A control system for an intense pulsed light therapy device

Abstract

The present invention discloses a control system for an intense pulsed light therapy device, which relates to the technical field of therapy device control. The present invention includes a dynamic spectral energy adaptation module, a three-dimensional contact pressure field calibration module, a domain gradient temperature management module, and a multimodal safety interlock module. Through a multispectral sensor and a convolutional neural network, the skin absorption coefficient is dynamically calculated and the pulse light parameters are optimized to achieve precise energy adaptation; the contact pressure field uniformity is calibrated using a pressure-sensitive matrix and an ANFIS network combined with a robotic arm adjustment; with the help of an annular temperature sensor array and a finite element model, regional temperature monitoring and dynamic adjustment are achieved; and through a multi-level interlocking mechanism and a multi-dimensional scoring model, the safety and comfort of the treatment process are guaranteed. The present invention significantly improves the therapeutic effect, personalization level, and intelligence level of the intense pulsed light therapy device, providing a more efficient and safe solution for clinical applications.

Description

Technical Field

[0001] The present invention relates to the technical field of therapeutic apparatus control, in particular to a control system of an intense pulsed light therapeutic apparatus. Background Art

[0002] Intense pulsed light (IPL) therapy devices are a common medical cosmetic device widely used in skin treatments such as hair removal, skin rejuvenation, and treatment of vascular lesions. However, traditional IPL devices present several challenges during use. For one thing, their energy output is typically fixed and cannot be precisely adapted to the patient's specific skin conditions (such as skin tone, skin type, and treatment area), which can easily lead to suboptimal treatment results and even skin damage. Furthermore, the contact pressure between the treatment head and the skin is difficult to distribute evenly. Uneven contact can affect the effective transmission of light energy, reducing treatment efficiency. Excessive or insufficient local pressure can also affect the safety and comfort of treatment. Furthermore, the heat generated during treatment needs to be strictly controlled to avoid thermal damage, but existing devices often lack precise temperature management, making it impossible to monitor and dynamically adjust the temperature of each area in real time.

[0003] In terms of control technology, the control system functions of traditional intense pulsed light therapy devices are relatively simple and lack an intelligent multi-module collaborative control mechanism. For example, energy regulation mainly relies on manual settings or simple preset modes, and cannot be dynamically optimized based on skin feedback data in real time. In terms of contact pressure calibration, there is a lack of effective real-time monitoring and automatic adjustment methods. The only way to improve the contact effect is to manually adjust the position of the treatment head, but this method is inefficient and difficult to achieve ideal uniformity. For temperature management, although some devices are equipped with cooling systems, they cannot achieve precise temperature control in different areas, and cannot dynamically adjust energy output and cooling intensity according to the actual temperature conditions in each area. These shortcomings limit the treatment effect, safety and comfort of intense pulsed light therapy devices, and cannot meet the growing demand for personalized medical care.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of the present invention is to solve the existing problems and to provide a control system for an intense pulsed light therapy device.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A control system for an intense pulsed light therapy device, comprising:

[0008] Dynamic spectral energy adaptation module, used to calculate the absorption coefficient of each layer of skin, simulate energy deposition, optimize the energy distribution of the target and surrounding tissues through the objective function, and adapt the spectral energy;

[0009] The 3D contact pressure field calibration module is used to input pressure-sensitive data into the ANFIS network. With the goal of minimizing the pressure standard deviation, it combines the robot arm joint weight strategy to calibrate the contact pressure field uniformity.

[0010] The regional gradient temperature management module is used to monitor the temperature gradient by region, dynamically adjust energy to control the temperature of each region, balance the therapeutic thermal effect and tissue protection, and achieve precise temperature control;

[0011] The multimodal safety interlock module is used to achieve multi-dimensional safety monitoring and adaptive decision-making during the treatment process through comfort score constraints.

[0012] Furthermore, the execution process of the dynamic spectrum energy adaptation module is as follows:

[0013] A multispectral sensor array is integrated into the treatment head to collect the reflected spectrum data of the treatment area in real time, and simultaneously activate the skin impedance detection electrode to obtain the epidermal electrical impedance parameters;

[0014] The convolutional neural network is used to extract features from spectral data, identify the characteristic values ​​of skin melanin content, hemoglobin distribution and stratum corneum thickness, and calculate the comprehensive absorption coefficient of the skin by combining impedance parameters;

[0015] A pulse energy dynamic adaptation model is established, and the absorption coefficient is input into a preset Bayesian optimization algorithm to output the optimal pulse width, spectral band, and energy density;

[0016] The DSP chip is controlled to generate a dynamic modulation waveform, driving the semiconductor laser array to output according to the optimized parameters. At the same time, closed-loop monitoring is started. When the real-time reflectivity change exceeds the threshold of ±5%, parameter re-optimization is triggered.

