Programmable pulse power control system for microwave medical equipment

Abstract

The invention relates to the field of instrument control, in particular to a programmable pulse power control system for microwave medical equipment, which comprises a target protocol module, a pulse power compiling module, an organization state identification module and a control strategy module. The target protocol module verifies the programming parameters to generate a high-level treatment target protocol; the pulse power compiling module inputs a protocol into the thermodynamic model, and an initial power reference trajectory and an auxiliary pulse modulation function are generated through a finite element method; the tissue state identification module collects multi-dimensional data, extracts tissue feature vectors after Wilkokson symbol rank test, and outputs tissue state indexes through an identification model; and the control strategy module searches a control strategy through a galaxy group optimization algorithm, adjusts a gravitation coefficient in combination with fuzzy logic, and outputs an optimal control strategy to the programmable microwave pulse generator. Programmable control of microwave power is achieved, dynamic changes of tissue states are adapted, and treatment safety and effectiveness are guaranteed.

Description

Technical Field

[0001] This invention pertains to the field of medical device control, specifically a programmable pulse power control system for microwave medical equipment. Background Technology

[0002] Microwave medical devices, with their advantages of high energy focusing and treatment efficiency, are widely used in clinical scenarios such as tumor ablation. Their core requirement is to achieve effective tissue ablation and safe protection by controlling pulse power. However, existing power control systems for microwave medical devices face numerous technical bottlenecks.

[0003] At the control strategy level, existing algorithms mostly rely on single-objective optimization, which cannot balance tissue ablation effect, temperature safety threshold and impedance matching efficiency. Furthermore, core parameters such as gravity coefficient in the algorithm are usually fixed, resulting in an imbalance between global search and local optimization capabilities. This can easily lead to the risk of incomplete ablation or excessive damage. The verification mechanism for doctor input parameters is not perfect and lacks rapid matching with the clinical safety knowledge base, which may cause treatment safety issues due to parameter violations.

[0004] Therefore, a programmable pulse power control system for microwave medical devices is needed to solve the above problems. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, the present invention provides a programmable pulse power control system for microwave medical devices.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] This invention provides a programmable pulse power control system for microwave medical devices, including a target protocol module, a pulse power compilation module, a tissue state identification module, a control strategy module, and a database.

[0008] The target protocol definition module receives and stores the various programming parameters input by the doctor. If the parameters pass verification, a high-level treatment target protocol is generated. The specific steps are as follows:

[0009] A microwave therapy parameter knowledge base is constructed, specifically the safe parameter ranges corresponding to different tissue types and lesion sizes. Doctors fill in various programming parameters in the graphical interface or script interface, including tissue type, total treatment time, target total energy, maximum safe temperature, minimum safe temperature and maximum temperature rise rate. The programming parameters are matched with the microwave therapy parameter knowledge base through the Rete algorithm to retrieve illegal parameters and output the illegal type and reference threshold.

[0010] Based on the total treatment duration, maximum safe temperature, minimum safe temperature, and maximum temperature rise rate, a target temperature curve is generated. The target temperature curve and input programming parameters are converted into a standard parameter format using a named entity recognition algorithm, and a preset protocol template is embedded to obtain a high-level treatment target protocol, which is then sent to the pulse power compilation module.

[0011] The pulse power compilation module inputs the high-level treatment target protocol into the standard tissue thermodynamic model and solves it using the finite element method to generate the initial power reference trajectory and auxiliary pulse modulation function. The specific steps are as follows:

[0012] Based on the metabolic heat production rate UC, density q, specific heat capacity c, and thermal conductivity e corresponding to the tissue type, a standard tissue thermodynamic model is established. The model construction logic is as follows: ,in Here, UY represents the tissue spatial coordinates, UY represents the microwave power deposition density, and T represents the predicted tissue temperature field. The temperature gradient vector, The divergence operator is used to describe the divergent characteristics of heat flux. The total treatment time is divided into several time steps, and an implicit finite element solver is used to solve the predicted tissue temperature field of the standard tissue thermodynamic model at each time step. The calculation logic is as follows: Where Fb is the biological thermal load vector, and k is the time step number. The length of the time step; the predicted tissue temperature field and target temperature curve based on each time step. The power objective function is established, and its calculation logic is as follows: ,in To organize the computational domain, i.e. the three-dimensional spatial region, HC is the objective function value and Ec is the initial power. The objective function value is minimized by the conjugate gradient method and the initial power at each time step is solved. The initial power of each time step is spliced ​​together to obtain the initial power reference trajectory EQ.

