An optimization method for the temperature controller of an infant incubator

Abstract

The present invention discloses an optimization method for the temperature controller of an infant incubator, belonging to the technical field of PID control optimization. The specific steps are as follows: Step 1, construct a temperature controller system for the infant incubator; Step 2, improve the sewing training optimization algorithm. The specific improvement strategies are: D1, use a multi-position adaptive update strategy to improve the mathematical model in the training stage of the sewing training optimization algorithm; D2, use a dynamic threshold penalty update strategy to improve the mathematical model in the practice stage of the sewing training optimization algorithm; Step 3, use the improved sewing training optimization algorithm to optimize the PID controller in the temperature controller system of the infant incubator; Step 4, use the values of the control parameters Kp, Ki, and Kd obtained by optimization in Step 3 for the PID controller in the temperature controller system of the infant incubator; Through the optimization of the temperature controller system of the infant incubator by the improved algorithm, the robustness and control performance of the system are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of PID control optimization, and particularly relates to an optimization method for the temperature controller of an infant incubator. Background Art

[0002] The temperature controller of an infant incubator is a key device specifically designed to maintain an appropriate temperature for newborns in the incubator. Especially for premature infants and newborns with too low body weight, their thermoregulatory ability is relatively poor and they are easily affected by changes in environmental temperature. Therefore, the temperature control of the incubator is particularly important. The basic working principle of this controller is to continuously monitor the temperature inside the incubator through a high-precision temperature sensor, compare it with the set target temperature, and automatically adjust the power of the heater to ensure that the temperature inside the box remains within a constant and safe range. To achieve this goal, a proportional-integral-derivative (PID) controller is usually adopted. This kind of controller can accurately adjust according to the current temperature, set temperature and temperature change rate, thereby effectively reducing temperature fluctuations and ensuring the comfort and safety of newborns in the incubator.

[0003] A PID controller is a feedback controller widely used in industrial control systems, aiming to adjust the control input to make the system output reach the set target. It consists of three parts: proportional (P), integral (I), and derivative (D). The proportional part adjusts according to the current error (the difference between the set value and the actual value), the integral part accumulates past errors to eliminate steady-state errors, and the derivative part predicts future error changes to improve the response speed of the system. By reasonably adjusting these three parameters, the PID controller can achieve a fast and stable system response and is widely used in various control applications such as temperature, pressure, and flow.

[0004] The Sewing Training Optimization Algorithm (STBO) is a new human-based meta-heuristic algorithm. Its basic inspiration is to teach novice tailors the sewing process. Mathematically, it is modeled in three stages: training, imitating the skills of the teacher, and practice. The Sewing Training Optimization Algorithm has a high ability to balance search and exploitation. Summary of the Invention

[0005] The object of the invention is to improve the Sewing Training Optimization Algorithm, and use the improved Sewing Training Optimization Algorithm to optimize the PID controller in the temperature control system of the infant incubator, so as to improve the temperature control accuracy and response speed of the system, automatically adapt to external environmental changes, quickly find the optimal control parameters, reduce the dependence on manual adjustment, ensure the stability of the temperature in the infant incubator, avoid health risks to infants caused by too high or too low temperature, improve the operation efficiency and safety of the equipment, and ultimately ensure the comfort and safety of infants in the incubator and improve the reliability and accuracy of medical equipment.

[0006] To achieve the above object, the present invention adopts the following technical solutions.

[0007] Step 1: Construct a temperature controller system for an infant incubator. The system includes: an instruction operation module, a temperature difference calculation module, a PID controller, an improved sewing training optimization algorithm module, a temperature adjustment module, and a temperature sensor module.

[0008] Step 2: Improve the sewing training optimization algorithm, specifically as follows:

[0009] D1. Use a multi-position adaptive update strategy to improve the mathematical model in the training stage of the sewing training optimization algorithm. The multi-position adaptive update strategy guides the current population individuals through the optimal position and the average position in the population. The average position of the population is adaptively adjusted through the iteration stage of the algorithm, and the current position adaptively adjusts the update step size according to the optimal position and the average position, so as to update the position.

[0010] D2. Use a dynamic threshold penalty update strategy to improve the mathematical model in the practical stage of the sewing training optimization algorithm. The dynamic threshold penalty update strategy dynamically generates a threshold according to the current iteration stage and the fitness value of the individuals in the population, calculates the penalty factor P of the current individual through the threshold, and the penalty factor P affects the quality of the population position update. If the position of the individual is closer to the optimal position of the current population, the fitness value is rewarded and the position is retained. If the individual position deviates from the optimal position in the current population, the fitness value is punished, and the individual position is updated through the penalty factor P.