[0017] Skin multi-layer optical modeling and Monte Carlo photon transport simulation accurately quantify the spatial distribution of light energy in skin tissue through mathematical and physical models, and dynamically optimize pulse parameters based on real-time feedback data.

[0018] Furthermore, the specific operation steps of the skin multi-layer optical modeling and Monte Carlo photon transmission simulation in the dynamic spectral energy adaptation module are as follows:

[0019] The skin is defined as a three-layer structure, and the optical properties of each layer are determined by the absorption coefficient , scattering coefficient and anisotropy factor g; integrating multispectral sensor and impedance detection data, dynamically calculating the parameters of each layer:

[0020] in, For skin tissue at wavelength The total absorption coefficient under For melanin at wavelength The total absorption coefficient under For hemoglobin at wavelength The absorption coefficient under The concentration of melanin in the skin and the concentration of hemoglobin in the skin are extracted by the convolutional neural network. This is the basic absorption value of the skin;

[0021] The propagation path of photons in the skin is simulated based on random walk theory, and the energy deposition distribution is calculated:

[0022] ,in In space coordinates and the energy deposition distribution at time t, is the photon weight, are the absorption coefficient and scattering coefficient respectively, s is the path length, is the total number of simulated photons;

[0023] The energy distribution output by the Monte Carlo simulation is combined with the Bayesian optimization algorithm to construct the objective function:

[0024] ,in, They are total energy output, energy in the target area, energy in the surrounding tissue, and the safe energy threshold for the surrounding tissue. Optimization variables include pulse width, spectral band combination, and energy density to ensure energy focusing in the target area while suppressing the risk of overheating in the surrounding tissue.

[0025] In closed-loop monitoring, the Monte Carlo model is dynamically modified using reflectivity change data. The error threshold is set to ±2% to improve the matching accuracy between the model and the patient's individual skin.

[0026] Furthermore, the specific operation steps of the three-dimensional contact pressure field calibration module are as follows:

[0027] A 64-point piezoresistive matrix is ​​arranged on the base of the treatment head to acquire contact pressure distribution data at a sampling rate of 100 Hz;

[0028] A three-dimensional pressure field model is constructed through spatial interpolation algorithm to calculate the pressure gradient distribution and contact area ratio;

[0029] When the maximum pressure difference is greater than 20kPa or the effective contact area is less than 85%, the six-axis robot arm is activated to adjust the posture: first, the Z-axis fine-tunes the downward pressure to the target value, then the position deviation is compensated by the XY-axis translation, and finally the joint is rotated to adjust the incident angle;

[0030] Recalibrate the pressure field uniformity after each adjustment until the preset conditions of pressure standard deviation <3kPa and contact area >95% are met;

[0031] By using multi-scale contact mechanics modeling and dynamic fuzzy decision-making mechanism, the contact state between the treatment head and the skin is optimized through the collaborative use of physical models and intelligent algorithms, significantly improving the pressure field uniformity and calibration efficiency.

[0032] Furthermore, the specific operation steps of the multi-scale contact mechanics modeling and dynamic fuzzy decision-making mechanism are as follows:

[0033] Based on the Hertz contact theory, a stress distribution model of the skin-treatment head contact area is established, and the equivalent curvature radius of the contact surface is defined. and elastic modulus : ,in are the equivalent curvature radius of the contact area between the skin and the treatment head, the curvature radius of the treatment head, and the curvature radius of the skin surface, respectively. The material parameters of the treatment head are the elastic modulus and Poisson's ratio of the treatment head material. are the skin elasticity parameters, namely the elastic modulus and Poisson's ratio of the skin;

[0034] The 64-point pressure-sensitive matrix data is input into the ANFIS network, and the training input is the pressure distribution , contact area A and equivalent curvature radius , the output is the optimal adjustment vector ,in They are Z-axis displacement, X-axis translation, Y-axis translation and rotation angle respectively;

[0035] The network objective function is to minimize the pressure standard deviation: , where N is the number of pressure-sensitive matrix data points, is the pressure value of the i-th point, is the average pressure; constraints: , the maximum pressure difference in the pressure distribution ,ANFIS updates the fuzzy rule base online by hybrid gradient descent and least squares algorithm to achieve adaptive decision-making;