[0013] It should be noted that the duty cycle affects the local distribution and diffusion of heat. A high duty cycle is conducive to heat accumulation and deep conduction, while a low duty cycle (short pulse) is conducive to the diffusion of heat within the pulse interval, reducing the risk of surface overheating. The microwave frequency mainly affects the energy penetration depth and focusing ability. The lower the frequency, the deeper the penetration, and the higher the frequency, the shallower the energy deposition.

[0014] Obtain the rated peak power, and calculate the proportion of the empty function using the rated peak power Es and the initial power reference trajectory. Its calculation logic is as follows: Furthermore, based on the organization's dielectric loss factor and fundamental frequency... Get frequency function Its calculation logic is as follows: ,in To organize the relative permittivity loss factor, As the reference dielectric loss factor, For the average temperature of the tissue, This is the frequency adjustment coefficient; the proportion empty function and the frequency function are integrated into an auxiliary pulse modulation function.

[0015] The tissue state identification module triggers the sensor based on the initial power reference trajectory and the auxiliary pulse modulation function to measure multi-dimensional data in real time. Then, it obtains a multi-dimensional detection dataset by performing a Wilcoxon signed-rank test on the multi-dimensional data. Tissue feature vectors are extracted from the multi-dimensional detection dataset and input into the tissue state identification model to obtain the tissue state index. The specific steps are as follows:

[0016] The sensor is triggered based on the initial power reference trajectory and the auxiliary pulse modulation function. The sensor includes a directional coupler, an RF power meter, a broadband impedance analyzer and an optical fiber temperature sensor. The sensor collects microwave forward power, reflected power, real part impedance detection value, imaginary part impedance detection value and multi-point temperature in real time. The real-time time series data of each parameter is captured by a sliding time window to form a dataset to be tested.

[0017] For each dimension of the dataset to be tested, a Wilcoxon signed-rank test is performed against a preset benchmark. Specifically, the difference between each data point within the window and the benchmark is calculated to obtain the data difference. The absolute values ​​of the data differences, ds, are sorted in ascending order and assigned a rank rc. If there are data points with the same absolute value, the average rank is taken. The signed rank is then obtained based on the ranks of the data differences. Its calculation logic is as follows: Where g is the index of the data difference, and the sum of the positive signed ranks is obtained based on the rank and signed rank. The sum of the negative sign rank and its calculation logic is as follows: , Then, the Wilcoxon signed-rank statistic is calculated using the sum of the positive and negative signed ranks. Its calculation logic is as follows: Extract the preset critical value within the data. If the Wilcoxon signed-rank statistic is less than or equal to the preset critical value, it is determined that there is a statistically significant difference between the current data and the baseline state. The median of the data within the window is taken as the detection data for that dimension. Otherwise, the fluctuation of the current data is determined to be random noise. The detection data is replaced with the preset baseline value. The detection data of each parameter are integrated to generate a multi-dimensional detection dataset.

[0018] Extract the real part of the impedance detection value RC and the imaginary part of the impedance detection value FK from the multi-dimensional detection dataset. Obtain the instantaneous load impedance using the real part and imaginary part of the impedance detection value. The calculation logic is as follows: Based on the instantaneous load impedance KH and the preset power amplifier standard load KZ in the database, the current voltage reflection coefficient magnitude is obtained. Its calculation logic is as follows: The current load matching degree is obtained by using the magnitude of the voltage reflection coefficient. The calculation logic is as follows: The power reflection coefficient is obtained by dividing the reflected power by the corresponding forward power. Then, the instantaneous load impedance, voltage reflection coefficient magnitude, load matching degree and power reflection coefficient are integrated into an organization feature vector.

[0019] The tissue feature vector is input into the tissue state identification model, which incorporates a phase space reconstruction algorithm. The autocorrelation function of the time series of the tissue feature vector is calculated, and the time delay at which the autocorrelation function decreases to a preset initial value is marked as the optimal delay. The time series is then used to construct reconstructed vectors based on the initial embedding dimension. The distances to the nearest neighbors of each reconstructed vector are calculated, the initial embedding dimension is incremented by one, and a new distance is calculated. If the new distance is less than a preset distance, the current embedding dimension is marked as the optimal embedding dimension; otherwise, the initial embedding dimension is incremented by two and compared until it is less than the preset distance. Based on the optimal embedding dimension and optimal delay, a phase space reconstruction matrix of the tissue feature vector is constructed. The covariance matrix of the phase space reconstruction matrix is ​​obtained, and the eigenvalues ​​of the covariance matrix are calculated. The first C principal components are selected to generate a dimension reduction matrix. The tissue state identification model extracts the dynamic features of the tissue state from the dimension reduction matrix through a CNN layer, outputting probability distributions for four types of tissue states: unablated state probability, partially ablated state probability, completely ablated state probability, and excessively damaged state probability. The weighted sum of these probability distributions yields the tissue state index.

