Method and system for dynamically regulating and controlling irradiance of phototherapy equipment based on neonatal jaundice index

Abstract

The invention discloses a phototherapy equipment irradiance dynamic regulation and control method and system based on a neonatal jaundice index, relates to the field of irradiance dynamic regulation and control, and pre-estimates a bilirubin decline rate in an ideal state through a preset phototherapy kinetic model in combination with child patient individual baseline characteristics. In the treatment process, the real bilirubin decline rate of a child patient is monitored in real time and calculated, and the real bilirubin decline rate and an expected target value are subjected to difference comparison of physiological response dimensions. And reversely deducing the current physical attenuation coefficient of the hardware by using the difference value. And finally, acting the attenuation coefficient on an original set value, and correcting the drive output of hardware in real time by dynamically generating an irradiance compensation instruction. Therefore, automatic closed-loop regulation and control from clinical effect evaluation to hardware power compensation are realized, the influence of hardware performance fluctuation on treatment accuracy is eliminated fundamentally, and continuous stability and high efficiency of phototherapy irradiance are ensured.

Description

Technical Field

[0001] This invention belongs to the field of dynamic irradiance control, specifically relating to a method and system for dynamic irradiance control of phototherapy equipment based on neonatal jaundice index. Background Technology

[0002] Neonatal hyperbilirubinemia (jaundice) is one of the most common clinical problems in the neonatal period. If not treated promptly, it can lead to serious neurological damage such as bilirubin encephalopathy. Phototherapy, as the most commonly used and effective first-line treatment, works by using specific wavelengths of light to promote the isomerization of unconjugated bilirubin in the skin and mucous membranes, which is then excreted through bile and urine. The effectiveness of phototherapy is directly related to the irradiance (irradiance), and the rate of decrease often depends on the degree of coupling between bilirubin levels and light energy. Therefore, developing a dynamic irradiance control scheme for phototherapy equipment based on the neonatal jaundice index has clinical value and industrial application significance for improving treatment efficiency, precisely controlling the rate of decrease in serum total bilirubin (TSB), and shortening the length of hospital stay for infants.

[0003] In current neonatal phototherapy practices, irradiance control primarily relies on pre-defined clinical guidelines. Healthcare professionals typically manually set or simply select the light source's output intensity based on the newborn's gestational age, weight, age, and whether high-risk factors such as hemolysis or infection are present, referring to standard phototherapy threshold curves. However, this experience-based or fixed-threshold control scheme has revealed significant shortcomings in long-term clinical application. Existing irradiance control algorithms generally lack real-time sensing capabilities for the health status of phototherapy equipment hardware, particularly neglecting the natural decay and aging of light sources (such as LED modules) over time. In departments like the NICU, phototherapy equipment operates at high frequency and for extended periods, making light source decay an unavoidable physical phenomenon. Current solutions often involve periodic manual calibration, but between maintenance intervals, the displayed power level may deviate significantly from the actual effective irradiance received by the infant. This implicit reduction in energy output leads to frequent clinical occurrences where equipment parameters appear normal, but the infant's treatment outcome is poor—that is, the actual rate of bilirubin reduction is significantly lower than the expected response at baseline. Because existing solutions fail to establish a closed-loop link between physiological indicator feedback, hardware attenuation inference, and adaptive output compensation, the system cannot identify hardware performance degradation based on individualized treatment data of newborns. This lack of real-time hardware performance monitoring and the absence of an adaptive compensation mechanism make it difficult for phototherapy equipment to maintain a consistent treatment intensity throughout the entire lifespan. The disconnect between the actual output of the equipment and the set value directly affects the prognosis and safety of the child.

[0004] Therefore, there is an urgent need in this field for a dynamic irradiance control scheme that can avoid the limitations of traditional open-loop control and achieve real-time monitoring and adaptive compensation of hardware attenuation, so as to ensure that phototherapy equipment can provide stable and accurate therapeutic doses throughout its entire life cycle. Summary of the Invention

[0005] This invention application provides a method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index.

[0006] The technical solution of the present invention is as follows: A method for dynamic control of irradiance of a phototherapy device based on neonatal jaundice index, comprising: S1: acquiring baseline physiological characteristic data of the newborn and setting an irradiance; S2: extracting a neonatal baseline feature set from the neonatal baseline physiological characteristic data, and estimating the expected response of the set irradiance based on the neonatal baseline feature set to obtain the expected bilirubin decrease rate under ideal output conditions without device aging; S3: continuously monitoring the baseline physiological characteristic data, and when a new round of bilirubin measurement value is detected, using the initial bilirubin concentration and initial timestamp in the neonatal baseline feature set, performing differential verification and matching on the newly acquired current bilirubin concentration and current timestamp to obtain the actual bilirubin decrease rate of the newborn; S4: performing reverse inference of the hardware attenuation coefficient between the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn to obtain a verified attenuation coefficient; S5: performing dynamic irradiance compensation on the set irradiance based on the verified attenuation coefficient to obtain a compensated irradiance control command, and sending the compensated irradiance control command to the hardware LED driver module.

[0007] This invention application provides another dynamic control system for the irradiance of a phototherapy device based on the neonatal jaundice index.

[0008] The technical solution of this invention is as follows: A dynamic irradiance control system for phototherapy equipment based on neonatal jaundice index, comprising: a neonatal subject data acquisition module, used to acquire baseline physiological characteristic data of the neonatal subject and set an irradiance; an expected bilirubin decrease rate calculation module, used to extract a neonatal baseline feature set from the neonatal subject's baseline physiological characteristic data, and perform expected response estimation on the set irradiance based on the neonatal baseline feature set to obtain the expected bilirubin decrease rate under ideal output conditions without equipment aging; and an actual bilirubin decrease rate calculation module, used to continuously monitor the baseline physiological characteristic data, and when a new round of bilirubin decrease is detected... During bilirubin measurement, the initial bilirubin concentration and initial timestamp from the neonatal baseline feature set are used to perform differential verification and matching on the newly acquired current bilirubin concentration and current timestamp to obtain the actual bilirubin decline rate of the newborn; the attenuation coefficient analysis module is used to perform hardware attenuation coefficient inverse inference on the actual bilirubin decline rate and the expected bilirubin decline rate of the newborn to obtain the verified attenuation coefficient; the irradiance control command generation module is used to perform dynamic irradiance compensation on the set irradiance based on the verified attenuation coefficient to obtain the compensated irradiance control command, and the compensated irradiance control command is sent to the hardware LED driver module.

[0009] Compared with existing technologies, the beneficial effects of this invention are as follows: Addressing the shortcomings mentioned in the background technology where hardware aging and light source attenuation lead to deviations from expected treatment effects, this solution uses a pre-set phototherapy dynamics model combined with the individual baseline characteristics of the child to predict the ideal bilirubin reduction rate. During treatment, the actual bilirubin reduction rate of the child is monitored and calculated in real time, and compared with the expected target value using a physiological response dimension difference. This difference is used to infer the current physical attenuation coefficient of the hardware, thereby quantifying the power loss caused by equipment aging. Finally, this attenuation coefficient is applied to the original set value, and the hardware drive output is corrected in real time by dynamically generating irradiance compensation commands. This method breaks the open-loop limitation of traditional equipment relying solely on set parameters, achieving automatic closed-loop control from clinical effect evaluation to hardware power compensation, fundamentally eliminating the impact of hardware performance fluctuations on treatment accuracy, and ensuring the continuous stability and high efficiency of phototherapy irradiance. Attached Figure Description

[0010] Figure 1 This is a flowchart of a phototherapy device irradiance dynamic control method based on neonatal jaundice index according to an embodiment of this application.

