Intelligent bladder irrigation dynamic regulation method based on multi-modal physiological signal fusion

Abstract

This invention provides an intelligent dynamic control method for bladder irrigation based on multimodal physiological signal fusion, belonging to the field of multimodal information technology. It involves acquiring and fusing signals from bladder wall mechanoreceptors, core body temperature, and lower abdominal skin electrical activity to extract multimodal features and calculate a physiological balance index to assess the overall system state. Based on this index, physiological regulation patterns are identified and a combined temperature and viscosity regulation strategy is matched. During execution, the regulation parameters are dynamically corrected by monitoring the response trajectory to drive the control device to process the irrigation fluid. This invention achieves personalized, adaptive, and precise control of the bladder irrigation process, improving treatment safety and comfort.

Description

Technical Field

[0001] This invention relates to the field of multimodal information technology, and in particular to a method for intelligent dynamic control of bladder irrigation based on multimodal physiological signal fusion. Background Technology

[0002] Bladder irrigation is an important clinical procedure in urological surgery and in the treatment of certain diseases. Its core purpose is to remove foreign bodies such as blood clots, mucus, and tissue debris from the bladder, maintain unobstructed drainage, and prevent infection. Traditional bladder irrigation typically involves continuous or intermittent irrigation with physiological saline or a specific irrigation solution at a fixed flow rate and temperature (close to body temperature). The operation mainly relies on manual control by experienced medical staff or automated control using pre-set simple programs, such as starting and stopping irrigation based on the balance of irrigation fluid inflow and outflow or timed rules. In order to improve the safety and comfort of irrigation, some studies have attempted to introduce single-dimensional physiological signals for auxiliary monitoring, such as monitoring intrabladder pressure to prevent overfilling or monitoring irrigation fluid temperature to maintain a constant level.

[0003] The aforementioned conventional practices have significant limitations. The bladder, as an organ regulated by a complex autonomic nervous system, is not determined by a single parameter, but rather by the synergistic effect of multiple systems. Relying solely on irrigation pressure, flow rate, or a single temperature parameter for regulation cannot comprehensively and sensitively reflect the combined impact of the irrigation process on the patient's overall physiological state, particularly on local nerve reflexes of the bladder wall, the thermoregulatory system, and the level of sympathetic nerve stress. Fixed irrigation strategies lack individual adaptability and struggle to address the diverse physiological responses of different patients due to differences in surgical trauma, pain sensitivity, and autonomic nervous function. This can lead to low irrigation efficiency or complications such as patient discomfort, bladder spasms, and abnormal autonomic reflexes, affecting the treatment process and the patient's recovery experience. Therefore, there is an urgent need for a bladder irrigation method that can more comprehensively perceive the patient's physiological state and achieve precise, adaptive regulation. Summary of the Invention

[0004] This invention provides a method and system for intelligent dynamic control of bladder irrigation based on multimodal physiological signal fusion, which can solve the problems in the prior art.

[0005] A first aspect of this invention provides a method for intelligent dynamic control of bladder irrigation based on multimodal physiological signal fusion, comprising: Acquire bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals during bladder irrigation. The pulse sequence of the bladder wall mechanoreceptor signal is decoded to extract the pulse firing frequency and pulse interval distribution features. The core body temperature signal is trend-separated to extract the body temperature baseline drift and the amplitude of periodic body temperature fluctuations. The lower abdominal electrodermal signal is subjected to time-domain statistical analysis to extract the skin conductance level and the number of skin electrical responses. Each feature is mapped to a standardized feature space. The dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is calculated. Based on the physiological balance index, the patient's current physiological regulation mode is identified. For the identified physiological regulation mode, a matching temperature and viscosity joint regulation strategy is selected from a preset multi-mode regulation strategy library. During the execution of the temperature and viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored. By calculating the deviation between the actual response trajectory and the expected response trajectory, the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected. The flushing fluid is processed by the temperature control device and viscosity control device according to the corrected adjustment parameters and then infused into the bladder.

