Description:Field of the Invention
The present invention technical field relates to an insulin delivery system, in particular to a system to insert sufficient insulin in terms of stability and speed for controlling glucose concentration in a DT-1 patient.
Background of the Invention
According to the International Diabetes Federation (IDF), 537 million persons globally had diabetes mellitus in 2021, with Malaysia, China, and Cambodia having the highest rates of the disease. This figure is expected to increase to 783 million by 2045. These figures and forecasts motivate effective research to be done to boost DT-1 patient prevention, enhance techniques, and preferably find a cure.
Dubey, Vandana, Harsh Goud, and Prakash C. Sharma. ‘Role of PID Control Techniques in Process Control System: A Review’. In Data Engineering for Smart Systems, 659–70. Lecture Notes in Networks and Systems. Singapore: Springer Singapore, 2022. https://doi.org/10.1007/978-981-16-2641-8_62 discloses that FO-PID requires sophisticated tuning because it contains 5 tuning parameters. The five parameters that need to be tuned are Kp (proportional), Ki (integral), Kd (derivative), ? (integral component of fractional order), and µ (derivative component of fractional order).
Farahmand, Bahare, Maryam Dehghani, and Navid Vafamand. ‘Fuzzy Model-Based Controller for Blood Glucose Control in Type 1 Diabetes: An LMI Approach’. Biomedical Signal Processing and Control 54, no. 101627 (September 2019): 101627. https://doi.org/10.1016/j.bspc.2019.101627.
Goud, Harsh, Prakash Chandra Sharma, Pankaj Swarnkar, Venkatesh Gauri Shankar, Vijay Prakash Sharma, and Anil Kumar Sahu. "A comparative analysis of conventional PID tuning techniques for single link robotic arm." Solid State Technology 64, no. 2 (2021): 565-574.
Garg, Himanshu, and Jyoti Yadav. ‘Optimal Fuzzy-PID Controller Design by Grey Wolf Optimization for Renewable Energy-Hybrid Power System’. In Control Applications in Modern Power System, 77–88. Lecture Notes in Electrical Engineering. Singapore: Springer Singapore, 2021. https://doi.org/10.1007/978-981-15-8815-0_7.
Ebrahimi, Nahid, Sadjaad Ozgoli, and Amin Ramezani. ‘Model Free Sliding Mode Controller for Blood Glucose Control: Towards Artificial Pancreas without Need to Mathematical Model of the System’. Computer Methods and Programs in Biomedicine 195, no. 105663 (October 2020): 105663. https://doi.org/10.1016/j.cmpb.2020.105663.
Franco, Roberto, Alejandra Ferreira de Loza, Hector Rios, Louis Cassany, David Gucik-Derigny, Jerome Cieslak, David Henry, and Loic Olcomendy. ‘Output-Feedback Sliding-Mode Controller for Blood Glucose Regulation in Critically Ill Patients Affected by Type 1 Diabetes’. IEEE Transactions on Control Systems Technology: A Publication of the IEEE Control Systems Society 29, no. 6 (November 2021): 2704–11. https://doi.org/10.1109/tcst.2020.3046420.
Swarnkar, Pankaj, and Harsh Goud. ‘Design of Fuzzy Adaptive Pi Controller for Inherently Unstable System’. SSRN Electronic Journal, 2020. https://doi.org/10.2139/ssrn.3623747.
Goud, Harsh, and Pankaj Swarnkar. ‘Signal Synthesis Model Reference Adaptive Controller with Genetic Algorithm for a Control of Chemical Tank Reactor’. International Journal of Chemical Reactor Engineering 17, no. 7 (26 July 2019). https://doi.org/10.1515/ijcre-2018-0199.