[0011] Step 3: Use the improved sewing training optimization algorithm to optimize the PID controller in the infant incubator temperature controller system, and obtain an optimal set of control parameters Kp, Ki, and Kd through the improved algorithm.

[0012] Step 4: Use the values of the control parameters Kp, Ki, and Kd optimized in Step 3 for the PID controller in the infant incubator temperature controller system, and adjust the control quantity to reach the target temperature.

[0013] Preferably, the specific execution process of the temperature control system for the infant incubator constructed in Step 1 is as follows: First, set the target temperature value through the instruction operation module, input the target temperature value and the temperature monitored by the temperature sensor in real time into the temperature difference calculation module to obtain the temperature difference e(t), input the temperature difference e(t) into the PID controller, optimize the PID control through the improved sewing training optimization algorithm module to obtain the control parameters Kp, Ki, and Kd, output the control quantity u(t) according to the optimized control parameters, and the temperature adjustment module adjusts the temperature in the incubator according to the control quantity u(t). The mathematical model of the temperature adjustment module is:

[0014] (1);

[0015] In formula (1), T represents the temperature inside the incubator, and C represents the specific heat capacity of the incubator. It is indicated that the heat power of the heater is regulated by a PID controller, specifically as shown in formula (2). It is indicated that the temperature loss is related to the ambient temperature and heat exchange.

[0016] (2);

[0017] In formula (2), Kp represents the proportionality coefficient, Ki represents the integral coefficient, Kd represents the differential coefficient, I represents the total running time of the system, and e(t) represents the real-time temperature difference.

[0018] Preferably, the mathematical model in D1 for improving the training stage of the sewing training optimization algorithm using a multi-position adaptive update strategy is as follows:

[0019] (3);

[0020] In formula (3), X(iter + 1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r1 represents a random number between [0, 1], Xavg(iter) represents the average position in the population, the calculation formula is as shown in formula (4), uniform(-1, 1) represents a uniform distribution between [-1, 1], and Xbest represents the best position in the population.

[0021] (4);

[0022] In formula (4), Xavg(iter + 1) represents the updated average position, N represents the number of individuals in the population, X(i) represents the position of the i-th individual, a represents the update constant, K(iter) represents the adaptive attenuation coefficient, the calculation formula is as shown in formula (5), Xbest represents the best position in the population, and Xavg(iter) represents the average position at the current iteration number.

[0023] (5);

[0024] In formula (5), K(iter) represents the adaptive attenuation coefficient, k0 represents the initial coefficient, λ represents the adjustment factor and takes a random number between [0.1, 0.5], iter represents the current iteration number, max_iter represents the maximum iteration number, and randn(0, 1) represents a random number generated by the standard normal distribution.

[0025] Preferably, by combining the optimal position and average position information, the algorithm can better balance global search and local search. The adjustment of the adaptive decay coefficient and step size enables the algorithm to dynamically adjust the search strategy according to the current iteration stage, improving the convergence speed and accuracy, helping the algorithm to jump out of the local optimal solution, and enhancing the robustness and stability of the algorithm.

[0026] Preferably, the mathematical model in the practice stage of improving the sewing training optimization algorithm using a dynamic threshold penalty update strategy described in D2 first generates a dynamic threshold P, and the specific mathematical model is:

[0027] (6);

[0028] In formula (6), Thresh(iter) represents the dynamic threshold, t0 represents the initial threshold, iter represents the current iteration number, max_iter represents the maximum iteration number, ω represents the threshold change influence factor, Fbest represents the optimal fitness value at the current iteration number, and Favg represents the average fitness value at the current iteration number;

[0029] Calculate the penalty factor P of the individual through the value of the threshold Thresh(iter), and the specific calculation formula is:

[0030] (7);

[0031] In formula (7), μ represents the coefficient factor controlling the penalty intensity, F represents the fitness value of the current individual, and Thresh represents the dynamic threshold as shown in formula (6);

[0032] Update the position of the current individual through the obtained penalty factor P, and the specific calculation formula is:

[0033] (8);