[0006] Pulse sequence decoding of bladder wall mechanoreceptor signals, trend separation of core body temperature signals, and time-domain statistical analysis of lower abdominal electrodermal signals were performed, including: The bladder wall mechanoreceptor signal is preprocessed with bandpass filtering to remove baseline drift and high-frequency noise interference. A threshold detection method is used to identify effective pulse events. The pulse firing frequency within a unit time window is statistically analyzed. The mean and standard deviation of the interval between adjacent pulses are calculated as pulse interval distribution characteristics. The core body temperature signal is filtered by a moving average filter to extract long-period trend components as the body temperature baseline. The body temperature fluctuation component is obtained by the difference between the core body temperature signal and the body temperature baseline. The body temperature fluctuation component is subjected to spectral analysis to identify the dominant period and extract the fluctuation amplitude corresponding to the dominant period. The average conductivity value of the lower abdominal skin electrical signal within a preset time period is calculated as the skin electrical conductivity level. The abrupt change point of the skin electrical signal is detected by the first-order difference method. The number of occurrences of the abrupt change point is counted as the number of skin electrical response. The peak amplitude and rise time of each skin electrical response are recorded. The extracted pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and number of skin conductance responses are used as the original feature set.

[0007] Each feature is mapped to a standardized feature space. The dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is calculated, including: The pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and skin conductance response frequency are normalized respectively, and the numerical range of each feature is mapped to a unified standardized feature space. A multi-channel time series matrix is ​​established, wherein the rows of the multi-channel time series matrix correspond to different standardized features, the columns correspond to continuous time sampling points, and each matrix element records the standardized value of a specific feature at a specific time. For any two feature channels in the multi-channel time series matrix, calculate the time delay cross-correlation coefficient between them. The time delay cross-correlation quantifies the influence of one feature on another feature under different time delays. A temporal correlation matrix is ​​constructed, wherein the rows and columns of the temporal correlation matrix correspond to each normalized feature, and the value of the matrix element is the maximum time delay cross-correlation coefficient between the corresponding feature pairs. The temporal correlation matrix characterizes the dynamic coupling strength and temporal dependency between multimodal signals. The temporal correlation matrix is ​​decomposed to extract principal components that represent the overall coordinated mode of multimodal signals, and the physiological balance index is calculated based on the principal components.

[0008] Matrix decomposition is performed on the temporal correlation matrix to extract principal components representing the overall coordinated mode of multimodal signals. Based on these principal components, a physiological balance index is calculated, including: Singular value decomposition is performed on the time-series correlation matrix to obtain a left singular vector matrix, a singular value diagonal matrix, and a right singular vector matrix. The singular value sequence is extracted from the singular value diagonal matrix, and the decreasing slope of adjacent singular values ​​in the singular value sequence is calculated. Identify the inflection points in the descending slope sequence where the absolute value of the slope is less than a preset slope threshold. Determine the singular values ​​before the inflection point and their corresponding left singular vectors as effective principal components. The number of effective principal components is directly determined by the index of the inflection point. The real-time values ​​of each standardized feature at the current moment are extracted from the multi-channel time series matrix. The real-time values ​​are then multiplied by matrix-vector multiplication with each column vector in the left singular vector matrix corresponding to the effective principal components to obtain the component values ​​of each effective principal component space dimension at the current moment. A piecewise nonlinear mapping function is constructed, and the component values ​​of each effective principal component space dimension are respectively input into the piecewise nonlinear mapping function. When the component value is located in the preset central interval, a linear mapping is applied, and when the component value is not in the central interval, a saturation suppression mapping is applied. The output values ​​of each dimension after mapping are weighted and summed according to the square root of the corresponding singular value. The weighted summation result is normalized to a preset numerical range as the physiological balance index.

[0009] Based on the physiological balance index, the patient's current physiological regulation mode is identified, and a matching temperature-viscosity combined regulation strategy is selected from a pre-set multi-mode regulation strategy library, including: A classification rule for physiological regulation modes is pre-established. The classification rule for physiological regulation modes defines the numerical range and rate of change characteristics of physiological balance index corresponding to different physiological regulation modes. The physiological regulation modes include thermal balance maintenance mode, stress compensation mode and imbalance early warning mode. Obtain the rate of change of the physiological balance index at the current moment within a preset time window; The physiological balance index is compared with the numerical range defined in the physiological regulation mode classification rules, and the rate of change is matched with the rate of change feature defined in the physiological regulation mode classification rules. Based on the comparison and matching results, the patient's current physiological regulation mode is identified. Access the multi-mode regulation strategy library, which stores temperature and viscosity joint regulation strategies corresponding to each physiological regulation mode. Each temperature and viscosity joint regulation strategy specifies the adjustment range of the flushing fluid temperature, the adjustment range of the flushing fluid viscosity, and the timing relationship between the two. Based on the identified physiological regulation pattern, a temperature-viscosity joint regulation strategy that matches the physiological regulation pattern is retrieved from the multi-mode regulation strategy library and selected as the current regulation strategy to be executed.