[0036] A dynamic weight distribution strategy is introduced in the robot arm posture adjustment to adjust the priority of each joint movement according to the direction of the pressure gradient:

[0037] Z-axis downforce: Weight , give priority to eliminating local high-pressure areas;

[0038] XY Translation: Weight , is the pressure gradient, the larger the gradient, the higher the translation compensation weight;

[0039] Revolute Joints: Weights , is the pressure distribution skewness, used to correct asymmetric contact;

[0040] Define the contact quality index CQI to comprehensively evaluate the pressure field performance: , target value ≥ 0.97;

[0041] When the CQI does not meet the standard, a secondary optimization cycle is triggered until the preset conditions are met.

[0042] Furthermore, the specific operation steps of the domain gradient temperature management module are as follows:

[0043] A ring-shaped temperature sensor array is arranged inside the treatment head to monitor the temperature of each area at 50ms intervals;

[0044] Establish a three-dimensional heat conduction finite element model to predict the epidermal temperature distribution and dermal heat diffusion trend in real time;

[0045] When the instantaneous temperature in a certain area exceeds 42°C, a graded response is initiated:

[0046] The first-level response is to reduce the output energy of the area by 30% at 42-45℃, and simultaneously turn on the semiconductor cooling plate at the corresponding position;

[0047] Secondary response: when the temperature is greater than 45°C, treatment is suspended and the liquid cooling circulation system is activated;

[0048] After the temperature recovers, the gradient restart strategy is implemented, giving priority to resuming treatment in the low-temperature area, and setting an energy increment gradient for every 5°C temperature difference until the entire area reaches a balanced state.

[0049] Furthermore, the specific operation steps of the multi-modal safety interlock module are as follows:

[0050] Establish a data bus between modules to exchange energy parameters, pressure distribution and temperature field data in real time;

[0051] When any module detects an abnormality, a multi-level interlock response is triggered: Level 1 alarm means a single parameter exceeds the limit, parameter compensation is initiated, Level 2 alarm means two modules are abnormal at the same time, treatment is suspended and a fault code is fed back, Level 3 alarm means system-level fault, the main power is cut off and the treatment head is ejected;

[0052] After the abnormality is resolved, the system self-check process is executed: the calibration of the optical components, the linearity of the pressure sensor and the performance of the cooling system are checked in sequence;

[0053] Generate a treatment quality assessment report, including energy adaptability, contact uniformity coefficient, and thermal injury risk index, and store it in the medical database for subsequent treatment reference;

[0054] By establishing a multi-dimensional scoring model and decision tree analysis algorithm, we can achieve objective evaluation of the comfort of a single treatment and automatic tracing of abnormal causes.

[0055] Furthermore, the specific steps for establishing a multi-dimensional scoring model and a decision tree analysis algorithm in the multimodal safety interlock module are as follows:

[0056] Based on the three core parameters of the quality assessment report, including energy fitness , contact uniformity coefficient , heat injury risk index , construct a weighted scoring function: ,in Score the treatment comfort. They are energy adaptability weight, contact uniformity coefficient weight and thermal damage risk index weight, with values ​​of 0.4, 0.3 and 0.3 respectively. The scoring range is Between 0-100, when When the trigger is triggered, cause analysis;

[0057] Automatically locate the source of the problem using the following rules:

[0058] when , then it is judged that the energy adaptation deviation is too large; when , and the pressure standard deviation is greater than 3kPa, it is judged that the contact pressure is uneven. , and the pressure standard deviation is less than 3kPa, then the treatment head posture deviation is determined; when When the maximum temperature is greater than 15℃, the cooling response is judged to be delayed. When the maximum temperature is less than 15°C, it is judged that the area of ​​insufficient blood perfusion is not identified;

[0059] Match the preset optimization strategy library based on the diagnosis results;

[0060] The scores, reasons, and recommendations are appended to the assessment report, and two types of storage tags are generated:

[0061] Patient tags: record individual comfort trends for personalized parameter settings; device tags: count the causes of high-frequency abnormalities to guide hardware or algorithm iterations.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] (1) The present invention, through a dynamic spectral energy adaptation module, can calculate and optimize the energy parameters of pulsed light, including pulse width, spectral band, and energy density, in real time based on the patient's skin type, melanin content, hemoglobin distribution, and stratum corneum thickness. This precise energy adaptation ensures the effective deposition of light energy in the target area (such as hair follicles or diseased tissue), while minimizing thermal damage to surrounding normal tissues, thereby significantly improving the treatment effect. Through multi-layer optical modeling of the skin and Monte Carlo photon transmission simulation, combined with real-time feedback data to dynamically optimize pulse parameters, the level of treatment personalization is further improved, which can better meet the treatment needs of different patients;