[0020] The control strategy module, based on the current tissue state index, predicted tissue temperature field, and instantaneous load impedance, selects the corresponding control strategy path using a galaxy cluster optimization algorithm. Then, a fuzzy logic control algorithm is used to adjust the gravitational coefficient of the galaxy cluster, and the optimal control strategy is sent to the programmable microwave pulse generator. The specific steps are as follows:

[0021] Extract the target tissue state index, target temperature field, and target instantaneous load impedance. Subtract the target tissue state index from the tissue state index to obtain the tissue state deviation. Subtract the target temperature field from the predicted tissue temperature field to obtain the temperature field deviation. Subtract the target instantaneous load impedance from the instantaneous load impedance to obtain the impedance difference. Weighted sum the absolute values ​​of the tissue state deviation, temperature field deviation, and impedance difference to obtain the cost function Ys for the galaxy position.

[0022] Construct a control strategy space, including microwave power, duty cycle, operating frequency, and pulse width. Map each control strategy space to a galaxy position and randomly generate several initial galaxy position vectors. The galaxy mass is calculated using a cost function for each initial galaxy position vector. The calculation logic is as follows: ,in Let j be the maximum value of the cost function, and j be the galaxy number. The total number of galaxies is given; then, the gravitational force between galaxies is calculated using the galaxy mass and the corrected gravitational coefficient DS. Its calculation logic is as follows: ,in Let be the Euclidean distance between galaxies i and j. and The position vectors of different galaxies are used to obtain the galactic acceleration UY based on the gravitational attraction between galaxies. The calculation logic is as follows: The galaxy position is updated iteratively through galaxy acceleration. When the number of iterations is reached, the iteration stops and the current optimal control strategy is output. The optimal control strategy is then sent to the programmable microwave pulse generator.

[0023] A fuzzy rule base is established, and tissue state deviation, temperature field deviation, and impedance difference are combined as fuzzy subsets. The fuzzy rule base is matched with IF-THEN rules to output a correction fuzzy set. This correction fuzzy set is then converted into a gravity coefficient correction value using the centroid method, and embedded with the gravity coefficient to obtain the corrected gravity coefficient.

[0024] Compared with the prior art, the beneficial effects of the present invention are:

[0025] By integrating galaxy cluster optimization algorithms with fuzzy logic control, core parameters such as microwave power and duty cycle are optimized through multi-objective cost functions. The gravity coefficient is dynamically adjusted to balance global search and local optimization capabilities, adapting in real time to dynamic changes in tissue state, temperature field, and impedance, thus absolving incomplete or excessive damage and balancing therapeutic efficacy and safety threshold.

[0026] The Wilcoxon signed-rank test is used to effectively filter data noise and detect the reliability of data from multiple dimensions. The phase space reconstruction algorithm is used to mine the dynamic evolution law of tissue feature vectors, and deep features are extracted by combining CNN layers. The probability distribution and state index of four types of tissue states are output, which solves the problems of lagging state perception and high misjudgment rate in traditional systems. Attached Figure Description

[0027] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. The following drawings are not drawn to scale according to the actual size, but are intended to illustrate the main idea of ​​the present invention.

[0028] Figure 1 This is a module connection diagram of the present invention. Detailed Implementation

[0029] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are also within the scope of protection of the present invention.

[0030] Please refer to Figure 1 As shown, the present invention provides a programmable pulse power control system for microwave medical devices, including a target protocol module, a pulse power compilation module, a tissue state identification module, a control strategy module, and a database.