Sachan, S., H. Goud, and P. Swarnkar. ‘Performance and Stability Analysis of Industrial Robot Manipulator’. In Intelligent Computing Techniques for Smart Energy Systems: Proceedings of ICTSES 2021, 473–81. Singapore; Singapore: Springer Nature, 2022. https://doi.org/10.1007/978-981-19-0252-9_43.
Goud, Harsh, Prakash Chandra Sharma, Kashif Nisar, Muhammad Reazul Haque, Ag Asri Ag. Ibrahim, Narendra Singh Yadav, Pankaj Swarnkar, Manoj Gupta, and Laxmi Chand. ‘Metaheuristics Algorithm for Tuning of PID Controller of Mobile Robot System’. Computers, Materials & Continua 72, no. 2 (2022): 3481–92. https://doi.org/10.32604/cmc.2022.019764.
Balakrishnan, Nagaraj, and K. Nisi. ‘A Deep Analysis on Optimization Techniques for Appropriate PID Tuning to Incline Efficient Artificial Pancreas’. Neural Computing & Applications 32, no. 12 (June 2020): 7587–96. https://doi.org/10.1007/s00521-018-3687-7.
Shijo, Johnson K., Thanaraj K. Palani, and S. Senthil Kumar. ‘Design of Controllers for T1DM Blood Glucose Insulin Dynamics Based on Constrained Firefly Algorithm’. In 2018 4th International Conference on Electrical Energy Systems (ICEES). IEEE, 2018. https://doi.org/10.1109/icees.2018.8443246.
Belmon, Anchana P., and Jeraldin Auxillia. ‘An Adaptive Technique Based Blood Glucose Control in Type-1 Diabetes Mellitus Patients’. International Journal for Numerical Methods in Biomedical Engineering 36, no. 8 (August 2020): e3371. https://doi.org/10.1002/cnm.3371.
The above prior art provide the solution of the insulin delivery problem. The usage of a fuzzy logic-based controller was suggested by Farahmand et al. [1] to control insulin infusion, and it was tested using the BMM model, the Tolic model, and the meal disruption. Numerous cutting-edge controllers are used by researchers in addition to the conventional PID controller [2, 3]. For managing Blood Sugar Levels, a variety of computing methods can be applied [4, 5]. These techniques are used to fine-tune the controller's parameters and control behavior [6]. Adaptive PID controller techniques might be applied in these circumstances to identify patients with potentially fatal conditions [7]. The use of metaheuristic optimization techniques is common in intelligent diagnostic systems that aim to improve classification accuracy [8, 9]. In [10], the researcher has focused on the configuration of various control methods to control blood glucose levels in DT-1 patients where the GWO-PID controller provides a promising result with rise time (0.32), settling time (0.983), overshoot (0), and integral-of-square-error-ISE (0.66). In [11], the researcher used a Firefly-PID controller with integral-of-time absolute-error-ITAE (125000), ISE (6431), rise time (0.8526), settling time (482.89), and overshoot (41.88) for glucose control. To enhance the parameters of the conventional PID controller to control the BSL, Parsa [12] presented a contemporary method Grasshopper in the year 2020 and contrasted it with other methods like Elephant Herding and PSO technique. PID controller, which is frequently employed in control systems, is most widely used in IDS (insulin delivery system) control mechanism as shown in Figure 5.
Due to the constraints of the FO-PID tuning, the present invention system focus on tuning only the 3 (Kp, Ki, and Kd) PID parameters. When compared to traditional tuning techniques, adaptive metaheuristic tuning techniques offer improved search efficiency.