[0064] (2) In the present invention, the three-dimensional contact pressure field calibration module and the domain gradient temperature management module jointly ensure the safety and comfort of the treatment process. The three-dimensional contact pressure field calibration module, through the pressure-sensitive matrix and ANFIS network, combined with the dynamic adjustment strategy of the robotic arm, can quickly calibrate the uniformity of the contact pressure field, ensure good contact between the treatment head and the skin, and avoid uneven light energy transmission and skin damage caused by uneven contact. The domain gradient temperature management module monitors the temperature changes of each area in real time and dynamically adjusts the energy output and cooling intensity according to the temperature conditions, thereby achieving precise temperature control and effectively preventing the occurrence of thermal damage. In addition, the multimodal safety interlock module conducts all-round safety monitoring of the treatment process by establishing an inter-module data bus and a multi-level interlock response mechanism, and takes timely measures in abnormal situations, further enhancing the safety of the treatment;

[0065] (3) The present invention uses a multi-dimensional scoring model and decision tree analysis algorithm to objectively evaluate the treatment process and trace the cause of abnormalities, automatically locating the source of the problem and matching the optimization strategy. This intelligent evaluation and optimization mechanism not only improves the adaptive ability of the equipment, but also provides reference and optimization suggestions for subsequent treatment. The treatment quality assessment report automatically generated by the system contains key parameters such as energy adaptability, contact uniformity coefficient, and thermal damage risk index, providing clinicians with detailed treatment data support, which helps to further optimize the treatment plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0067] Figure 1 This is the overall system block diagram of the present invention. DETAILED DESCRIPTION

[0068] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0069] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0070] It should also be understood that the terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" as used in this disclosure and the claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0071] like Figure 1 As shown, a control system of an intense pulsed light therapy device includes a dynamic spectrum energy adaptation module, a three-dimensional contact pressure field calibration module, a domain gradient temperature management module and a multi-modal safety interlock module;

[0072] The dynamic spectral energy adaptation module calculates the absorption coefficient of each layer of the skin, simulates energy deposition, optimizes the energy distribution of the target and surrounding tissues through the objective function, and adapts the spectral energy;

[0073] A multispectral sensor array is integrated into the treatment head to collect real-time reflectance spectral data of the treatment area and simultaneously activate the skin impedance detection electrodes to obtain epidermal electrical impedance parameters. A convolutional neural network is used to extract features from the spectral data, identify characteristic values ​​of skin melanin content, hemoglobin distribution, and stratum corneum thickness, and calculate the skin's comprehensive absorption coefficient based on the impedance parameters.

[0074] A dynamic pulse energy adaptation model is established, and the absorption coefficient is input into a preset Bayesian optimization algorithm to output the optimal pulse width (adjustable from 1 to 20 ms), spectral band (400-1200 nm segmented combination), and energy density (5-50 J / cm²). The DSP chip is controlled to generate a dynamic modulation waveform, driving the semiconductor laser array to output according to the optimized parameters. Closed-loop monitoring is also initiated, and parameter re-optimization is triggered when the real-time reflectivity change exceeds the threshold of ±5%.

[0075] Skin multi-layer optical modeling and Monte Carlo photon transport simulation accurately quantify the spatial distribution of light energy within skin tissue through mathematical and physical models, and dynamically optimize pulse parameters based on real-time feedback data. The specific steps are as follows:

[0076] The skin is defined as a three-layer structure (epidermis, dermis, and subcutaneous tissue), and the optical properties of each layer are determined by the absorption coefficient , scattering coefficient The parameters of each layer are dynamically calculated by integrating multispectral sensor and impedance detection data:

[0077] in, For skin tissue at wavelength The total absorption coefficient under For melanin at wavelength The total absorption coefficient under For hemoglobin at wavelength The absorption coefficient under The concentration of melanin in the skin and the concentration of hemoglobin in the skin are extracted by the convolutional neural network. This is the basic absorption value of the skin;

[0078] The propagation path of photons in the skin is simulated based on random walk theory, and the energy deposition distribution is calculated:

[0079] ,in In space coordinates and the energy deposition distribution at time t, is the photon weight, are the absorption coefficient and scattering coefficient respectively, s is the path length, is the total number of simulated photons (≥10 6 );

[0080] The energy distribution output by the Monte Carlo simulation is combined with the Bayesian optimization algorithm to construct the objective function:

[0081] ,in, They are total energy output, energy in the target area (such as hair follicles), energy in the surrounding tissues, and the safety energy threshold of the surrounding tissues; the optimization variables include pulse width, spectral band combination, and energy density to ensure energy focusing in the target area (such as hair follicles) while suppressing the risk of overheating of the surrounding tissues; in closed-loop monitoring, the reflectivity change data is used to dynamically correct the Monte Carlo model The error threshold is set to ±2% to improve the matching accuracy between the model and the patient's individual skin.

[0082] The 3D contact pressure field calibration module inputs pressure-sensitive data into the ANFIS network, and calibrates the uniformity of the contact pressure field by combining the robot arm joint weight strategy with the goal of minimizing the pressure standard deviation.

[0083] A 64-point piezoresistive matrix is ​​placed at the base of the treatment head, and contact pressure distribution data is acquired at a 100Hz sampling rate. A three-dimensional pressure field model is constructed using a spatial interpolation algorithm to calculate the pressure gradient distribution and contact area ratio. When the maximum pressure difference is greater than 20kPa or the effective contact area is less than 85%, the six-axis robotic arm is activated to adjust the position: first, the Z-axis fine-tunes the downward pressure to the target value (15±2N), then the XY-axis translation compensates for position deviation, and finally, the joint is rotated to adjust the incident angle (adjustable from 0-30°). After each adjustment, the pressure field uniformity is rechecked until the preset conditions of pressure standard deviation less than 3kPa and contact area greater than 95% are met.

[0084] By using multi-scale contact mechanics modeling and dynamic fuzzy decision-making mechanisms, the contact state between the treatment head and the skin is optimized through the synergistic combination of physical models and intelligent algorithms, significantly improving the uniformity of the pressure field and calibration efficiency. The specific steps are as follows:

[0085] Based on the Hertz contact theory, a stress distribution model of the skin-treatment head contact area is established, and the equivalent curvature radius of the contact surface is defined. and elastic modulus : ,in are the equivalent curvature radius of the contact area between the skin and the treatment head, the curvature radius of the treatment head, and the curvature radius of the skin surface, respectively. The material parameters of the treatment head are the elastic modulus and Poisson's ratio of the treatment head material. are the skin elasticity parameters (calculated by inversion of piezoresistive resistance data), namely the elastic modulus and Poisson's ratio of the skin;

[0086] The 64-point pressure-sensitive matrix data is input into the ANFIS network, and the training input is the pressure distribution , contact area A and equivalent curvature radius , the output is the optimal adjustment vector ,in They are Z-axis displacement, X-axis translation, Y-axis translation and rotation angle respectively; the network objective function is to minimize the pressure standard deviation: , where N is the number of pressure-sensitive matrix data points, is the pressure value of the i-th point, is the average pressure; constraints: , the maximum pressure difference in the pressure distribution ANFIS achieves adaptive decision-making by online updating of the fuzzy rule base through a hybrid gradient descent and least squares algorithm. It also introduces a dynamic weight distribution strategy in the position adjustment of the robotic arm to adjust the motion priority of each joint according to the direction of the pressure gradient:

[0087] Z-axis downforce: Weight , prioritize eliminating local high-pressure areas; XY axis translation: weight , is the pressure gradient. The larger the gradient, the higher the translation compensation weight. Rotational joint: weight , is the pressure distribution skewness, used to correct asymmetric contact.

[0088] Define the Contact Quality Index (CQI) to comprehensively evaluate the pressure field performance: , target value ≥ 0.97; when CQI does not meet the standard, a secondary optimization cycle is triggered until the preset conditions are met.

[0089] The zone-specific gradient temperature management module monitors temperature gradients by region, dynamically adjusts energy to control the temperature of each region, balances therapeutic thermal effects with tissue protection, and achieves precise temperature control.

[0090] A 16-channel ring-shaped temperature sensor array is placed inside the treatment head to monitor the temperature of each area at 50ms intervals. A three-dimensional heat conduction finite element model is established to predict the epidermal temperature distribution and dermal heat diffusion trend in real time. When the instantaneous temperature of a certain area exceeds 42°C, a graded response is initiated: the first response (42-45°C) reduces the output energy of the area by 30% and simultaneously activates the corresponding semiconductor cooling chip; the second response (>45°C) suspends treatment and activates the liquid cooling circulation system.