[0011] Figure 2 This is a schematic diagram of the data flow of a phototherapy device irradiance dynamic control method based on neonatal jaundice index according to an embodiment of this application.

[0012] Figure 3This is a flowchart of step S4 in the method for dynamic control of irradiance of a phototherapy device based on neonatal jaundice index according to an embodiment of this application.

[0013] Figure 4 This is a block diagram of a phototherapy device irradiance dynamic control system based on neonatal jaundice index, according to an embodiment of this application. Detailed Implementation

[0014] The technical solutions of 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. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0015] In view of the shortcomings of the prior art, this application proposes a method for dynamic control of the irradiance of a phototherapy device based on the neonatal jaundice index. For example... Figure 1 and Figure 2 As shown, Figure 1 This is a flowchart of a phototherapy device irradiance dynamic control method based on neonatal jaundice index according to an embodiment of this application. Figure 2 This is a schematic diagram of the data flow of a phototherapy device irradiance dynamic control method based on neonatal jaundice index according to an embodiment of this application.

[0016] Specifically, step S1 involves acquiring baseline physiological characteristic data of the newborn and setting the irradiance. It should be understood that during phototherapy for neonatal jaundice, the gestational age, weight, and age of the newborn, as well as accompanying pathological characteristics (such as the presence of hemolysis, sepsis, etc.), directly determine their biological sensitivity to light energy. Traditional fixed-frequency or fixed-power treatment modes fail to consider the impact of individual differences in bilirubin metabolism efficiency, resulting in the inability to establish precise phototherapy kinetics predictions. Therefore, by executing step S1, this application can establish a standardized individual physiological anchor point for subsequent precision treatment, transforming static clinical indicators into calculable feature vectors. This provides the necessary input parameters for subsequently calculating the expected bilirubin decrease rate under ideal, non-aging conditions, enabling the control logic to shift from single hardware parameter control to closed-loop feedback control centered on the individual response of the newborn.

[0017] In a specific embodiment of this application, step S1 is implemented as follows: First, the original medical data of the newborn is obtained through a system interface or human-computer interaction interface. This original medical data contains a large amount of redundant clinical information, such as the child's name, hospital admission number, bed information, and other non-algorithm-related items. Therefore, this step first removes the aforementioned irrelevant information through preliminary formatting, cleaning, and noise reduction processing, thereby accurately separating and precipitating the baseline physiological characteristic data. Specifically, the baseline physiological characteristic data of the newborn in this application includes initial bilirubin concentration, initial timestamp, child weight, and risk factor weighted score. Specifically, the initial bilirubin concentration... Obtained by transcutaneous bilirubin monitoring or biochemical blood test, in mg / dL; initial timestamp The precise moment when phototherapy officially began was recorded; the child's weight. The value, expressed in kilograms, reflects the correlation between body surface area and light coverage. The most crucial feature processing involves the risk factor weighting score. The score is constructed using a multi-factor evaluation model, and its mathematical expression is as follows: In the above formula, This represents the total number of high-risk factors of clinical concern, such as preterm birth, low birth weight, hemolysis, and sepsis. As a binary variable, when the child has the i-th high-risk factor, The value is 1 if it is set to 1, and 0 otherwise. This refers to the corresponding risk weighting factor. Risk weighting factor These parameters, obtained through logistic regression training on a massive amount of historical clinical medical records, represent the contribution of each factor to the resistance to bilirubin reduction. For example, in the clinical preset, the weight of hemolysis can be set to 2.5, while the weight of simple mild premature gestation can be set to 1.2. This weighted summation operation transforms complex qualitative pathological descriptions into quantitative numerical characteristics. Simultaneously, the irradiance is set... It is based on the target light intensity preset in clinical guidelines, for example, set to 35 μW / cm. 2 / nm. In the data stream processing architecture, the execution module will extract the initial bilirubin concentration. Initial timestamp The child's weight and the calculated weighted scores of risk factors Data is packaged to generate a standardized baseline feature set. Taking a newborn weighing 3.2 kg with sepsis, a weight of 2.0, and an initial bilirubin concentration of 18.5 mg / dL as an example, the execution module obtains this information at 08:00:00 on March 3, 2026. This is converted into corresponding timestamp values. This feature set is then pushed into a memory buffer and used as a global variable in the expected response estimation of the subsequent step S2.

[0018] Specifically, step S2 involves extracting a neonatal baseline feature set from the neonatal subject's baseline physiological characteristic data, and estimating the expected response to a set irradiance based on this feature set to obtain the expected bilirubin reduction rate under ideal output conditions without device aging. Correspondingly, during neonatal phototherapy, the metabolic efficiency of bilirubin is highly constrained by individual physiological state and pathological background. Even under the same irradiance, neonates with different constitutions exhibit significant differences in biological responses, making a single irradiance setpoint unsuitable as the sole criterion for measuring treatment effectiveness. Given the inevitable physical degradation of light source performance due to long-term device operation, without a theoretical benchmark for a specific infant, the control unit cannot distinguish whether the slowdown in actual therapeutic effect stems from the infant's pathological resistance or from a hidden drop in hardware output power. This application introduces step S2 to establish a theoretical therapeutic effect model for a specific infant under ideal conditions without device aging, based on the infant's baseline physiological characteristics. This step transforms static individual characteristics into dynamic, expected bilirubin decline rates, providing a precise reference coordinate system for subsequent reverse inference of the actual hardware degradation level. It is the core of the algorithm for realizing hardware performance adaptive compensation logic.

[0019] In one specific embodiment of this application, step S2 includes: S21, using analytical extraction methods to perform feature stripping and encapsulation on the baseline physiological characteristic data of the newborn subject, so as to extract the initial bilirubin concentration, initial timestamp, infant weight and risk factor weighted score and package them to generate a baseline feature set; S22, based on the phototherapy kinetic response model, normalizing the baseline feature set and the set irradiance and calculating the expected response to obtain the expected bilirubin decrease rate.

[0020] The implementation process is as follows: First, sub-step S21 is executed. This sub-step aims to transform the discrete clinical parameters obtained in step S1 into a standard input vector that the algorithm can efficiently call. The parsing extraction method uses a pre-defined data dictionary mapping to accurately extract the initial bilirubin concentration from the previously cleaned and transformed baseline physiological feature data. Initial timestamp The child's weight Weighted score of risk factors These discrete feature dimensions are then structurally combined by the execution module and encapsulated into a strongly correlated data object, namely the baseline feature set. This encapsulation process not only ensures the atomicity of the data but also provides a standardized input interface for the subsequent parallel computation of the dynamic model. Taking the example in step S1, the baseline feature set at this time includes the initial bilirubin concentration of 18.5 mg / dL, the timestamp value at the start time, the body weight of 3.2 kg, and a weighted score of 2.0 reflecting the risk of sepsis.