Object of the Invention
The focal objective of the adaptive PID controller is to create a control law u(t) which enables the system to attain the necessary performance as shown in Figure 4.

Drawings
Figure 1: Circuit Diagram of Automatic Insulin Delivery System
Figure 2: Step Response comparison of IDS DT-1 using various control technologies. (a) IDS DT-1 with ISE (b) IDS DT-1 with ITAE
Figure 3: Impulse Response comparison of IDS DT-1 using various control technologies. (a) IDS DT-1 with ISE. (b) IDS DT-1 with ITAE
Figure 4: PID Controller with IDS DT-1
Figure 5: IDS Control Schematic
Figure 6: Simulation Setup of Automatic Insulin Delivery System

Detailed Description of the Invention

The present invention relates to an insulin delivery system for Type-1 diabetic patients. The present invention delivery system comprises an adaptive PID (Proportional Integral Derivative) controller and several metaheuristic computations including PSO (Particle-Swarm-Optimization), ACO (Ant-Colony-Optimization), FPO (Flower Pollination-Optimization), and GWO (Grey-Wolf-Optimization) for controlling insulin delivery in a DT-1 patient. These computing techniques improve the parameters of the adaptive controller, and a comparison is made by taking both step and impulsive reactions into account.
On embodiment of the invention propose a step-size strategy that is adaptive and that reduces the step size whenever adequate progress is not obtained. The simulated results demonstrate that the adaptive step-size technique maintains the rapid initial convergence speed of the traditional metaheuristic PID tuning approaches for Insulin Delivery System (IDS) controlling while greatly enhancing the stability and robustness of the reconstruction toward IDS's error. The IDS’s errors are much less when IDS is controlled by the AGWO-PID controller in comparison to the other tuning techniques. The IDS’s stability can be enhanced by the AGWO-PID controller with optimum values of ITAE and ISE respectively. Results from simulations demonstrate that the suggested AGWO-PID controller can guarantee control of the IDS DT-1 system's performance while lowering expenses. The IDS DT-1 system with the AGWO-PID controller also displayed outstanding robustness to variations in hormone and food sensitivity as well as a greater capacity for subject variability adaptation.
To demonstrate the working of the PID method in an insulin delivery system (IDS), we have developed hardware. The circuit diagram for automatic IDS is shown in Figure 1. This hardware will continuously monitor the blood glucose and will deliver the appropriate amount of insulin to the patient's body. The proposed hardware consists of a blood glucose sensor that gives analog signal output, and the output goes to the processor, we have used an Arduino board for processing purposes. The Processing Unit will drive a bipolar stepper motor and a mechanism that converts the rotary motion of the motor to linear motion and will push the injection in a controlled manner.
To design adaptive PID controllers, this paper changes the configuration of traditional metaheuristic tuning techniques during the search by using the optimum step size strategy and also uses the linear simulator for the same and gets more optimized PID controlling parameters for IDS controlling. The responses in Figure 2 and Figure 3 demonstrate that without a controller the IDS is not responding properly to the step input as well as for impulse input respectively. These findings clearly show how unstable the system is, necessitating the use of a controller capable of managing and controlling the IDS in order to provide patients with an adequate supply of insulin. The response of the adaptive metaheuristic PID controllers is much better than the other traditional metaheuristic PID controllers. Simulation results of the adaptive GWO-PID controller beat the other adaptive state-of-the-art controllers.
Detail about Computing Technique: Adaptive Grey-Wolf-Optimization (AGWO):
The headship structure and chasing strategy of grey wolves in nature are modeled in the GWO method. The headship ladder is simulated using four different sorts of grey wolves, including alpha (spearhead), beta (subsidiary), delta, and omega (lowermost). This research work expands the search space as well as search parameters to make the GWO-PID controller an adaptive GWO-PID controller for IDS controlling and get better outcomes than other states of art controllers.
Key-Terms: Wolf i, Population N set of entire i, SS [lb,ub], Location X, The hunting is guided by: Alpha (a) is the first-best answer, beta (ß) is second-best, and delta (d) is third best. As the iterations progress, the component (a) linearly drops from 2 to 0.
Tuning Method for the Automation of IDS:
i. Set all the initial input parameters like FF, N, number of iteration T, coefficient value (a), and number of PID parameters (Kp, Ki, and Kd.).
ii. Start a loop for t=1: N; and Find , , .

Start an inner loop for n=1: number of PID parameters; after that calculate the value of N, , and .

iii. Update the position of N, find the tuned value of Kp, Ki, and Kd., also calculate the new minimized value of IDS’s error.
While neither the minimum IDS’s error condition nor the maximum iterations are met, then start a loop again.
iv. Print the ideal response.
PID controller’s constraints for both FFs (ITAE and ISE) as well as for conventional PID which were found by the offline tuning method (CHR: Chien-Hrones-Reswick) and intelligent tuning methods are given in Table 1, which are calculated through MATLAB simulation environment. Table 2 shows the BSL rating of the insulin system. Simulation Setup of our automatic Insulin Delivery System has been shown in figure 6.
Table 1: PID constraints of conventional and intelligent techniques for IDS DT-1
Control Techniques

Constraints
Conventional (CHR)–PID Controller Existing Metaheuristic Computing Method based PID Controller Proposed Adaptive Metaheuristic Computing Method based PID Controller
PSO-PID (ITAE/
ISE) ACO-PID (ITAE/
ISE) FPA-PID (ITAE/
ISE) GWO-PID (ITAE/
ISE) APSO-PID
(ITAE/
ISE) AACO-PID (ITAE/
ISE) AFPA-PID (ITAE/
ISE) AGWO-PID (ITAE/
ISE)