[0034] In formula (8), X(iter + 1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r3 represents a random number between [0, 1], lb represents the lower limit of the search space, ub represents the upper limit of the search space, b represents the adjustment factor controlling the fitness value difference, F represents the fitness value of the current individual, Fbest represents the optimal fitness value at the current iteration number, and P represents the penalty factor as shown in formula (7);

[0035] Finally, take reward and punishment measures for the fitness value of the updated individual, and the specific calculation formula is as follows:

[0036] (9);

[0037] In Equation (9), F represents the fitness value of an individual after update, P represents the penalty factor as shown in Equation (7), Fbest represents the optimal fitness value at the current iteration, and Thresh represents the dynamic threshold, as shown in Equation (6).

[0038] Preferably, by introducing the dynamic threshold Thresh(iter), this method can adjust the fitness threshold according to the current iteration, thereby adaptively balancing exploration and exploitation during the optimization process. At the initial stage of the algorithm, a larger threshold range provides a larger search space for individuals, encouraging extensive exploration; while in the later stage of the optimization process, the threshold gradually decreases, making the search more focused on the optimization of local accuracy and avoiding premature convergence to local optimal solutions. By calculating the difference between the fitness of each individual and the threshold and combining the penalty mechanism, individuals whose fitness values exceed the threshold will be penalized. This penalty mechanism effectively reduces the impact of inappropriate individuals on the overall optimization process and at the same time prompts individuals to gradually approach the global optimal solution.

[0039] Preferably, in step three, the improved sewing training optimization algorithm is used to optimize the PID controller in the baby incubator temperature control system. The specific steps are as follows:

[0040] S1. Initialize the parameters of the improved sewing training optimization algorithm, including the number of population individuals N, the problem dimension dim, the maximum number of iterations max_iter, the upper bound ub of the search space, and the lower bound lb of the search space. Generate the initial positions of the population through the initial parameters;

[0041] S2. Establish a mapping relationship between the improved sewing training optimization algorithm and the PID controller in the baby incubator temperature control system, convert the temperature control of the actual system into a mathematical model to be optimized, and correspond the three dimensions of the individual position vector of the algorithm to the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd of the PID controller, that is, X(x1, x2, x3) = (Kp, Ki, Kd). Through continuous optimization of the algorithm, the position vector of the individual is continuously updated, thereby dynamically adjusting the PID control parameters and finally obtaining the optimal set of PID control parameters for the system;

[0042] S3. Construct the fitness value function of the improved sewing training optimization algorithm, calculate the fitness values of the individuals in the population through the fitness value function, and select the optimal individual and the optimal fitness value. The specific fitness value function is:

[0043] (10);

[0044] In Equation (10), J represents the fitness value, L represents the total running time of the system, and e(t) represents the real-time temperature difference;

[0045] S4. Update the individual position through the mathematics of the improved sewing training optimization algorithm. The specific steps are as follows:

[0046] S41. Execute the training stage of the algorithm. The specific mathematical model is shown in Equation (3);

[0047] S42. Execute the stage of imitating the teacher's skills of the algorithm. Generate a set of imitation variables according to the number of iterations. Take the positions of the individuals in the population whose individual fitness values are less than the current individual fitness value as the mentor positions to generate a mentor set. Each mentor is a three-dimensional vector, and each dimension represents a skill. The updated individual selects three variables through the imitation variable set and combines them with the dimensions in the corresponding mentor set to generate a new position as the updated position;

[0048] S43. Execute the practice stage of the algorithm. The specific mathematical models are shown in Equations (6)-(8);

[0049] S5. Determine whether the current number of iterations has reached the maximum number of iterations. If not, continue the optimization. If so, output the optimal individual vector in the population as the optimal parameter.

[0050] By adopting the above technical solutions, the beneficial effects of the present invention are as follows: By improving the sewing training optimization algorithm, the adaptability of the algorithm in the optimization process is enhanced, the optimization capabilities of the exploration and development stages of the algorithm are balanced, the convergence speed and optimization accuracy of the algorithm are improved, better robustness is shown when dealing with complex problems, and applying the improved sewing training optimization algorithm to the PID controller of the baby incubator temperature control system can improve the control accuracy of the PID controller, thereby improving the control ability and performance of the baby incubator temperature control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flowchart of an optimization method for a baby incubator temperature controller.

[0052] Figure 2 It is a model diagram of a baby incubator temperature control system.