[0031] The target protocol definition module receives and stores the various programming parameters input by the doctor. If the parameters pass verification, a high-level treatment target protocol is generated. The specific steps are as follows:

[0032] A microwave therapy parameter knowledge base is constructed, specifically outlining the safe parameter ranges corresponding to different tissue types and lesion sizes. Doctors input various programmed parameters into a graphical or script-based interface, including tissue type, total treatment time, target total energy, maximum safe temperature, minimum safe temperature, and maximum temperature rise rate. These programmed parameters are then matched against the microwave therapy parameter knowledge base using the Rete algorithm to retrieve any non-compliant parameters and output the violation type and reference threshold. It should be noted that the Rete algorithm is a rule-based matching algorithm used to quickly retrieve the matching relationship between the doctor's input parameters and the microwave therapy parameter knowledge base, achieving accurate parameter verification.

[0033] Based on the total treatment duration, maximum safe temperature, minimum safe temperature, and maximum temperature rise rate, a target temperature curve is generated. The target temperature curve and input programming parameters are converted into a standard parameter format using a named entity recognition algorithm, and a preset protocol template is embedded to obtain a high-level treatment target protocol, which is then sent to the pulse power compilation module.

[0034] The pulse power compilation module inputs the high-level treatment target protocol into the standard tissue thermodynamic model and solves it using the finite element method to generate the initial power reference trajectory and auxiliary pulse modulation function. The specific steps are as follows:

[0035] Based on the metabolic heat production rate UC, density q, specific heat capacity c, and thermal conductivity e corresponding to the tissue type, a standard tissue thermodynamic model is established. The model construction logic is as follows: ,in Here, UY represents the tissue spatial coordinates, UY represents the microwave power deposition density, and T represents the predicted tissue temperature field. The temperature gradient vector, The divergence operator is used to describe the divergent characteristics of heat flux. The total treatment time is divided into several time steps, and an implicit finite element solver is used to solve the predicted tissue temperature field of the standard tissue thermodynamic model at each time step. The calculation logic is as follows: Where Fb is the biological thermal load vector, and k is the time step number. The length of the time step; the predicted tissue temperature field and target temperature curve based on each time step. The power objective function is established, and its calculation logic is as follows: ,in To organize the computational domain, i.e. the three-dimensional spatial region, HC is the objective function value and Ec is the initial power. The objective function value is minimized by the conjugate gradient method and the initial power at each time step is solved. The initial power of each time step is spliced ​​together to obtain the initial power reference trajectory EQ.

[0036] It should be noted that the duty cycle affects the local distribution and diffusion of heat. A high duty cycle is conducive to heat accumulation and deep conduction, while a low duty cycle (short pulse) is conducive to the diffusion of heat within the pulse interval, reducing the risk of surface overheating. The microwave frequency mainly affects the energy penetration depth and focusing ability. The lower the frequency, the deeper the penetration, and the higher the frequency, the shallower the energy deposition.

[0037] Obtain the rated peak power, and calculate the proportion of the empty function using the rated peak power Es and the initial power reference trajectory. Its calculation logic is as follows: Furthermore, based on the organization's dielectric loss factor and fundamental frequency... Get frequency function Its calculation logic is as follows: ,in To organize the relative permittivity loss factor, As the reference dielectric loss factor, For the average temperature of the tissue, This is the frequency adjustment coefficient; the proportion empty function and the frequency function are integrated into an auxiliary pulse modulation function.

[0038] The tissue state identification module triggers the sensor based on the initial power reference trajectory and the auxiliary pulse modulation function to measure multi-dimensional data in real time. Then, it obtains a multi-dimensional detection dataset by performing a Wilcoxon signed-rank test on the multi-dimensional data. Tissue feature vectors are extracted from the multi-dimensional detection dataset and input into the tissue state identification model to obtain the tissue state index. The specific steps are as follows:

[0039] The sensor is triggered based on the initial power reference trajectory and the auxiliary pulse modulation function. The sensor includes a directional coupler, an RF power meter, a broadband impedance analyzer and an optical fiber temperature sensor. The sensor collects microwave forward power, reflected power, real part impedance detection value, imaginary part impedance detection value and multi-point temperature in real time. The real-time time series data of each parameter is captured by a sliding time window to form a dataset to be tested.