[0021] Then, sub-step S22 is executed. The phototherapy kinetic response model, in its algorithmic implementation, is represented by the formula: The model logically corresponds to three core functional modules: a normalization preprocessing layer, an energy coupling calculation layer, and a physiological correction layer. The first layer of the model is the normalization preprocessing layer, whose execution logic is as follows: In this layer, the initial bilirubin concentration... This represents the density of unconverted biochemical substrates in the blood, while the child's weight... This serves as a proxy variable for body surface area and circulating blood volume, reflecting the effectiveness of light coverage and the response volume. Through the quotient of these two variables, this layer can transform raw physiological data with significant individual differences into a substrate distribution baseline per unit body weight. In other words, the total metabolic burden represented by the same concentration of bilirubin differs drastically in children of different weights; normalization eliminates the incomparability of children of different body types in the initial state, providing a unified computational basis for subsequent energy coupling. Next, the energy coupling calculation layer is entered, which combines the normalized physiological baseline with the set irradiance. and dynamic constants Perform multiplicative coupling. Set the irradiance. This represents the expected physical light energy input and is the core driving force behind photobiochemical reactions. The kinetic constant... The core control parameters for this layer are obtained through nonlinear regression training on tens of thousands of phototherapy cases where the equipment is in brand-new condition (without aging). During training, the least squares method is used for iterative optimization to determine the biochemical conversion efficiency of a unit light intensity on a unit bilirubin level under standard physical conditions. The computational essence of this layer is to transform the abstract light energy input into a specific biological metabolic driving force, determining the theoretical rate of decline after excluding pathological interference. Finally, the model calibrates the above calculation results for clinical dimensions through a physiological correction layer, which logically corresponds to multiplying by the risk factor weighted score. The reciprocal of the expected bilirubin reduction rate. Because children may have accompanying pathological conditions such as hemolysis and sepsis that affect bilirubin production or excretion, these complex clinical factors can significantly inhibit or alter the actual marginal effect of phototherapy; that is, the greater the pathological resistance, the slower the expected bilirubin reduction rate should be. Therefore, the physiological correction layer performs non-linear scaling using the risk factor weighted score as the denominator. The physiological correction layer introduces a risk weight factor determined by logistic regression training to non-linearly scale the expected response intensity, ensuring that the calculation results closely match the child's actual pathological resistance or metabolic sensitivity. Thus, this three-layer architecture, through rigorous data flow, deeply integrates the originally isolated hardware parameters with individual physiological characteristics, ultimately outputting the expected bilirubin reduction rate. It can serve as a precise benchmark for measuring the performance of a device under ideal, un-aged output conditions. Specifically, The set irradiance obtained from step S1, such as 35 μW / cm². 2 / nm represents the energy input value under ideal conditions. Risk factors are weighted and scored to provide a biological correction for metabolic resistance in pathological states. The aforementioned dynamic constants, solidified through training, are preset to 0.00124 in this embodiment based on the historical ideal model. The process is then deduced using specific numerical values: Baseline feature set input... =18.5, =3.2, =2.0, set irradiance =35. First, the calculation unit obtained a normalized physiological baseline of 5.78125; then, it was multiplied by the set irradiance, the reciprocal of the risk factor score, and the kinetic constant. The expected rate of bilirubin decrease was then calculated. Approximately 0.1255 mg / dL / h. This value represents the bilirubin concentration that the current child should experience per hour under a specific light intensity, assuming no hardware aging. This calculation result is then pushed onto the memory stack as a target baseline, awaiting differential comparison with the actual rate of decrease observed in real time in subsequent step S3.

[0022] Specifically, step S3 involves continuously monitoring baseline physiological characteristic data. When a new round of bilirubin measurements is detected, the initial bilirubin concentration and initial timestamp from the neonatal baseline feature set are used to perform differential verification and matching with the newly acquired current bilirubin concentration and current timestamp to obtain the actual rate of bilirubin decline in the newborn. It is understandable that during the clinical implementation of neonatal phototherapy, the metabolism of bilirubin in the infant's body is a biochemical process that changes continuously over time. Although the phototherapy device is set with a target power at the initial startup, during the actual treatment cycle, the infant's serum bilirubin concentration is affected by a combination of factors, including light energy absorption, the infant's own metabolic rate, and fluctuations in the actual hardware output, exhibiting a non-linear decreasing trend. Since clinical bilirubin monitoring is done through discrete blood tests or percutaneous detection, instantaneous metabolic efficiency cannot be directly obtained. Therefore, step S3 is performed to transform these discrete clinical measurement points into quantifiable real-time treatment response indicators. By establishing a continuous monitoring mechanism and differential calculation model, this step can capture the trajectory of bilirubin changes in children under real physical conditions, thereby obtaining actual response data that reflects the true output performance of the equipment, providing the most crucial empirical basis for subsequent identification of hardware performance degradation.

[0023] In one specific embodiment of this application, step S3 is implemented as follows: First, the execution module starts a resident background monitoring process that continuously scans the baseline physiological characteristic data stream in the interface or device data buffer. When medical staff complete a new round of percutaneous bilirubin measurement or blood test result entry, the monitoring process immediately captures this data update event and triggers the subsequent parsing program. The parsing program accurately extracts the current bilirubin concentration after phototherapy from the newly entered message. and the associated generated current timestamp After obtaining the latest observations, the execution module enters the crucial differential verification and matching stage. At this point, the processing unit retrieves the initial bilirubin concentration from the baseline feature set generated in step S2. With the initial timestamp To ensure the calculated rate of decline is statistically significant and to eliminate spurious signals caused by sensor errors or short-term physiological fluctuations, the execution module introduces a minimum kinetic statistical interval verification mechanism. The parameters of this mechanism are determined based on large-scale data analysis of bilirubin metabolic kinetics, such as a preset value of 2 hours. The processing unit calculates the current timestamp. With the initial timestamp The absolute difference between them is considered valid only if the difference is greater than or equal to the preset minimum dynamic statistical interval. Only samples that meet the time interval requirements are considered valid for subsequent calculations. If the time interval is insufficient, the sample is suspended until the next monitoring cycle to ensure that the calculation results accurately reflect the sustained effects of phototherapy. Once the time interval is validated, the execution module uses the difference equation to calculate the actual rate of bilirubin decrease in newborns. The calculation. In a specific embodiment of this application, step S3 includes: performing differential verification and matching between the newly acquired current bilirubin concentration and the current timestamp using the following formula to obtain the actual bilirubin decrease rate in the newborn, wherein the formula is: in, This represents the initial bilirubin concentration, indicating the concentration at the start of phototherapy, expressed in mg / dL. This represents the current bilirubin concentration. This is the current timestamp for the current bilirubin concentration; This is the initial timestamp for the initial bilirubin concentration. This represents the actual rate of bilirubin decrease in the newborn. In the above formula, the numerator term... The absolute decrease in total bilirubin under light exposure was calculated. (Denominator term) This represents the total irradiation time (in hours) between the two measurements. Through this ratio calculation, the formula converts the discrete concentration difference into a metabolic rate per unit time, reflecting the biological absorption feedback of the child to the current phototherapy energy in the real world, implicitly including potential power losses from the hardware. Continuing from step S2, a specific numerical example: initial bilirubin concentration. The concentration was 18.5 mg / dL, with an initial timestamp. The record is for March 3, 2026, at 08:00:00. If the monitoring module detects a new round of bilirubin measurements at 14:00:00 on the same day... The concentration is 17.8 mg / dL, and the current timestamp is [data missing]. The time difference from the initial timestamp is 6 hours. Since 6 hours is greater than the preset minimum dynamic statistical interval of 2 hours, this data point meets the difference verification condition. Substituting the values ​​into the formula for calculation: =(18.5-17.8) / (14.0-8.0)≈0.1167mg / dL / h, the calculated actual rate of decrease of bilirubin in newborns is 0.1167mg / dL / h.