[0053] Figure 3 It is a comparison diagram of the change of fitness values during the optimization process between the improved sewing training optimization algorithm and the standard sewing training optimization algorithm.

[0054] Figure 4 It is a comparison diagram of the response of optimizing the PID controller between the improved sewing training optimization algorithm and the standard sewing training optimization algorithm. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0056] The present invention provides a technical solution: an optimization method for the temperature controller of an infant incubator, which specifically includes the following steps, as Figure 1 shown.

[0057] Step 1: Construct a temperature controller system for an infant incubator. The system includes: an instruction operation module, a temperature difference calculation module, a PID controller, an improved sewing training optimization algorithm module, a temperature adjustment module, and a temperature sensor module, as Figure 2 shown.

[0058] Furthermore, the specific execution process of the temperature control system for the infant incubator constructed in Step 1 is as follows: First, set the target temperature value through the instruction operation module, input the target temperature value and the temperature monitored by the temperature sensor in real time into the temperature difference calculation module to obtain the temperature difference e(t), input the temperature difference e(t) into the PID controller, optimize the PID control through the improved sewing training optimization algorithm module to obtain the control parameters Kp, Ki, Kd, output the control quantity u(t) according to the optimized control parameters, and the temperature adjustment module adjusts the temperature in the incubator according to the control quantity u(t). The mathematical model of the temperature adjustment module is:

[0059] (1);

[0060] In formula (1), T represents the temperature in the incubator, C represents the specific heat capacity of the incubator, represents the heat power of the heater regulated by the PID controller, as specifically shown in formula (2), represents that the temperature loss is related to the ambient temperature and heat exchange;

[0061] (2);

[0062] In formula (2), Kp represents the proportional coefficient, Ki represents the integral coefficient, Kd represents the differential coefficient, I represents the total running time of the system, and e(t) represents the real-time temperature difference.

[0063] Step 2: Improve the sewing training optimization algorithm, and the specific improvement is as follows:

[0064] D1. Improve the mathematical model in the training stage of the sewing training optimization algorithm using a multi-position adaptive update strategy. The multi-position adaptive update strategy guides the individuals in the current population through the optimal position and the average position in the population. The average position of the population is adaptively adjusted in the iterative stage of the algorithm, and the current position adaptively adjusts the size of the update step according to the optimal position and the average position, thereby updating the position.

[0065] D2. Improve the mathematical model in the practical stage of the sewing training optimization algorithm using a dynamic threshold penalty update strategy. The dynamic threshold penalty update strategy dynamically generates a threshold according to the current iteration stage and the fitness values of the individuals in the population, calculates the penalty factor P of the current individual through the threshold, and the penalty factor P affects the quality of the population position update. If the position of an individual is closer to the optimal position of the current population, its fitness value is rewarded and the position is retained. If the position of an individual deviates from the optimal position in the current population, its fitness value is penalized, and the individual position is updated through the penalty factor P.

[0066] Furthermore, for the improvement of the mathematical model in the training stage of the sewing training optimization algorithm using a multi-position adaptive update strategy in D1, the improved mathematical model is specifically as follows:

[0067] (3);

[0068] In formula (3), X(iter + 1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r1 represents a random number between [0, 1], Xavg(iter) represents the average position in the population, and the calculation formula is as shown in formula (4). uniform(-1, 1) represents a uniform distribution between [-1, 1], and Xbest represents the best position in the population.

[0069] (4);

[0070] In formula (4), Xavg(iter + 1) represents the updated average position, N represents the number of population individuals, X(i) represents the position of the i-th individual, a represents the update constant, K(iter) represents the adaptive decay coefficient, and the calculation formula is as shown in formula (5). Xbest represents the best position in the population, and Xavg(iter) represents the average position at the current iteration number.

[0071] (5);

[0072] In Equation (5), K(iter) represents the adaptive attenuation coefficient, k0 represents the initial coefficient, λ represents the adjustment factor which takes a random number between [0.1, 0.5], iter represents the current iteration number, max_iter represents the maximum iteration number, and randn(0,1) represents a random number generated from the standard normal distribution.