[0040] For each dimension of the dataset to be tested, a Wilcoxon signed-rank test is performed against a preset benchmark. Specifically, the difference between each data point within the window and the benchmark is calculated to obtain the data difference. The absolute values ​​of the data differences, ds, are sorted in ascending order and assigned a rank rc. If there are data points with the same absolute value, the average rank is taken. The signed rank is then obtained based on the ranks of the data differences. Its calculation logic is as follows: Where g is the index of the data difference, it should be noted that the rank refers to the position number of each data point after sorting a set of data in ascending order. Its core function is to replace the absolute value of the original data with a relative order; the sum of the positive signed ranks is obtained based on the ranks and the signed rank. The sum of the negative sign rank and its calculation logic is as follows: , Then, the Wilcoxon signed-rank statistic is calculated using the sum of the positive and negative signed ranks. Its calculation logic is as follows: Extract the preset critical value within the data. If the Wilcoxon signed-rank statistic is less than or equal to the preset critical value, it is determined that there is a statistically significant difference between the current data and the baseline state. The median of the data within the window is taken as the detection data for that dimension. Otherwise, the fluctuation of the current data is determined to be random noise. The detection data is replaced with the preset baseline value. The detection data of each parameter are integrated to generate a multi-dimensional detection dataset.

[0041] Extract the real part of the impedance detection value RC and the imaginary part of the impedance detection value FK from the multi-dimensional detection dataset. Obtain the instantaneous load impedance using the real part and imaginary part of the impedance detection value. The calculation logic is as follows: Based on the instantaneous load impedance KH and the preset power amplifier standard load KZ in the database, the current voltage reflection coefficient magnitude is obtained. Its calculation logic is as follows: The current load matching degree is obtained by using the magnitude of the voltage reflection coefficient. The calculation logic is as follows: The power reflection coefficient is obtained by dividing the reflected power by the corresponding forward power. Then, the instantaneous load impedance, voltage reflection coefficient magnitude, load matching degree and power reflection coefficient are integrated into an organization feature vector.

[0042] The tissue feature vector is input into the tissue state identification model, which incorporates a phase space reconstruction algorithm. The autocorrelation function of the time series of the tissue feature vector is calculated, and the time delay at which the autocorrelation function decreases to a preset initial value is marked as the optimal delay. The time series is then used to construct reconstructed vectors based on the initial embedding dimension. The distances to the nearest neighbors of each reconstructed vector are calculated, the initial embedding dimension is incremented by one, and a new distance is calculated. If the new distance is less than a preset distance, the current embedding dimension is marked as the optimal embedding dimension; otherwise, the initial embedding dimension is incremented by two and compared until it is less than the preset distance. Based on the optimal embedding dimension and optimal delay, a phase space reconstruction matrix of the tissue feature vector is constructed. The covariance matrix of the phase space reconstruction matrix is ​​obtained, and the eigenvalues ​​of the covariance matrix are calculated. The first C principal components are selected to generate a dimension reduction matrix. The tissue state identification model extracts the dynamic features of the tissue state from the dimension reduction matrix through a CNN layer, outputting probability distributions for four types of tissue states: unablated state probability, partially ablated state probability, completely ablated state probability, and excessively damaged state probability. The weighted sum of these probability distributions yields the tissue state index.

[0043] It should be noted that the time delay represents the time interval between adjacent vectors during reconstruction, which is used to avoid feature redundancy. The preset initial value is the value of the autocorrelation function when the time delay is 0. The embedding dimension represents the dimension of the reconstructed phase space, ensuring that the dynamic trajectory of the organizational state can be fully depicted.

[0044] The control strategy module, based on the current tissue state index, predicted tissue temperature field, and instantaneous load impedance, selects the corresponding control strategy path using a galaxy cluster optimization algorithm. Then, a fuzzy logic control algorithm is used to adjust the gravitational coefficient of the galaxy cluster, and the optimal control strategy is sent to the programmable microwave pulse generator. The specific steps are as follows:

[0045] The fuzzy logic control algorithm solves the problem of unbalanced exploration capabilities caused by fixed gravity coefficients in the galaxy cluster optimization algorithm (such as the need for global exploration of optimal power in the early stage of ablation and the need for local development of stable strategies in the later stage of ablation). It dynamically adjusts the galaxy cluster optimization parameters through fuzzy logic to adapt to different organizational states.

[0046] Extract the target tissue state index, target temperature field, and target instantaneous load impedance. Subtract the target tissue state index from the tissue state index to obtain the tissue state deviation. Subtract the target temperature field from the predicted tissue temperature field to obtain the temperature field deviation. Subtract the target instantaneous load impedance from the instantaneous load impedance to obtain the impedance difference. Weighted sum the absolute values ​​of the tissue state deviation, temperature field deviation, and impedance difference to obtain the cost function Ys for the galaxy position.