[0024] Specifically, step S4: The actual and expected bilirubin decrease rates in the newborn are compared using a hardware attenuation coefficient to obtain a verified attenuation coefficient. It is understandable that during long-term clinical operation of phototherapy equipment, the photoelectric conversion efficiency of the LED light source inevitably degrades with increasing usage hours. Since phototherapy often lasts for several days and is conducted in a closed or semi-closed incubator, medical staff cannot visually observe subtle decreases in light output intensity. Frequent manual calibration using an external radiometer would interfere with the treatment process and increase the risk of cross-infection. When the hardware output capacity decreases, even if the control system's setpoint remains high, the effective photon flux received by the child is significantly insufficient. Therefore, this application introduces step S4 to transform the individual child's physiological metabolic feedback into a precise perception of the hardware's physical state. Through a reverse mapping logic of human-machine interaction, the current energy attenuation level of the device is quantitatively inferred without interrupting treatment.

[0025] Figure 3 This is a flowchart of step S4 in the method for dynamic control of irradiance of a phototherapy device based on the neonatal jaundice index according to an embodiment of this application. Figure 3 As shown, in a specific embodiment of this application, step S4 includes: S41, performing initial reverse inference of the hardware attenuation coefficient between the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn to obtain the original inferred attenuation coefficient; S42, pushing the original inferred attenuation coefficient into the historical inferred attenuation coefficient queue, and performing moving average filtering on the updated historical inferred queue to obtain the verified attenuation coefficient.

[0026] The implementation process is as follows: Step S4 first performs an initial reverse inference of the hardware attenuation coefficient between the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn through sub-step S41, thereby obtaining the original inferred attenuation coefficient. In a specific embodiment of this application, step S41, performing an initial reverse inference of the hardware attenuation coefficient between the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn to obtain the original inferred attenuation coefficient, includes: S411, comparing the physiological response difference between the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn to obtain the physiological response ratio; S412, mapping the physiological response ratio and the set irradiance to a physical attenuation state to obtain the original inferred attenuation coefficient.

[0027] In step S411, the actual rate of decline in neonatal bilirubin calculated in step S3 is received. and the expected rate of bilirubin decrease calculated from step S2 The physiological response difference was compared and calculated for these two key parameters. The specific mathematical expression is as follows: In the above formula, Defined as the physiological response ratio. This formula constructs a dimensionless index to measure the deviation between actual therapeutic effect and theoretical ideal efficacy. The numerator represents the actual metabolic performance after being affected by the actual hardware performance, while the denominator represents the ideal physiological benchmark after excluding hardware aging interference. When this ratio is less than 1.0, it logically excludes individual pathological fluctuation factors of the child (because pathological factors have already been...). The risk factor weighting score is corrected, directly pointing to the lack of hardware output capability. This ratio calculation realizes the transformation from absolute rate value to relative efficiency ratio, and is a logical bridge connecting biochemical indicators and physical power parameters. Then, step S412 is executed to obtain the physiological response ratio. With set irradiance Physical attenuation state mapping is performed. During this process, the irradiance will be set. This serves as a calibration reference system for the ideal, age-free state of phototherapy equipment. In the kinetic logic architecture, the set irradiance is mapped to a physical reference coefficient of 1.0, meaning that when the hardware is in a brand-new state, the actual output should be completely equivalent to the set command. Since the output dose of photons and the isomerization rate of bilirubin molecules exhibit a highly linear positive correlation within the clinically commonly used dose range in phototherapy kinetics, the aforementioned physiological response ratio is directly and equivalently mapped to the actual ratio of the hardware's light output power to the set power, based on the inverse mapping solution logic. The calculation logic is as follows: In the formula This is the original inferred attenuation coefficient. Through this mapping, the originally abstract deviation in the bilirubin decrease rate is successfully transformed into a specific percentage of hardware physical output efficiency. For example, if the ratio is 0.9, it means that the actual output power of the hardware is only 90% of the nominal value. This logic, by setting irradiance data, confirms the premise of the current physiological ratio mapping and ensures the consistency between the physical inference process and the clinically pre-set logic. Specifically, the expected bilirubin decrease rate calculated in step S2... The rate was 0.1255 mg / dL / h, while the actual rate of bilirubin decrease in newborns obtained through continuous monitoring and differential verification in step S3 was... The value was 0.1167 mg / dL / h. The S411 comparison calculation was performed to obtain the physiological response ratio. =0.1167 / 0.1255≈0.93. Then, the S412 mapping process begins, with an irradiance set at 35 μW / cm². 2 / nm is anchored as a reference for the ideal output state, and this ratio of 0.93 is directly mapped to the original inferred attenuation coefficient. =0.93. This value clearly reveals that due to LED aging or other hardware wear and tear, the actual output efficiency of phototherapy equipment has dropped to 93% of its ideal state.

[0028] To further eliminate random noise caused by single bilirubin measurement errors or short-term physiological fluctuations, the execution module then executes sub-step S42. In a specific embodiment of this application, step S42, which pushes the original inferred attenuation coefficient into the historical inferred attenuation coefficient queue and performs moving average filtering on the updated historical inferred queue to obtain the verified attenuation coefficient, includes: S421, performing data stacking and out-of-bounds removal processing on the original inferred attenuation coefficient and historical inferred attenuation coefficient queues to obtain the historical inferred queue that has completed the state update operation; S422, performing physiological noise filtering and physical feature solidification on the historical inferred queue based on the moving average filtering module to obtain the verified attenuation coefficient.

[0029] First, sub-step S421 is executed. The historical inferred attenuation coefficient queue is a first-in-first-out (FIFO) data structure stored in local solid-state memory. Essentially, it's a time-sliding sampling window that records the hardware performance data inferred from different patients' feedback over a period of time. First, the queue is retrieved from memory, then a push operation is performed, inserting the calculated original inferred attenuation coefficients at the end of the sliding queue. During this data injection process, the newly generated original coefficients are utilized and integrated into the existing data structure, preventing any isolated data objects and ensuring the continuity of data flow. Subsequently, the total number of elements in the historical inferred attenuation coefficient queue is monitored in real-time to determine if it exceeds the preset maximum window threshold. The preset maximum window threshold is a key control parameter, set based on a comprehensive consideration of clinical sampling frequency and hardware aging time constant. In this embodiment, the threshold is preset to 50. The logic behind setting it to 50 is that if bilirubin monitoring during phototherapy is performed every 12 hours, this window can cover the device's operating status for the past 25 days, sufficient to filter short-term physiological fluctuations while capturing trend-based physical attenuation. If the queue length exceeds 50, a dequeue operation is automatically performed, removing the oldest historical element from the head of the queue to maintain a constant queue length. The queue after completing the above state update operation is defined as the history inference queue.