[0073] Furthermore, for the mathematical model in the practice stage of improving the sewing training optimization algorithm using a dynamic threshold penalty update strategy in D2, first generate the dynamic threshold P, and the specific mathematical model is:

[0074] (6);

[0075] In Equation (6), Thresh(iter) represents the dynamic threshold, t0 represents the initial threshold, iter represents the current iteration number, max_iter represents the maximum iteration number, ω represents the threshold change influence factor, Fbest represents the optimal fitness value at the current iteration number, and Favg represents the average fitness value at the current iteration number;

[0076] Calculate the penalty factor P of the individual through the value of the threshold Thresh(iter), and the specific calculation formula is:

[0077] (7);

[0078] In Equation (7), μ represents the coefficient factor controlling the penalty intensity, F represents the fitness value of the current individual, Thresh represents the dynamic threshold as shown in Equation (6);

[0079] Update the position of the current individual through the obtained penalty factor P, and the specific calculation formula is:

[0080] (8);

[0081] In Equation (8), X(iter + 1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r3 represents a random number between [0, 1], lb represents the lower limit of the search space, ub represents the upper limit of the search space, b represents the adjustment factor controlling the difference in fitness values, F represents the fitness value of the current individual, Fbest represents the optimal fitness value at the current iteration number, and P represents the penalty factor as shown in Equation (7);

[0082] Finally, take reward and punishment measures for the fitness value of the updated individual, and the specific calculation formula is as follows:

[0083] (9);

[0084] In Equation (9), F represents the fitness value after individual update, P represents the penalty factor as shown in Equation (7), Fbest represents the optimal fitness value at the current iteration, and Thresh represents the dynamic threshold, as shown in Equation (6).

[0085] Step 3: Optimize the PID controller in the temperature controller system of the infant incubator using the improved sewing training optimization algorithm, and obtain the optimal set of control parameters Kp, Ki, and Kd through the improved algorithm.

[0086] Furthermore, in the above Step 3, the specific steps for optimizing the PID controller in the temperature controller system of the infant incubator using the improved sewing training optimization algorithm are as follows:

[0087] S1: Initialize the parameters of the improved sewing training optimization algorithm, including the number of population individuals N, the problem dimension dim, the maximum number of iterations max_iter, the upper bound ub of the search space, and the lower bound lb of the search space, and generate the initial positions of the population through the initial parameters.

[0088] S2: Establish a mapping relationship between the improved sewing training optimization algorithm and the PID controller in the temperature controller system of the infant incubator, convert the temperature control of the actual system into a mathematical model to be optimized, and correspond the three dimensions of the individual position vector of the algorithm to the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd of the PID controller, that is, X(x1, x2, x3) = (Kp, Ki, Kd). Through the continuous optimization of the algorithm, the position vector of the individual is continuously updated, thereby dynamically adjusting the PID control parameters, and finally obtaining the optimal set of PID control parameters for the system.

[0089] S3: Construct the fitness value function of the improved sewing training optimization algorithm, calculate the fitness values of the individuals in the population through the fitness value function, and select the optimal individual and the optimal fitness value. The specific fitness value function is as follows:

[0090] (10);

[0091] In Equation (10), J represents the fitness value, L represents the total running time of the system, and e(t) represents the real-time temperature difference.

[0092] S4: Update the individual positions through the mathematics of the improved sewing training optimization algorithm. The specific steps are as follows:

[0093] S41: Execute the training stage of the algorithm. The specific mathematical model is as shown in Equation (3).

[0094] S42. In the stage of imitating the teacher's skills for executing the algorithm, a set of imitation variables is generated according to the number of iterations. The positions of the individuals in the population whose fitness values are less than the fitness value of the current individual are used as the positions of the tutors to generate a tutor set. Each tutor is a three-dimensional vector, and each dimension represents a skill. The updated individual is generated by selecting three variables from the set of imitation variables and combining them with the dimensions in the corresponding tutor set to generate a new position as the updated position;

[0095] S43. In the practical stage of the algorithm, the specific mathematical model is shown in Equations (6)-(8);

[0096] S5. Determine whether the current number of iterations has reached the maximum number of iterations. If not, continue the optimization. If so, output the optimal individual vector in the population as the optimal parameter.

[0097] Step 4: Use the values of the control parameters Kp, Ki, and Kd optimized in Step 3 for the PID controller in the baby incubator temperature controller system, and adjust the control quantity to reach the target temperature.