[0047] Construct a control strategy space, including microwave power, duty cycle, operating frequency, and pulse width. Map each control strategy space to a galaxy position and randomly generate several initial galaxy position vectors. The galaxy mass is calculated using a cost function for each initial galaxy position vector. The calculation logic is as follows: ,in Let j be the maximum value of the cost function, and j be the galaxy number. The total number of galaxies is given; then, the gravitational force between galaxies is calculated using the galaxy mass and the corrected gravitational coefficient DS. Its calculation logic is as follows: ,in Let be the Euclidean distance between galaxies i and j. and The position vectors of different galaxies are used to obtain the galactic acceleration UY based on the gravitational attraction between galaxies. The calculation logic is as follows: The galaxy position is updated iteratively through galaxy acceleration. When the number of iterations is reached, the iteration stops and the current optimal control strategy is output. The optimal control strategy is then sent to the programmable microwave pulse generator.

[0048] A fuzzy rule base is established, and the tissue state deviation, temperature field deviation, and impedance difference are combined as fuzzy subsets. The fuzzy rule base is matched with the IF-THEN rule to output the correction fuzzy set. For example, if the tissue state deviation is positive AND the temperature field deviation is positive AND the impedance difference is positive THEN correction amount is significantly increased, the correction fuzzy set is converted into the gravity coefficient correction amount through the centroid method, and the correction gravity coefficient is obtained by embedding the gravity coefficient.

[0049] All formulas involved in this invention are dimensionless numerical calculations. Dimensionlessness can be achieved using conventional techniques such as standardization, which will not be elaborated upon here. Each formula is obtained through software simulation and fitting based on a large amount of measured data, and can best fit the actual operating scenarios of the programmable microwave pulse power control system of this invention. The preset parameters in the formulas can be flexibly set by those skilled in the art according to specific treatment needs, equipment conditions, and clinical scenarios.

[0050] The embodiments of the present invention can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented in software form, it can be entirely or partially embodied as a computer program product. This computer program product includes one or more computer instructions or computer programs. When these computer instructions or computer programs are loaded and executed on a computer, they fully or partially implement the closed-loop control flow and core functions described in the embodiments of the present invention, covering the entire chain of logic including programming parameter verification, power trajectory generation, organizational state identification, control strategy optimization, and pulse generator driving. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device, and must be adapted to the collaborative operation logic of the various modules of the present invention (target protocol module, pulse power compilation module, etc.).

[0051] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, from one website, computer, server, or data center to another via wired (e.g., cable) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium accessible to a computer, or a data storage device such as a server or data center containing one or more sets of available media. Available media include magnetic media (e.g., floppy disks, ATA hard drives, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state ATA hard drives).

[0052] It should be understood that in the various embodiments of the present invention, the sequence number of each process does not represent the order of execution. The execution order should be based on its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0053] Those skilled in the art will recognize that the modules and algorithm steps described in conjunction with the embodiments of the present invention can be implemented through electronic hardware, or a combination of computer software and electronic hardware. The implementation of these functions depends on the specific application and design constraints of the technical solution. Skilled personnel can adopt different implementation methods for different application scenarios, but such implementations should not exceed the scope of protection of this invention.

[0054] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other forms. For example, the apparatus embodiments are merely illustrative, and the division of modules is only a logical functional division. In actual implementation, other division methods may exist, such as multiple modules or components being combined or integrated into another system, or some features being ignored or not executed. Furthermore, the mutual coupling, direct coupling, or communication connection shown or discussed can be implemented through interfaces, and the indirect coupling or communication connection of apparatuses or modules can be electrical, mechanical, or other forms.

[0055] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0056] In addition, the functional modules in the various embodiments of the present invention can be integrated into a processing unit, or each module can exist physically separately, or two or more modules can be integrated into a unit.

[0057] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied as a software product. This computer software product is stored in a storage medium and contains several instructions to cause a computer device (such as a personal computer, server, network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present invention. The storage medium includes media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0058] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A programmable pulse power control system for microwave medical equipment, comprising a target protocol module, a pulse power compilation module, a tissue state identification module, a control strategy module, and a database, characterized in that: The target protocol definition module is used to receive and store the various programming parameters input by the doctor. If the verification is successful, a high-level treatment target protocol is generated. The pulse power compilation module inputs the high-level treatment target protocol into the standard tissue thermodynamic model and solves it using the finite element method to generate the initial power reference trajectory and auxiliary pulse modulation function; The tissue state identification module triggers the sensor based on the initial power reference trajectory and the auxiliary pulse modulation function to measure multi-dimensional data in real time. Then, it obtains a multi-dimensional detection dataset by performing a Wilcoxon signed-rank test on the multi-dimensional data. The tissue feature vector is extracted from the multi-dimensional detection dataset and input into the tissue state identification model to obtain the tissue state index. The control strategy module selects the corresponding control strategy path based on the current tissue state index, predicted tissue temperature field and instantaneous load impedance through the galaxy cluster optimization algorithm, and then adjusts the gravitational coefficient of the galaxy cluster using the fuzzy logic control algorithm, and sends the optimal control strategy to the programmable microwave pulse generator.