[0030] Next, sub-step S422 is executed. The moving average filtering module is a signal processing architecture based on time-domain weighted smoothing. Its core logic lies in decoupling two drastically different variable characteristics: the physiological differences of the children manifest as bidirectional zero-mean random fluctuations around a baseline, meaning that due to instantaneous changes in the individual's metabolic microenvironment, the attenuation coefficient inferred in a single instance may be too high or too low, but statistically these fluctuations cancel each other out; while the aging and degradation of hardware exhibits a unidirectional, slow decline in physical characteristics, with obvious trends and irreversibility. This module performs noise reduction feature decoupling calculations, using the arithmetic mean method to cancel out high-frequency physiological noise, thereby extracting low-frequency smooth features that reflect the true state of the hardware. The specific calculation process is performed according to the following formula: In the above formula, This represents the total number of actual elements contained in the current historical inference queue. This value is used as the denominator in the calculation, and its dynamic range is between 1 and 50. This represents the j-th single-inference decay coefficient sample contained in the historical inference queue. First, a summation operation is performed on all samples in the queue to obtain the cumulative performance value within this time window. Then, this value is divided by the denominator. The arithmetic mean within the current time window is calculated. This process transforms the discrete raw coefficients, which contain physiological random errors, into smooth values ​​representing the steady-state performance of the hardware, and finally solidifies and assigns them to the verified decay coefficient. Continuing with the numerical example of the previous steps, if the raw inferred decay coefficient obtained in step S41 is 0.93, and the historical inferred decay coefficient queue now contains 49 samples from the past period, with values ​​fluctuating between 0.91 and 0.95, the processing unit first executes operation S421, pushing 0.93 to the end of the queue. At this point, the total number of samples in the queue reaches 50, which does not exceed the maximum window threshold, so the dequeue operation is not performed temporarily. Subsequently, in stage S422, the moving average filtering module sums these 50 samples. If the sum of these 50 samples is 46.25, dividing the sum by the denominator 50 yields the verified decay coefficient of 0.925. The attenuation coefficient of 0.925 after verification indicates that, after multiple sample verifications, the output efficiency of the device's LED light source has indeed decreased to 92.5% of the nominal power. Compared to the 0.93 obtained from a single inference, this filtered value eliminates the influence of single blood sampling errors or the instantaneous metabolic fluctuations of the child, and more realistically depicts the physical performance degradation of the hardware.

[0031] It is understandable that when phototherapy devices perform dynamic irradiance compensation, their underlying logic is to increase the output current of the hardware LED driver module to compensate for the power deficit caused by the aging of the light source. However, semiconductor light-emitting elements have physical limits to their current carrying capacity and heat dissipation efficiency. When the physical attenuation of the light source exceeds a certain limit, simply increasing the driving current not only fails to linearly increase irradiance but also leads to severe heat generation problems due to a surge in power consumption, potentially inducing LED thermal runaway or spectral drift, thereby threatening the treatment safety of children and possibly causing permanent damage to the equipment. Performing a safety assessment of the validated attenuation coefficient is to establish a safety barrier between the automated compensation algorithm and the physical tolerance of the hardware. This step is to interrupt dangerous automatic compensation logic and guide manual intervention in time before hardware performance degrades to an uncontrollable range, thereby ensuring that the phototherapy process always remains within the physical safety envelope and guaranteeing the long-term stability and medical safety of the equipment.

[0032] In one specific embodiment of this application, the method further includes: performing a safety assessment on the attenuation coefficient after verification, and triggering a hardware maintenance warning signal when the attenuation coefficient after verification is lower than a preset limit tolerance threshold.

[0033] In the specific implementation process, this security assessment logic immediately follows the moving average filtering in step S42. The received, solidified verification attenuation coefficient... This is used as the sole input to the security assessment module. At this point, a preset tolerance threshold is retrieved from the system's read-only memory. The limit tolerance threshold is a critical indicator for measuring the health of hardware. In this embodiment, based on the correlation model between device luminous efficiency and thermal load, this threshold is scientifically set to 0.65. This value was obtained through large-scale experimental training on the driving power and temperature rise curves of the LED module under different attenuation levels. During the training process, temperature sensor arrays were used to collect LED core temperature data. By fitting a model of the relationship between attenuation coefficient, driving current, and junction temperature, it was determined that when the coefficient is below 0.65, i.e., the light source attenuation exceeds 35%, the current required to maintain the set irradiance will trigger the thermal protection threshold. Therefore, this threshold represents the minimum energy efficiency ratio at which the hardware can safely perform closed-loop compensation. The safety assessment is performed using the following judgment formula: In the above formula, This is the attenuation coefficient calculated in the previous stage after verification. The preset tolerance threshold is set, such as 0.65. A logical comparison operation is performed; if... If the value is greater than or equal to the threshold, the current hardware state is determined to be in a compensable safe zone, and the system continues to generate subsequent dynamic compensation instructions. If the current is below the threshold, a hardware interrupt logic is directly triggered. This logic uses an internal watchdog mechanism to intercept current increase commands sent to the LED driver module in real time, forcibly stopping automatic compensation to prevent hardware overheating. When the judgment result is Alert, a hardware maintenance warning signal is generated and activated simultaneously. This signal has multi-dimensional characteristics. First, the signal is pushed to the central monitoring screen of the nursing station via the local area network communication protocol, displaying a message in the form of a bright pop-up indicating that the equipment hardware is severely aging and that the light source should be replaced immediately, ensuring that medical staff can be informed of the risk immediately. Second, the operation panel on the device will simultaneously trigger the underlying I / O control, causing the normally off or normally green warning light to light up bright yellow, providing a visual alarm at the bedside. Specifically, the attenuation coefficient calculated in step S42 is 0.925, which is then compared with the limit tolerance threshold of 0.65. Since 0.925 is much greater than 0.65, the safety assessment module determines that although the device has slight attenuation, it is still within the safe control range, therefore no warning is triggered, and the system is allowed to continue performing irradiance compensation. However, if, after several months of equipment operation, the light source ages rapidly and the attenuation coefficient, as verified by step S42, drops to 0.62, which is below the limit tolerance threshold of 0.65, the safety assessment module will instantly determine it as an unsafe state, immediately execute watchdog interception, interrupt compensation commands, and send hardware maintenance warning signals to the nursing station and equipment, forcibly requiring the replacement of the light source module.