[0091] After the temperature recovers, the gradient restart strategy is implemented, giving priority to resuming treatment in the low-temperature area, and setting an energy increase gradient for every 5°C temperature difference (increasing energy by 10% per gradient) until the entire area reaches a balanced state.

[0092] The multimodal safety interlock module enables multi-dimensional safety monitoring and adaptive decision-making during the treatment process through constraints such as comfort scores (energy adaptability, contact uniformity, and thermal injury risk);

[0093] Establish a data bus between modules to exchange energy parameters, pressure distribution, and temperature field data in real time. When any module detects an abnormality, a multi-level interlock response is triggered: Level 1 alarm (single parameter exceeds the limit) initiates parameter compensation; Level 2 alarm (two modules are abnormal at the same time) suspends treatment and returns a fault code; Level 3 alarm (system-level fault) cuts off the main power supply and ejects the treatment head.

[0094] After the anomaly is resolved, a system self-check process is performed: the optical component calibration (deviation <2%), pressure sensor linearity (R²>0.99), and heat dissipation system efficiency (cooling rate>3°C / s) are verified in sequence. A treatment quality assessment report is generated, including energy adaptability (±3%), contact uniformity coefficient (>0.92), and thermal damage risk index (<0.05), and stored in the medical database for subsequent treatment reference.

[0095] By establishing a multi-dimensional scoring model and decision tree analysis algorithm, we can achieve objective evaluation of the comfort level of a single treatment and automatically trace the causes of abnormalities. The specific steps are as follows:

[0096] Based on the three core parameters of the quality assessment report (energy adaptation , contact uniformity coefficient , heat injury risk index ), construct a weighted scoring function: ,in Score the treatment comfort. They are energy adaptability weight, contact uniformity coefficient weight and thermal damage risk index weight, with values ​​of 0.4, 0.3 and 0.3 respectively. The scoring range is Between 0-100, when When the trigger is triggered, cause analysis;

[0097] Automatically locate the main source of the problem through the following rules: , then it is judged that the energy adaptation deviation is too large; when , and the pressure standard deviation is greater than 3kPa, it is judged that the contact pressure is uneven. , and the pressure standard deviation is less than 3kPa, then the treatment head posture deviation is determined; when When the maximum temperature is greater than 15℃, the cooling response is judged to be delayed. When the maximum temperature is less than 15°C, it is judged that the area of ​​insufficient blood perfusion is not identified;

[0098] Match the preset optimization strategy library based on the diagnosis results, for example:

[0099] If the cause is "large energy adaptation deviation," the recommendation is: "calibrate the spectral sensor before the next treatment and increase the number of Monte Carlo model iterations." If the cause is "uneven contact pressure," the recommendation is: adjust the dynamic weight parameters of the robotic arm.

[0100] The scores, reasons, and suggestions are appended to the evaluation report, and two types of storage tags are generated: patient tags: record individual comfort trends for personalized parameter presets; device tags: count the causes of high-frequency abnormalities to guide hardware / algorithm iterations.

[0101] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A control system for an intense pulsed light therapy device, characterized in that: include: Dynamic spectral energy adaptation module, used to calculate the absorption coefficient of each layer of skin, simulate energy deposition, optimize the energy distribution of the target and surrounding tissues through the objective function, and adapt the spectral energy; The 3D contact pressure field calibration module is used to input pressure-sensitive data into the ANFIS network. With the goal of minimizing the pressure standard deviation, it combines the robot arm joint weight strategy to calibrate the contact pressure field uniformity. The regional gradient temperature management module is used to monitor the temperature gradient by region, dynamically adjust energy to control the temperature of each region, balance the therapeutic thermal effect and tissue protection, and achieve precise temperature control; A multimodal safety interlock module is used to achieve multi-dimensional safety monitoring and adaptive decision-making during the treatment process through comfort score constraints; The execution process of the dynamic spectrum energy adaptation module is as follows: A multispectral sensor array is integrated into the treatment head to collect the reflected spectrum data of the treatment area in real time, and simultaneously activate the skin impedance detection electrode to obtain the epidermal electrical impedance parameters; The convolutional neural network is used to extract features from spectral data, identify the characteristic values ​​of skin melanin content, hemoglobin distribution and stratum corneum thickness, and calculate the comprehensive absorption coefficient of the skin by combining impedance parameters; A pulse energy dynamic adaptation model is established, and the absorption coefficient is input into a preset Bayesian optimization algorithm to output the optimal pulse width, spectral band, and energy density; Control the DSP chip to generate a dynamic modulation waveform, drive the semiconductor laser array to output according to the optimized parameters, and start closed-loop monitoring at the same time. When the real-time reflectivity change exceeds the threshold of ±5%, it triggers parameter re-optimization; Skin multi-layer optical modeling and Monte Carlo photon transport simulation accurately quantify the spatial distribution of light energy in skin tissue through mathematical and physical models, and dynamically optimize pulse parameters based on real-time feedback data.