[0098] Furthermore, the mathematical model of the sewing training optimization algorithm is improved through Matlab, the code design and writing of the experiment are completed, the algorithm parameters are initialized, the maximum number of iterations is 50, the problem dimension is 3, the population size is 100, and the range of the search space is [200, 0.001]. A simulation model of the baby incubator temperature controller system is constructed through the simulation software Simulink, including an instruction operation module, a temperature difference calculation module, a PID controller, an improved sewing training optimization algorithm module, a temperature regulation module, and a temperature sensor module. Among them, the improved sewing training optimization algorithm model is connected to the Matlab code. Through the optimization of the algorithm mathematical model, the obtained parameters are input into Simulink to complete the control of the system. The mathematical model of the temperature regulation module is Laplace-transformed, and the transformed function is used as the controlled function of the model to simulate the actual operating state of the system. The controlled function is: , s is a complex variable. Run the experimental code and the model, and the experimental results are as Figures 3 - 4 shown.

[0099] Furthermore, Figure 3 is a comparison chart of the change of the fitness value during the optimization process of the improved sewing training optimization algorithm and the standard sewing training optimization algorithm. It can be seen from the figure that the optimization accuracy of the improved algorithm is higher and the fitness value of the obtained individuals is lower. Figure 4Comparison chart of the response of the optimized PID controller between the improved sewing training optimization algorithm and the standard sewing training optimization algorithm. The target value is set to 1 unit. It can be seen from the figure that when using the improved algorithm to optimize the PID controller, the overshoot is smaller and the stability is better. The response curve of the PID controller optimized by the improved algorithm reaches the stable state of the target value at a faster speed.

Claims

1. A method for optimizing a temperature controller of an infant incubator, characterized in that: The specific steps are as follows: Step 1: construct a baby incubator temperature controller system, the system comprising: an instruction operation module, a temperature difference calculation module, a PID controller, an improved sewing training optimization algorithm module, a temperature adjustment module, and a temperature sensor module; Step 2: Improve the sewing training optimization algorithm. The specific improvements are as follows: D1. A mathematical model of the training phase of the sewing training optimization algorithm is improved using a multi-position adaptive update strategy, wherein the multi-position adaptive update strategy guides the current population individuals through the optimal position and the average position in the population, wherein the average position of the population is adaptively adjusted through the iterative phase of the algorithm, and the current position is adaptively adjusted according to the optimal position and the average position to update the step size, thereby updating the position; D2. A mathematical model of the practice phase of the sewing training optimization algorithm is improved by using a dynamic threshold penalty update strategy. The dynamic threshold penalty update strategy dynamically generates a threshold according to the current iteration phase and the fitness value of the individual in the population. The penalty factor P of the current individual is calculated by the threshold. The penalty factor P affects the quality of the population position update. If the individual position is closer to the optimal position of the current population, its fitness value is rewarded to retain the position. If the individual position deviates from the optimal position in the current population, its fitness value is penalized and the individual position is updated by the penalty factor P. Step 3: Optimize the PID controller in the temperature controller system of the baby incubator using the improved sewing training optimization algorithm, and obtain the optimal set of control parameters Kp, Ki, and Kd through the improved algorithm optimization; Step 4: Use the values ​​of the control parameters Kp, Ki, and Kd optimized in step 3 for the PID controller in the temperature controller system of the infant incubator, and adjust the control amount to achieve the target temperature.

2. The method for optimizing the temperature controller of an infant incubator according to claim 1, characterized in that: The baby incubator temperature control system constructed in the step 1 has the following specific execution process: first, the target temperature value is set through the instruction operation module, the target temperature value and the temperature monitored in real time by the temperature sensor are input into the temperature difference calculation module to obtain the temperature difference e(t), the temperature difference e(t) is input into the PID controller, and the PID control is optimized by improving the sewing training optimization algorithm module to obtain the control parameters Kp, Ki, and Kd.

3. The method for optimizing the temperature controller of an infant incubator according to claim 2, characterized in that: In D1, a multi-position adaptive update strategy is used to improve the mathematical model of the training phase of the sewing training optimization algorithm. The improved mathematical model is specifically: (3)； In formula (3), X(iter+1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r1 represents a random number between [0,1], Xavg(iter) represents the average position in the population, and the calculation formula is shown in formula (4). Uniform(-1,1) represents the uniform distribution between [-1,1], and Xbest represents the best position in the population. (4)； In formula (4), Xavg(iter+1) represents the updated average position, N represents the number of individuals in the population, X(i) represents the position of the i-th individual, a represents the update constant, K(iter) represents the adaptive attenuation coefficient, and the calculation formula is shown in formula (5). Xbest represents the best position in the population, and Xavg(iter) represents the average position under the current number of iterations; (5)； In formula (5), K(iter) represents the adaptive attenuation coefficient, k0 represents the initial coefficient, λ represents the adjustment factor, which is a random number between [0.1, 0.5], iter represents the current number of iterations, max_iter represents the maximum number of iterations, and randn(0, 1) represents the random number generated by the standard normal distribution.