[0034] Specifically, step S5: Based on the verified attenuation coefficient, dynamic irradiance compensation is performed on the set irradiance to obtain the compensated irradiance control command, which is then sent to the hardware LED driver module. In other words, during the long-term clinical application of phototherapy equipment, although medical personnel set the target irradiance according to clinical guidelines, the actual output gap caused by the physical attenuation of the light source is dynamic and implicit. If the hardware is driven solely according to the original set value, the actual photon energy received by the child will remain below the treatment threshold, leading to a slow decrease in bilirubin and even recurrence of the condition. Therefore, this application ultimately introduces step S5 to accurately convert the hardware performance loss inferred in the aforementioned steps into physical gain compensation. Through this step, the driving energy level of the hardware can be increased in reverse according to the current real-time health status of the device, thereby physically offsetting the power drop caused by aging. This ensures that the actual output irradiance of the device throughout its entire lifespan remains highly consistent with the preset clinical command, providing constant and precise therapeutic energy for newborns.

[0035] In one specific embodiment of this application, step S5 is implemented as follows: Example 1 Obtain the original set irradiance from step S1. And obtain the verification attenuation coefficient after solidification by moving mean filtering from step S4. Before generating the compensation command, it is first verified whether the current post-verification attenuation coefficient is within the safe envelope, i.e., by retrieving the preset limit tolerance threshold. A preliminary assessment is performed. If the attenuation coefficient is below this threshold, it indicates that the light source is aging too severely, and further increasing the drive current may trigger thermal runaway. After verification, the attenuation coefficient is confirmed to be within a safe range. Then, dynamic irradiance compensation calculation is performed, and its mathematical expression is as follows: In the above formula, The target irradiance set for clinical use is expressed in μW / cm². 2 / nm, To verify the attenuation coefficient, the core of the formula analysis lies in using the reciprocal proportional relationship to achieve power gain. Since the verified attenuation coefficient is a proportional value less than or equal to 1, such as 0.925, using it as the denominator in the calculation will inevitably cause the numerator to... It is magnified proportionally, thus resulting in a higher internal driving target value. This directly utilizes hardware degradation data, using mathematical compensation to replenish the percentage of physical loss in the drive instructions. Subsequently, the calculated... The signal is passed to the output mapping module. This module contains a pre-calibrated power-current mapping model. This model's architecture is based on a lookup table method combined with a linear interpolation algorithm. Its training process involves recording standard output intensities under different drive currents using a high-precision irradiance meter before the device leaves the factory. Based on the compensated target value, the mapping module calculates the corresponding pulse width modulation (PWM) duty cycle or drive current value and encapsulates it into a compensated irradiance control command. Continuing with the numerical example of the preceding steps: setting the irradiance... 35 μW / cm 2 / nm, attenuation coefficient after verification The value is 0.925. Perform the calculation: =35 / 0.925≈37.84μW / cm 2The originally set irradiance of 35 units was dynamically increased to 37.84 units per nm. This means that the control unit actively boosted the driving energy level by approximately 8.1% to compensate for the 7.5% hardware attenuation. The resulting compensated irradiance control command was sent to the hardware LED driver module in real time. Upon receiving this command, the hardware LED driver module adjusted the on-time or current of the power transistor to drive the light source to emit the enhanced beam. Due to the aging of the light source itself, this enhanced beam, after attenuation, will precisely return the actual value irradiated onto the newborn's skin to the clinically required 35 μW / cm². 2 / nm.

[0036] Example 2 Understandably, in the clinical application of phototherapy for neonatal jaundice, the core treatment hardware heavily relies on high-power LED arrays emitting specific peak wavelengths of 450nm to 460nm. When performing dynamic compensation calculations and generating instructions, if the driving energy is amplified solely by a simple mathematical division of the set irradiance by the verified attenuation coefficient, as in the first embodiment, this processing logic has a fatal underlying flaw. This purely linear approach completely ignores the nonlinear relationship of electro-thermal-optical multi-physics coupling that inevitably arises in semiconductor light-emitting devices under aging and high-load driving conditions. Specifically, when physical aging of the device leads to light decay, the non-radiative recombination ratio inside the LED will surge significantly. At this point, blindly performing linear current amplification compensation will disproportionately convert the additional injected electrical energy into Joule heat. This unpredictable and drastic temperature rise can directly trigger two types of secondary physical disasters: firstly, a thermal drop, where an abnormally high junction temperature causes a second, precipitous decrease in the LED's internal quantum efficiency, making simple current amplification unable to achieve a proportional gain in optical power, resulting in ineffective compensation; secondly, a spectral redshift, where the high accumulation of heat forces the LED's emission peak wavelength to drift towards longer wavelengths, severely deviating from the optimal biochemical absorption peak required for bilirubin isomerization. This means that even if the total physical optical power appears to meet the standard, the actual effective photon dose for clinical treatment of neonatal jaundice is significantly lost, which not only easily induces thermal runaway and even fire risks in the underlying hardware, but also significantly reduces the efficacy of jaundice reduction due to compensation failure. To address this, this application introduces cross-correction based on thermal and spectral dynamics dimensions to avoid the blind spot in the first embodiment's optical power meter being unable to distinguish between effective and ineffective wavelength photons. Through proactive pre-calculation of thermal dynamics and spectral weighting mapping, the medical safety hazards caused by high temperatures are eliminated from the underlying physical logic, ensuring the absolute therapeutic dose accuracy of the phototherapy device.

[0037] Based on this, in the second embodiment of this application, step S5 includes: Based on the verified attenuation coefficient and hardware thermal resistance factor, a forward estimation of the junction temperature rise is performed on the set irradiance to obtain the predicted junction temperature rise. In other words, it is necessary to pre-quantify the surge in non-radiative composite heat generation due to device aging before actually issuing the incremental compensation current with potential overheating risk. Implementing this step achieves the technical effect of accurately pinpointing the physical risk of thermal runaway without adding any additional physical temperature probes. In the specific implementation process, based on a thermodynamic model, the hardware thermal resistance factor reflecting the degree of dust accumulation or thermal grease aging on the heatsink over time, the set irradiance required by the current clinical order, and the verified attenuation coefficient characterizing the current physical aging level of the hardware are received. By converting the verified attenuation coefficient into an exponential term representing aging dissipation and multiplying it by the thermal resistance factor and the set irradiance, the difference in junction temperature change that would inevitably be caused by forcibly performing a pure current compensation operation is calculated in advance. The mathematical formula for this step is expressed as: In the above formula, This represents the predicted junction temperature rise, characterizing the additional increase in the internal junction temperature of the semiconductor expected after the application of a compensation drive. The thermal resistance factor represents the hardware thermal resistance factor. It should be clarified that the thermal resistance factor here is not the traditional absolute thermal resistance, but a 'system-level equivalent thermal mapping factor' after the device structure has been reduced and packaged. It characterizes the thermal conduction resistance constant of the heat dissipation system of the phototherapy device that has solidified over time. This represents the set irradiance, characterizing the target phototherapy intensity expected to be achieved clinically. This represents the attenuation coefficient after verification, extracted through feature filtering. Hardware thermal resistance factor. In this specific embodiment, it is set to a constant value of 0.5℃·cm. 2 •nm / μW is μW / cm² used to adapt the irradiance of phototherapy equipment. 2 The system-level equivalent thermal mapping normalization process performed at / nm is determined by fitting the decay curve of the heat dissipation system during the maintenance cycle based on the temperature rise load calibration experiment before the equipment leaves the factory. Continuing with the actual data example of the aforementioned process, the irradiance is set at this time. The attenuation coefficient is 35 after verification. The value is 0.925. First, the exponential term representing aging dissipation is calculated. Its value is approximately 1.0779. Substituting these values ​​into the formula, the predicted junction temperature rise is calculated. Equals 0.5 multiplied by 35 multiplied by 1.0779, the final result is approximately 18.86 degrees Celsius. This value precisely reveals that under the current aging conditions, performing compensation would generate an additional severe heat load of nearly 19 degrees Celsius.

