is a significant impact of occlusions, air bubbles, and leaks, on the voltage. In this case, it is possible to conduct successive signal amplification and then use a simple 12 bit, or even 8 bit ADC to determine the failure mode. Note that detecting air bubbles implies that it is possible to detect both the beginning of the air bubble and its end, potentially taking this into account to achieve proper dosage.

[0049] In the example shown in Fig. 7, failure modes targeted for detection are illustrated as they impact time-dependent voltage difference of a 20-stage nanocompressor powered by a 1.4 V power source. This voltage difference will increase if a larger-voltage battery is used, thus simplifying the detection, and offering an opportunity to increase the reliability of detection of various failure modes. This is illustrated in Fig. 8, where the voltage difference is shown not only for a fully occluded fluidic channel considered in previous examples, but also for several partially occluded channels. In the example shown in Fig. 8, a 2 power source is assumed to power the nanocompressor.

[0050] The disclosed detection algorithm based on the voltage difference between parts of the nanocompressor may be used to evaluate the size of air bubbles passing through the fluidic channel, as the voltage jump in the pre-seal stage will decrease when the size of an air bubble increases relative to the size of the pumping chamber. Whereas the example of Fig. 7 showed the case when the air bubble occupied the entire pumping chamber, the presence of smaller air bubbles may also be detected, as illustrated in Fig. 9 for a 2 V power source. As shown in this figure, air bubble as small as 50 nL can be detected.

[0051] The failure detection cases considered above assume the existence of only one failure mode at a time. The disclosed detection algorithm may also be applied to rare cases when there is more than one failure mode at a time. One such case is shown in Fig. 10, where the voltage difference is plotted versus time for the regular pumping with no failure modes, for a case with a 125 nL air bubble, and for two cases when there are an air bubble and a fluid leak at the same time. Two fluid loss rates are considered here, a big leak where all fluid is lost to a leak, and a small leak where only about 10% of the fluid is lost.

[0052] The above examples illustrate the detection of various failure modes in infusion sets by deflecting the membrane in the direction that displaces the pumped fluid towards its destination. It is also possible to use the deflection of the membrane in the opposite direction to detect an empty reservoir. Pumping will not be significantly deflecting the membrane in this case, thus resulting in changes of the actual versus expected voltage.

[0053] For the detection to be reliable, calibration is conducted of the regular pumping signal, such as in a self-test, which then will be used as a reference point for detection analysis, where the actual signal is compared with the expected signal. In addition to analyzing the rate of increase or decrease of the electric signal (voltage and / or current), one can also analyze the asymptotic signal for regular pumping and various failure modes.

Claims

Claims1 . An injection device for detecting failure modes including at least one of occlusions, air bubbles, leaks, and empty reservoir, based on analysis of an electric signal, said injection device comprises: a peristaltic fluidic pump including at least one chamber separated by a flexible membrane into a fluid compartment and an air compartment, at least one thermoelectric air compressor configured to compress air in the air compartment to deflect the flexible membrane and displace fluid in the fluid compartment toward its destination, an electric power supply configured to power the at least one thermoelectric air compressor, and means of reading and interpreting electric signals from an electric circuit including the air compressor.

2. The injection device claim 1 , wherein the electric signal to be interpreted is electric potential (voltage) across the air compressor and / or components of the air compressor.

3. The injection device of claim 1 , wherein the electric signal to be interpreted is electric current passing through the air compressor and / or components of the air compressor.

4. The injection device of claim 1 , further comprising: a fluidic inlet connected with a fluid reservoir; and a fluidic outlet connected with a cannula.

5. The injection device of claim 1 , further comprising: a fluidic channel connecting a fluidic inlet with a fluidic outlet via the fluid compartment of the at least one chamber.

6. The injection device of claim 1 , wherein the at least one thermoelectric air compressor is mounted at an entrance of an air channel that leads to an inlet air chamber.

7. The injection device of claim 5, wherein fluid is transported from the fluidic inlet to the fluidic outlet by applying electric current to the at least one thermoelectric air compressor.

8. The injection device of claim 1 , wherein the electric signals from the electric circuit change as functions of a rate of membrane deflection.

9. The injection device of claim 1 , wherein a deviation of membrane deflection rate from a preset profile affects electrical properties of the at least one thermoelectric air compressor.

10. The injection device of claim 1 , wherein the electric circuit includes (i) at least one differential amplifier, (ii) at least one analog-to-digital converter, and (iii) a central processing unit.11 . The injection device of claim 1 , further comprising: a locking mechanism that locks a first portion of the peristaltic fluid pump that includes the fluid compartment to a second portion of the peristaltic fluid pump that includes the air compartment.

12. The injection device of claim 1 , wherein the at least one chamber includes a pumping chamber, an inlet air chamber, and an outlet chamber, and wherein the inlet air chamber, the outlet chamber and the air compartment of the pumping chamber are controlled by respective thermoelectric air compressors.

13. The injection device of claim 1 , wherein the flexible membrane separates the peristaltic fluidic pump into an upper part and a lower part, and wherein the upper part includes the fluid compartment, and the lower part includes the air compartment.

14. The injection device of claim 13, wherein the upper part of the peristaltic fluidic pump includes (i) a fluidic inlet that supplies fluid and that connects with a fluid reservoir and (ii) a fluidic outlet that connects with a cannula.

Images