2. The control system of an intense pulsed light therapy device according to claim 1, characterized in that: The specific operation steps of skin multi-layer optical modeling and Monte Carlo photon transmission simulation in the dynamic spectral energy adaptation module are as follows: The skin is defined as a three-layer structure, and the optical properties of each layer are determined by the absorption coefficient , scattering coefficient and anisotropy factor g; integrating multispectral sensor and impedance detection data, dynamically calculating the parameters of each layer: in, For skin tissue at wavelength The total absorption coefficient under For melanin at wavelength The total absorption coefficient under For hemoglobin at wavelength The absorption coefficient under The concentration of melanin in the skin and the concentration of hemoglobin in the skin are extracted by the convolutional neural network. This is the basic absorption value of the skin; The propagation path of photons in the skin is simulated based on random walk theory, and the energy deposition distribution is calculated: ,in In space coordinates and the energy deposition distribution at time t, is the photon weight, are the absorption coefficient and scattering coefficient respectively, s is the path length, is the total number of simulated photons; The energy distribution output by the Monte Carlo simulation is combined with the Bayesian optimization algorithm to construct the objective function: ,in, They are total energy output, energy in the target area, energy in the surrounding tissue, and the safe energy threshold for the surrounding tissue. Optimization variables include pulse width, spectral band combination, and energy density to ensure energy focusing in the target area while suppressing the risk of overheating in the surrounding tissue. In closed-loop monitoring, reflectivity change data is used to dynamically correct the Monte Carlo model. The error threshold is set to ±2% to improve the matching accuracy between the model and the patient's individual skin.

3. The control system of an intense pulsed light therapy device according to claim 1, characterized in that: The specific operation steps of the three-dimensional contact pressure field calibration module are as follows: A 64-point piezoresistive matrix is ​​arranged on the base of the treatment head to acquire contact pressure distribution data at a sampling rate of 100 Hz; A three-dimensional pressure field model is constructed through spatial interpolation algorithm to calculate the pressure gradient distribution and contact area ratio; When the maximum pressure difference is greater than 20kPa or the effective contact area is less than 85%, the six-axis robot arm is activated to adjust the posture: first, the Z-axis fine-tunes the downward pressure to the target value, then the position deviation is compensated by the XY-axis translation, and finally the joint is rotated to adjust the incident angle; Recalibrate the pressure field uniformity after each adjustment until the preset conditions of pressure standard deviation <3kPa and contact area >95% are met; By using multi-scale contact mechanics modeling and dynamic fuzzy decision-making mechanism, the contact state between the treatment head and the skin is optimized through the collaborative use of physical models and intelligent algorithms, significantly improving the pressure field uniformity and calibration efficiency.

4. The control system of an intense pulsed light therapy device according to claim 3, characterized in that: The specific operation steps of the multi-scale contact mechanics modeling and dynamic fuzzy decision-making mechanism are as follows: Based on the Hertz contact theory, a stress distribution model of the skin-treatment head contact area is established, and the equivalent curvature radius of the contact surface is defined. and elastic modulus : ,in are the equivalent curvature radius of the contact area between the skin and the treatment head, the curvature radius of the treatment head, and the curvature radius of the skin surface, respectively. The material parameters of the treatment head are the elastic modulus and Poisson's ratio of the treatment head material. are the skin elasticity parameters, namely the elastic modulus and Poisson's ratio of the skin; The 64-point pressure-sensitive matrix data is input into the ANFIS network, and the training input is the pressure distribution , contact area A and equivalent curvature radius , the output is the optimal adjustment vector ,in They are Z-axis displacement, X-axis translation, Y-axis translation and rotation angle respectively; The network objective function is to minimize the pressure standard deviation: , where N is the number of pressure-sensitive matrix data points, is the pressure value of the i-th point, is the average pressure; constraint: , the maximum pressure difference in the pressure distribution ,ANFIS updates the fuzzy rule base online by hybrid gradient descent and least squares algorithm to achieve adaptive decision-making; A dynamic weight distribution strategy is introduced in the robot arm posture adjustment to adjust the priority of each joint movement according to the direction of the pressure gradient: Z-axis downforce: Weight , give priority to eliminating local high-pressure areas; XY Translation: Weight , is the pressure gradient, the larger the gradient, the higher the translation compensation weight; Revolute Joints: Weights , is the pressure distribution skewness, used to correct asymmetric contact; Define the contact quality index CQI to comprehensively evaluate the pressure field performance: , target value ≥ 0.97; When the CQI does not meet the standard, a secondary optimization cycle is triggered until the preset conditions are met.