4. The method for optimizing the temperature controller of an infant incubator according to claim 3, characterized in that: In D2, a dynamic threshold penalty update strategy is used to improve the mathematical model of the practice phase of the sewing training optimization algorithm. First, a dynamic threshold P is generated. The specific mathematical model is: (6)； In formula (6), Thresh(iter) represents the dynamic threshold, t0 represents the initial threshold, iter represents the current number of iterations, max_iter represents the maximum number of iterations, ω represents the threshold change influencing factor, Fbest represents the optimal fitness value under the current number of iterations, and Favg represents the average fitness value under the current number of iterations; The individual penalty factor P is calculated by the value of the threshold Thresh(iter). The specific calculation formula is: (7)； In formula (7), μ represents the coefficient factor controlling the penalty intensity, F represents the fitness value of the current individual, and Thresh represents the dynamic threshold, as shown in formula (6); The position of the current individual is updated by the obtained penalty factor P. The specific calculation formula is: (8)； In formula (8), X(iter+1) represents the updated individual position, iter represents the current iteration number, X(iter) represents the current individual position, r3 represents a random number between [0,1], lb represents the lower limit of the search space, ub represents the upper limit of the search space, b represents the adjustment factor for controlling the difference in fitness values, F represents the fitness value of the current individual, Fbest represents the optimal fitness value under the current iteration number, and P represents the penalty factor as shown in formula (7); Finally, rewards and punishments are taken for the fitness values ​​of the updated individuals. The specific calculation formula is as follows: (9)； In formula (9), F represents the fitness value after individual update, P represents the penalty factor as shown in formula (7), Fbest represents the optimal fitness value under the current number of iterations, and Thresh represents the dynamic threshold as shown in formula (6).

5. The method for optimizing the temperature controller of an infant incubator according to claim 4, characterized in that: In the step 3, the PID controller in the temperature controller system of the infant incubator is optimized using the improved sewing training optimization algorithm, and the specific steps are as follows: S1. Initialize the parameters of the improved sewing training optimization algorithm, including the number of individuals in the population N, the problem dimension dim, the maximum number of iterations max_iter, the upper limit ub of the search space, the lower limit lb of the search space, and generate the initial position of the population through the initial parameters; S2. Establish a mapping relationship between the improved sewing training optimization algorithm and the PID controller in the baby incubator temperature controller system, convert the temperature control of the actual system into a mathematical model to be optimized, and correspond the three dimensions of the individual position vector of the algorithm to the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd of the PID controller, that is, X(x1, x2, x3)=(Kp, Ki, Kd). Through the continuous optimization of the algorithm, the individual position vector is continuously updated, thereby dynamically adjusting the PID control parameters, and finally obtaining an optimal set of PID control parameters for the system; S3. Construct a fitness value function for improving the sewing training optimization algorithm, calculate the fitness value of individuals in the population through the fitness value function, select the best individual and the best fitness value, and the specific fitness value function is: (10)； In formula (10), J represents the fitness value, L represents the total operation time of the system, and e(t) represents the real-time temperature difference; S4. Update the individual positions through the mathematics of the improved sewing training optimization algorithm. The specific steps are: S41, executing the training phase of the algorithm, the specific mathematical model is shown in formula (3); S42, executing the algorithm's imitation teacher skill phase, generating a collection of imitation variables according to the number of iterations, taking the individual position in the population whose individual fitness value is less than the current individual fitness value as the tutor position to generate a tutor set, each tutor is a three-dimensional vector, each dimension represents a skill, and the updated individual is to select three variables from the imitation variable collection and combine them with the dimensions in the corresponding tutor set to generate a new position as the updated position; S43, the practice phase of executing the algorithm, the specific mathematical model is shown in formula (6) to formula (8); S5. Determine whether the current number of iterations has reached the maximum number of iterations. If not, continue to search for the optimal solution. If so, output the optimal individual vector in the population as the optimal parameter.

Images