2. The programmable pulse power control system for microwave medical equipment according to claim 1, characterized in that, The tissue state identification module triggers the sensor based on the initial power reference trajectory and the auxiliary pulse modulation function, measures multi-dimensional data in real time, and then obtains a multi-dimensional detection dataset through the Wilcoxon signed-rank test. The specific steps are as follows: The sensor is triggered based on the initial power reference trajectory and the auxiliary pulse modulation function. The sensor includes a directional coupler, an RF power meter, a broadband impedance analyzer, and an optical fiber temperature sensor. The sensor collects microwave forward power, reflected power, real part impedance detection value, imaginary part impedance detection value, and multi-point temperature in real time. The real-time time series data of each parameter is captured by a sliding time window to form a dataset to be tested. For each dimension of the dataset to be tested, a Wilcoxon signed-rank test is performed against a preset benchmark. Specifically, the difference between each data point within the window and the benchmark is calculated to obtain the data difference. The absolute values ​​of the data differences, ds, are sorted in ascending order and assigned ranks. If there are data points with the same absolute value, the average rank is taken. The signed rank is obtained based on the ranks of the data differences, and the sum of the positive signed ranks is obtained based on the ranks and the signed rank. The sum of the negative sign rank and the sum of the positive sign rank and the negative sign rank are used to calculate the Wilcoxon sign rank statistic. A preset critical value is extracted from the data. If the Wilcoxon sign rank statistic is less than or equal to the preset critical value, it is determined that there is a statistically significant difference between the current data and the baseline state. The median of the data in the window is taken as the detection data for this dimension. Otherwise, the fluctuation of the current data is determined to be random noise, and the detection data is replaced with the preset baseline value. The detection data of each parameter are integrated to generate a multi-dimensional detection dataset.

3. A programmable pulse power control system for microwave medical equipment according to claim 2, characterized in that, The organization state identification module extracts organization feature vectors from the multi-dimensional detection dataset and inputs them into the organization state identification model to obtain the organization state index. The specific process is as follows: Extract the real and imaginary impedance detection values ​​from the multi-dimensional detection dataset. Obtain the instantaneous load impedance using the real and imaginary impedance detection values. Based on the instantaneous load impedance and the preset power amplifier standard load in the database, obtain the current voltage reflection coefficient magnitude. Obtain the current load matching degree using the voltage reflection coefficient magnitude. Divide the reflected power by the corresponding forward power to obtain the power reflection coefficient. Finally, integrate the instantaneous load impedance, voltage reflection coefficient magnitude, load matching degree, and power reflection coefficient into an organized feature vector. The tissue feature vector is input into the tissue state identification model, which includes a phase space reconstruction algorithm. The autocorrelation function of the time series of the tissue feature vector is calculated. The time delay when the autocorrelation function decreases to a preset initial value is marked as the optimal delay. The time series is reconstructed into a reconstruction vector based on the initial embedding dimension. The distance between the nearest neighbors of each reconstruction vector is calculated. The initial embedding dimension is incremented by one, and the new distance is calculated. If the new distance is less than the preset distance, the current embedding dimension is marked as the optimal embedding dimension; otherwise, the initial embedding dimension is incremented by two, and the result is compared until it is less than the preset distance. Based on the optimal embedding dimension and the optimal delay, a phase space reconstruction matrix of the tissue feature vector is constructed. The covariance matrix of the phase space reconstruction matrix is ​​obtained, and the eigenvalues ​​of the covariance matrix are calculated. The first C principal components are selected to generate a dimension reduction matrix. The tissue state identification model extracts the dynamic features of tissue state from the dimensionality-reduced matrix through the CNN layer and outputs the probability distribution of four types of tissue state, including the probability of unablated state, the probability of partially ablated state, the probability of completely ablated state, and the probability of excessive damage state. The tissue state index is obtained by weighted summation of each probability distribution.