[0038] Based on the predicted junction temperature rise, a nonlinear correction factor is determined. Accordingly, the simple temperature difference calculated in the preceding steps cannot be directly used for issuing control commands; it needs to be decoupled and transformed into a physical loss weight that can accurately measure the effective medical photon retention rate. This successfully extracts the purity of medical photons that truly have a yellowing-reducing effect under complex thermal conditions, accurately quantifying the dual losses of luminous efficiency drop and wavelength drift. In specific implementation, the algorithm receives the output predicted junction temperature rise and introduces a pre-set thermal drop constant and spectral shift constant based on the photothermal coupling equation. The system uses fractional terms to fit the decay law of quantum efficiency with increasing temperature in semiconductor physics, while introducing a cosine function to map the biochemical mismatch rate caused by the redshift leading to spectral deviation from the optimal absorption peak of bilirubin, and finally performs a composite nonlinear product calculation. This calculation is achieved through the following formula: In the formula Representing the nonlinear correction factor, it characterizes the true effective photon attenuation loss rate after combining thermal degradation and spectral mismatch; Represents the thermal drop constant, which characterizes the intrinsic decay rate of the inherent quantum efficiency of the current LED light-emitting component as temperature increases; Represents the spectral shift constant, characterizing the conversion coefficient that transforms junction temperature changes into wavelength redshift and causes deviation from the biochemical absorption peak of bilirubin; This is the predicted junction temperature rise calculated in the previous step. Preset parameters. and The physical property constants are obtained by placing the same type of optical module in a constant temperature test chamber and performing tens of thousands of full-range scan regressions of photoelectric color parameters for the characteristic wavelength. In this embodiment, Set to 0.004. The value is set to 0.015 radians per degree Celsius. Using the previously calculated predicted junction temperature rise of 18.86 degrees Celsius as an example, we first calculate the heat depreciation fraction by dividing 1 by (1 plus 0.004 multiplied by 18.86), which is approximately 0.930. Then, we calculate the cosine term of the spectral mismatch, i.e., cos(0.015 × 18.86), which is approximately 0.960. Multiplying the two yields the nonlinear correction factor. The value is 0.893. This indicator means that, after the secondary loss due to the superimposed thermo-optical coupling field, the actual photon efficacy rate that can exert a therapeutic effect in the newborn is only 89.3%.

[0039] Based on the nonlinear correction factor and the verified attenuation coefficient, a multidimensional nonlinear control command is synthesized from the set irradiance to obtain the compensated irradiance control command. It should be understood that, after the calculations in the pre-processing stage, the primary physical losses caused by the physical aging of the hardware foundation and the secondary thermal losses caused by high-load thermo-optical coupling have been completely decoupled and extracted. At this point, the two need to be finally aggregated to generate a precise command that avoids overcompensation. This completely abandons the shallow and dangerous logic of simply increasing the drive current in the traditional approach, and constructs an absolutely safe adaptive underlying drive barrier through joint weighted mapping, cutting off the vicious cycle of hardware lifespan termination caused by overcompensation from the underlying logic. In the implementation process, the original clinically issued set irradiance is extracted as the numerator. The verified attenuation coefficient, representing the initial aging of the hardware foundation, is multiplied and fused with the nonlinear correction factor, representing the dynamic thermal performance degradation. This joint weight is used as the denominator for nonlinear comprehensive weighted mapping to generate the final signal value used to take over the underlying drive module. Its mathematical expression is: In this formula, The irradiance control command after compensation represents the adaptive target signal that is ultimately sent directly to the underlying photoelectric constant current unit. Represents the original target irradiance set by the transparent transmission; This represents the nonlinear correction factor calculated above. A coherent data stream is embedded into this synthesis formula to set the irradiance. The attenuation coefficient is 35 after verification. The nonlinear correction factor is 0.925. The solution is calculated to be 0.893. At this point, the comprehensive loss coefficient on the denominator side becomes 0.925 multiplied by 0.893, which equals 0.826. Dividing 35 by this coefficient yields the compensated irradiance control command. The value is approximately 42.37 units. Compared to the initial estimate of 37.84 units considering only physical aging, this control architecture accurately depicts that, to meet the clinically required effective therapeutic effect of 35 units, a drive control command of 42.37 units needs to be issued to the underlying layer, and this value must be strictly determined to be within the thermal safety threshold to fully compensate for the difference in photon loss throughout the treatment process caused by simple photon decay and secondary thermal redshift. Subsequently, this compensated control command is sent to the drive hardware for real-time dynamic output.

[0040] In summary, the dynamic irradiance control method for phototherapy devices based on the neonatal jaundice index, as described in this application, addresses the shortcomings of the prior art, where hardware aging and light source attenuation lead to deviations from expected treatment effects. This method uses a pre-set phototherapy dynamics model, combined with the individual baseline characteristics of the infant, to predict the ideal bilirubin reduction rate. During treatment, the actual bilirubin reduction rate of the infant is monitored and calculated in real time, and compared with the expected target value using a physiological response dimension difference. This difference is used to infer the current physical attenuation coefficient of the hardware, thereby quantifying the power loss caused by device aging. Finally, this attenuation coefficient is applied to the original set value, and irradiance compensation commands are dynamically generated to correct the hardware's drive output in real time. This method breaks the open-loop limitation of traditional devices relying solely on set parameters, achieving automatic closed-loop control from clinical effect evaluation to hardware power compensation. It fundamentally eliminates the impact of hardware performance fluctuations on treatment accuracy, ensuring the continuous stability and high efficiency of phototherapy irradiance.

[0041] Figure 4 This is a block diagram of a phototherapy device irradiance dynamic control system based on the neonatal jaundice index, according to an embodiment of this application. Figure 4 As shown, the phototherapy device irradiance dynamic control system 100 based on neonatal jaundice index according to an embodiment of this application includes: a neonatal subject data acquisition module 110, used to acquire baseline physiological characteristic data of the neonatal subject and set irradiance; an expected bilirubin decrease rate calculation module 120, used to extract a neonatal baseline feature set from the neonatal subject's baseline physiological characteristic data, and perform expected response estimation on the set irradiance based on the neonatal baseline feature set to obtain the expected bilirubin decrease rate under ideal output conditions without device aging; and an actual bilirubin decrease rate calculation module 130, used to continuously monitor the baseline physiological characteristic data, and when a new round of bilirubin decrease rate is detected... When measuring bilirubin, the initial bilirubin concentration and initial timestamp from the neonatal baseline feature set are used to perform differential verification and matching on the newly acquired current bilirubin concentration and current timestamp to obtain the actual bilirubin decrease rate of the newborn; the attenuation coefficient analysis module 140 is used to perform hardware attenuation coefficient inverse inference on the actual bilirubin decrease rate and the expected bilirubin decrease rate of the newborn to obtain the verified attenuation coefficient; the irradiance control command generation module 150 is used to perform irradiance dynamic compensation on the set irradiance based on the verified attenuation coefficient to obtain the compensated irradiance control command, and the compensated irradiance control command is sent to the hardware LED driver module.