5. The control system of an intense pulsed light therapy device according to claim 1, characterized in that: The specific operation steps of the domain gradient temperature management module are as follows: A ring-shaped temperature sensor array is arranged inside the treatment head to monitor the temperature of each area at 50ms intervals; Establish a three-dimensional heat conduction finite element model to predict the epidermal temperature distribution and dermal heat diffusion trend in real time; When the instantaneous temperature in a certain area exceeds 42°C, a graded response is initiated: The first-level response is to reduce the output energy of the area by 30% at 42-45℃, and simultaneously turn on the semiconductor cooling plate at the corresponding position; Secondary response: when the temperature is greater than 45°C, treatment is suspended and the liquid cooling circulation system is activated; After the temperature recovers, the gradient restart strategy is implemented, giving priority to resuming treatment in the low-temperature area, and setting an energy increment gradient for every 5°C temperature difference until the entire area reaches a balanced state.

6. The control system of an intense pulsed light therapy device according to claim 1, characterized in that: The specific operating steps of the multi-modal safety interlock module are as follows: Establish a data bus between modules to exchange energy parameters, pressure distribution and temperature field data in real time; When any module detects an abnormality, a multi-level interlock response is triggered: Level 1 alarm means a single parameter exceeds the limit, parameter compensation is initiated, Level 2 alarm means two modules are abnormal at the same time, treatment is suspended and a fault code is fed back, Level 3 alarm means system-level fault, the main power is cut off and the treatment head is ejected; After the abnormality is resolved, the system self-check process is executed: the calibration of the optical components, the linearity of the pressure sensor and the performance of the cooling system are checked in sequence; Generate a treatment quality assessment report, including energy adaptability, contact uniformity coefficient, and thermal injury risk index, and store it in the medical database for subsequent treatment reference; By establishing a multi-dimensional scoring model and decision tree analysis algorithm, we can achieve objective evaluation of the comfort of a single treatment and automatic tracing of abnormal causes.

7. The control system of an intense pulsed light therapy device according to claim 6, characterized in that: The specific steps for establishing a multi-dimensional scoring model and a decision tree analysis algorithm in the multimodal safety interlock module are as follows: Based on the three core parameters of the quality assessment report, including energy fitness , contact uniformity coefficient , heat injury risk index , construct a weighted scoring function: ,in Rate the treatment comfort. They are energy adaptability weight, contact uniformity coefficient weight and thermal damage risk index weight, with values ​​of 0.4, 0.3 and 0.3 respectively. The scoring range is Between 0-100, when When the trigger is triggered, cause analysis; Automatically locate the source of the problem using the following rules: when , then it is judged that the energy adaptation deviation is too large; when , and the pressure standard deviation is greater than 3kPa, it is judged that the contact pressure is uneven. , and the pressure standard deviation is less than 3kPa, then the treatment head posture deviation is determined; when When the maximum temperature is greater than 15℃, the cooling response is judged to be delayed. When the maximum temperature is less than 15°C, it is judged that the area of ​​insufficient blood perfusion is not identified; Match the preset optimization strategy library based on the diagnosis results; The scores, reasons, and recommendations are appended to the assessment report, and two types of storage tags are generated: Patient tags: record individual comfort trends for personalized parameter settings; device tags: count the causes of high-frequency abnormalities to guide hardware or algorithm iterations.

8. The control system of an intense pulsed light therapy device according to claim 1, characterized in that: The execution process of the dynamic spectrum energy adaptation module also includes: A color recognition unit is integrated into the treatment head to collect color data of the color marking area of ​​the magnetic cover in real time; The color data is converted into corresponding spectral band combinations and energy density parameters through a preset color-parameter mapping table; Combined with the skin absorption coefficient, the pulse light output parameters are optimized.

Images