4. A programmable pulse power control system for microwave medical equipment according to claim 1, characterized in that, The control strategy module, based on the current tissue state index, predicted tissue temperature field, and instantaneous load impedance, selects the corresponding control strategy path using a galaxy cluster optimization algorithm and sends the optimal control strategy to the programmable microwave pulse generator. The specific steps are as follows: Extract the target tissue state index, target temperature field, and target instantaneous load impedance. Subtract the target tissue state index from the tissue state index to obtain the tissue state deviation. Subtract the target temperature field from the predicted tissue temperature field to obtain the temperature field deviation. Subtract the target instantaneous load impedance from the instantaneous load impedance to obtain the impedance difference. Weighted sum the absolute values ​​of the tissue state deviation, temperature field deviation, and impedance difference to obtain the cost function for the galaxy position. A control strategy space is constructed, including microwave power, duty cycle, operating frequency, and pulse width. Each control strategy space is mapped to galaxy positions, and several initial galaxy position vectors are randomly generated. The galaxy mass of each initial galaxy position vector is calculated using a cost function. Then, the gravitational force between galaxies is calculated using the galaxy mass and the corrected gravitational coefficient. Based on the gravitational force between galaxies, the galaxy acceleration is obtained. The galaxy positions are iteratively updated using the galaxy acceleration. When the number of iterations is reached, the iteration stops and the current optimal control strategy is output. The optimal control strategy is then sent to the programmable microwave pulse generator.

5. A programmable pulse power control system for microwave medical equipment according to claim 4, characterized in that, The control strategy module adjusts the gravitational coefficient of the galaxy cluster using a fuzzy logic control algorithm. The specific steps are as follows: A fuzzy rule base is established, and the tissue state deviation, temperature field deviation and impedance difference are combined as fuzzy subsets. The fuzzy rule base is matched with the IF-THEN rule to output the correction fuzzy set. The correction fuzzy set is converted into the gravity coefficient correction amount by the centroid method, and the correction gravity coefficient is obtained by embedding the gravity coefficient.

6. A programmable pulse power control system for microwave medical equipment according to claim 1, characterized in that, The target protocol definition module is used to receive and store various programming parameters input by the doctor. If the parameters pass the verification, a high-level treatment target protocol is generated. The specific steps are as follows: A microwave therapy parameter knowledge base is constructed, specifically the safe parameter ranges corresponding to different tissue types and lesion sizes. Doctors fill in various programming parameters in the graphical interface or script interface, including tissue type, total treatment time, target total energy, maximum safe temperature, minimum safe temperature and maximum temperature rise rate. The programming parameters are matched with the microwave therapy parameter knowledge base through the Rete algorithm to retrieve illegal parameters and output the illegal type and reference threshold. Based on the total treatment duration, maximum safe temperature, minimum safe temperature, and maximum temperature rise rate, a target temperature curve is generated. The target temperature curve and input programming parameters are converted into a standard parameter format using a named entity recognition algorithm, and a preset protocol template is embedded to obtain a high-level treatment target protocol, which is then sent to the pulse power compilation module.

7. A programmable pulse power control system for microwave medical equipment according to claim 1, characterized in that, The pulse power compilation module inputs the high-level treatment target protocol into the standard tissue thermodynamic model and solves it using the finite element method to generate an initial power reference trajectory. The specific steps are as follows: Based on the metabolic heat production rate, density, specific heat capacity, and thermal conductivity corresponding to the tissue type, a standard tissue thermodynamic model is established. The total treatment time is divided into several time steps. An implicit finite element solver is used to solve the predicted tissue temperature field of each time step of the standard tissue thermodynamic model. Based on the predicted tissue temperature field and target temperature curve of each time step, a power objective function is established. The objective function value is minimized by the conjugate gradient method, and the initial power of each time step is solved. The initial power of each time step is spliced ​​to obtain the initial power reference trajectory.

8. A programmable pulse power control system for microwave medical equipment according to claim 7, characterized in that, The pulse power compilation module obtains the auxiliary pulse modulation function, and the specific steps are as follows: The rated peak power is obtained, and the proportion space function is calculated using the rated peak power and the initial power reference trajectory. Then, the frequency function is obtained based on the tissue dielectric loss factor and the fundamental frequency. The proportion space function and the frequency function are integrated into an auxiliary pulse modulation function.

Images