[0042] Here, the specific operation of each step in the above-mentioned dynamic control system for the irradiance of phototherapy equipment based on the neonatal jaundice index has been referenced above. Figures 1 to 3The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index has been described in detail, and therefore, its repeated description will be omitted.

[0043] The steps, measures, and solutions in the various operations, methods, and processes discussed in this application can be alternated, modified, combined, or deleted. Furthermore, other steps, measures, and solutions in the various operations, methods, and processes discussed in this application can also be alternated, modified, rearranged, decomposed, combined, or deleted. Furthermore, steps, measures, and solutions in the prior art that are identical to those disclosed in this application can also be alternated, modified, rearranged, decomposed, combined, or deleted. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combinations of these technical features do not contradict each other, they should be considered within the scope of this specification. The above-described embodiments are merely illustrative of several implementation methods of the present application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of protection of the embodiments of the present application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of the present application, and these all fall within the scope of protection of the embodiments of the present application. Therefore, the scope of protection of the embodiments of the present application should be determined by the appended claims. As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and detail can be made without departing from the spirit and scope of the present invention as defined in the appended claims.

[0044] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for dynamically adjusting the irradiance of a phototherapy device based on the neonatal jaundice index, characterized in that, include: S1: Obtain baseline physiological characteristic data of newborn subjects and set irradiance; S2: Extract the neonatal baseline feature set from the baseline physiological characteristic data of the newborn subjects, and estimate the expected response to the set irradiance based on the neonatal baseline feature set to obtain the expected bilirubin reduction rate under the ideal output state of the equipment without aging. S3: Continuously monitor baseline physiological characteristic data. When a new round of bilirubin measurement value is detected, use the initial bilirubin concentration and initial timestamp in the newborn baseline characteristic set to perform differential verification and matching on the newly acquired current bilirubin concentration and current timestamp to obtain the actual rate of bilirubin decrease in the newborn. S4: The actual bilirubin decrease rate and the expected bilirubin decrease rate in newborns are inferred in reverse by hardware attenuation coefficient to obtain the verified attenuation coefficient. S5: Based on the verified attenuation coefficient, the set irradiance is dynamically compensated to obtain the compensated irradiance control command, which is then sent to the hardware LED driver module.

2. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 1, characterized in that, Baseline physiological data for newborns include initial bilirubin concentration, initial timestamp, infant weight, and risk factor weighted score.

3. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 1, characterized in that, Also includes: A safety assessment is performed on the attenuation coefficient after verification. When the attenuation coefficient after verification is lower than the preset limit tolerance threshold, a hardware maintenance warning signal is triggered.

4. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 2, characterized in that, Step S2 includes: By using analytical extraction methods, the baseline physiological characteristic data of newborns are stripped and encapsulated to extract initial bilirubin concentration, initial timestamp, infant weight and risk factor weighted scores, and packaged to generate a baseline feature set; Based on the phototherapy kinetic response model, the baseline feature set and the set irradiance are normalized and the expected response is calculated to obtain the expected bilirubin reduction rate.

5. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 2, characterized in that, Step S3 includes: performing differential verification and matching between the newly acquired current bilirubin concentration and the current timestamp using the following formula to obtain the actual rate of bilirubin decrease in the newborn, wherein the formula is: in, This represents the initial bilirubin concentration. This represents the current bilirubin concentration. This is the current timestamp for the current bilirubin concentration; This is the initial timestamp for the initial bilirubin concentration. This represents the actual rate of decrease in bilirubin levels in newborns.

6. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 1, characterized in that, Step S4 includes: The hardware-based inverse inference of the actual bilirubin decrease rate and the expected bilirubin decrease rate in newborns is performed to obtain the original inferred inference inference coefficient. The original inferred attenuation coefficients are pushed into the historical inferred attenuation coefficient queue, and the updated historical inferred queue is subjected to moving average filtering to obtain the verified attenuation coefficients.

7. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 6, characterized in that, The actual rate of bilirubin decrease in newborns and the expected rate of bilirubin decrease are compared using hardware-based inverse inference of the decay coefficient to obtain the original inferred decay coefficient, including: The physiological response ratio was obtained by comparing the actual rate of bilirubin decrease in newborns with the expected rate of bilirubin decrease. The physiological response ratio and the set irradiance are physically attenuated to obtain the original inferred attenuation coefficient.

8. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 7, characterized in that, The original inferred attenuation coefficients are pushed into the historical inferred attenuation coefficient queue. A moving average filter is then applied to the updated historical inferred queue to obtain the validated attenuation coefficients, including: Data push and out-of-bounds removal are performed on the original inferred decay coefficient and historical inferred decay coefficient queues to obtain the historical inferred queue that has completed the state update operation. Based on the moving average filtering module, physiological noise is filtered out and physical characteristics are solidified in the historical inference queue to obtain the attenuation coefficient after verification.

9. The method for dynamic control of irradiance of phototherapy equipment based on neonatal jaundice index according to claim 1, characterized in that, Step S5 includes: Based on the verified attenuation coefficient and hardware thermal resistance factor, the junction temperature rise is prospectively estimated for a given irradiance to obtain the predicted junction temperature rise. Based on the predicted junction temperature rise, the nonlinear correction factor is determined; Based on the nonlinear correction factor and the verified attenuation coefficient, a multidimensional nonlinear control command is synthesized for the set irradiance to obtain the compensated irradiance control command.

10. A dynamic irradiance control system for phototherapy equipment based on neonatal jaundice index, characterized in that, include: The newborn subject data acquisition module is used to acquire baseline physiological characteristic data of newborn subjects and set irradiance. The expected bilirubin decrease rate calculation module is used to extract the neonatal baseline feature set from the baseline physiological characteristic data of the newborn subjects, and to estimate the expected response to a set irradiance based on the neonatal baseline feature set in order to obtain the expected bilirubin decrease rate under the ideal output state of the equipment without aging. The actual bilirubin decrease rate calculation module is used to continuously monitor baseline physiological characteristic data. When a new round of bilirubin measurement value is detected, the module uses the initial bilirubin concentration and initial timestamp in the neonatal baseline characteristic set to perform differential verification and matching on the newly acquired current bilirubin concentration and current timestamp to obtain the actual bilirubin decrease rate of the neonate. The attenuation coefficient analysis module is used to perform hardware attenuation coefficient inverse inference between the actual bilirubin decrease rate and the expected bilirubin decrease rate in newborns to obtain the verified attenuation coefficient. The irradiance control command generation module is used to perform dynamic irradiance compensation on the set irradiance based on the verified attenuation coefficient to obtain the compensated irradiance control command. The compensated irradiance control command is sent to the hardware LED driver module.

Images