4.8080 17.887/ 36.098 9.4/
9.9 9.628/ 3.131 30.00/
25.954 28.909/
20.260 2.9/
9.7 7.535/
9.282 11.26/
21.801

0.7491 1.447/ 6.109 1.8/
2.2 0.147/
0.531 17.258/
2.041 0.01/
0.201 0.01/
0.5 0.121/
0.249 0.011/
1.161

1.9256 29.554/30 2.8/
2.7 2.3789/ 2.783 16.533/
29.056 30/
18.065 2.1/
9.6 7.534/
6.356 12.431/
23.844

Table 2: BSL Rating of Insulin System
BSL Rating BSL Range
Normal Sugar 70 to 90 mg/dL
Pre Sugar 100 to 125 mg/dL
Sugar Above 126 mg/dL

In order to reduce the risks of hyperglycemia and hypoglycemia in DT-1 patients, the present system provides th econtrol measures. The zero-order hold method has been used to obtain the discrete-time transfer function of the insulin delivery system (IDS) and convert it into the continuous s-domain. Models of the IDS have been used to test the effectiveness of both conventional and non-conventional methods. Novel approaches are employed for calculating the PID controller parameters with the usage of the swarm, nature, and population-based methods. When it came to tracking the desired glucose level (70 mg/dL), the developed nonlinear control law i.e. AGWO-PID controller outperformed the linear control strategies. The suggested AGWO method can adaptively adjust the parameters for better performance and has successfully detected the proper dose of insulin in DT-1 patients. The proposed AGWO-PID controller integrates the new time-response specification with the Grey-Wolf-Optimization (GWO) method. The results graphically and analytically are analyzed with time response subjected to a unit step as well as impulse functions, frequency response bode plot, root locus, and Nyquist analysis. There is an optimal improvement in more petite rise time, settling time, maximum overshoot, peak time, average output levels, integral-of-square-error (ISE), integral-of-time absolute error (ITAE), and IDS stability with AGWO-PID controller. Future research may focus on controlling the IDS using adaptive intelligent hybrid controllers.
, Claims:1. An Insulin Delivery System for Type-1 Diabetic Patients, the system comprises of: an adaptive PID (Proportional Integral Derivative) controller and several metaheuristic computations.
2. The Insulin Delivery System for Type-1 Diabetic Patients as claimed in the claim1 wherein several metaheuristic computations includes PSO (Particle-Swarm-Optimization), ACO (Ant-Colony-Optimization), FPO (Flower Pollination-Optimization), and GWO (Grey-Wolf-Optimization) for controlling insulin delivery in a DT-1 patient.
3. The Insulin Delivery System for Type-1 Diabetic Patients as claimed in the claim1 wherein Adaptive step-size strategy is employed for calculating the PID controller parameters with the usage of the swarm, nature, and population-based techniques.
4. The Insulin Delivery System for Type-1 Diabetic Patients as claimed in the claim1 wherein the device of the system consists of:
a) a blood glucose sensor
b) Arduino board processor
c) bipolar stepper motor
d) bipolar stepper driver
e) Insulin injection pump
all the above (a),(b),(c),(d) and (e) are interlinked by circuit and built together in main device.
5. The device and Insulin Delivery System for Type-1 Diabetic Patients as claimed in claim 4 wherein blood glucose sensor gives analog signal output, and the output goes to the processor, processor and bipolar stepper motor converts the rotary motion of the motor to linear motion and will push the injection in a controlled manner.
6. The device and Insulin Delivery System for Type-1 Diabetic Patients as claimed in claim 4 wherein system used to get the design adaptive PID controllers, system modified the configuration of traditional metaheuristic tuning techniques during the search by using the optimum step size strategy and also uses the linear simulator for the same and gets more optimized PID controlling parameters for IDS controlling.
7. The device and Insulin Delivery System for Type-1 Diabetic Patients as claimed in claim 1 and 4 wherein system’s AGWO-PID controller integrates the new time-response specification with the Grey-Wolf-Optimization (GWO) method to tracking the desired glucose level (70 mg/dL).
8. The device and Insulin Delivery System for Type-1 Diabetic Patients as claimed in claim 1 and 4 wherein AGWO method can adaptively adjust the parameters for better performance and has successfully detected the proper dose of insulin in DT-1 patients.