[0010] During the execution of the temperature-viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored. By calculating the deviation between the actual response trajectory and the expected response trajectory, the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected, including: The current physiological balance index is obtained as the starting value. Based on the difference between the target physiological balance index and the starting value, the expected physiological balance index value at each sampling time is calculated in a linearly increasing manner over time. The expected physiological balance index values ​​are stored in chronological order as an array of expected response trajectories. The measured values ​​of the current physiological balance index are collected at fixed intervals. The measured values ​​at each sampling time are stored in chronological order as an actual response trajectory array. The timestamps of the actual response trajectory array correspond completely with those of the expected response trajectory array. Extract the expected value corresponding to the current time from the expected response trajectory array, calculate the difference between the current measured value and the expected value to obtain the deviation at the current time, count the deviation values ​​of the three most recent consecutive sampling times, and calculate the arithmetic mean of the three deviation values ​​as the average deviation; Determine whether the sign of the current deviation is consistent with that of the average deviation. If the signs are consistent and the absolute value of the average deviation is greater than a preset deviation threshold, divide the average deviation by the difference between the target physiological balance index and the initial value to obtain the normalized deviation ratio. Multiply the temperature adjustment range by a first correction factor to update the new temperature adjustment range. Multiply the viscosity adjustment range by a second correction factor to update the new viscosity adjustment range. The product of the first correction factor and the second correction factor is a fixed constant. The first correction factor increases monotonically as the normalized deviation ratio increases.

[0011] A second aspect of the present invention provides an intelligent bladder irrigation dynamic control system based on multimodal physiological signal fusion, comprising: The signal acquisition unit is used to acquire the bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals of the patient during bladder irrigation. The feature extraction unit is used to decode the pulse sequence of the bladder wall mechanoreceptor signal, extract the pulse firing frequency and pulse interval distribution features, perform trend separation on the core body temperature signal, extract the body temperature baseline drift and the amplitude of body temperature periodic fluctuations, perform time-domain statistical analysis on the lower abdominal skin electrical signal, extract skin conductance level and skin electrical response frequency, map each feature to a standardized feature space, capture the dynamic coupling relationship of multimodal signals by constructing a temporal correlation matrix between features, and calculate the physiological balance index characterizing the overall state of the bladder-nerve-body temperature system. The strategy control unit is used to identify the patient's current physiological regulation mode based on the physiological balance index, select a matching temperature and viscosity joint regulation strategy from a preset multi-mode regulation strategy library for the identified physiological regulation mode, continuously monitor the response trajectory of the physiological balance index during the execution of the temperature and viscosity joint regulation strategy, and dynamically correct the temperature regulation amplitude and the viscosity regulation amplitude by calculating the deviation between the actual response trajectory and the expected response trajectory. The output unit is used to drive the temperature control device and viscosity control device to process the flushing fluid according to the corrected adjustment parameters and then infuse it into the bladder.

[0012] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0014] This method comprehensively covers key physiological dimensions reflecting bladder filling pressure, autonomic nervous system response, and core body temperature regulation by simultaneously acquiring bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin conductance signals, providing a multimodal data foundation for subsequent precise analysis. Pulse sequence decoding, trend separation, and time-domain statistical analysis were performed on the three types of signals to extract deep features such as pulse firing frequency, baseline temperature drift, and skin conductance levels. These features can more fundamentally characterize changes in physiological state. Mapping the extracted features to a standardized space and constructing a temporal correlation matrix effectively captures the dynamic coupling relationships between different physiological signal systems, thereby calculating a physiological balance index that comprehensively reflects the overall coordinated state of the bladder-nervous-thermal system. This index can more comprehensively and stably assess the patient's real-time physiological load than a single signal indicator.

[0015] Based on the physiological balance index, the system identifies the patient's current physiological regulation pattern, distinguishing different physiological compensatory states, such as a sympathetic stress state or a relatively balanced adaptive state. For each identified pattern, a combined temperature and viscosity regulation strategy is intelligently matched from a pre-set strategy library, shifting from a "one-size-fits-all" approach to personalized, adaptive regulation. During strategy execution, the system continuously monitors the physiological balance index's response trajectory to the regulatory actions, and dynamically corrects deviations from the expected trajectory. This imbues the regulation process with self-learning and adaptive capabilities, continuously optimizing the regulatory amplitude based on the patient's real-time physiological feedback to ensure the intervention remains within the most appropriate range. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the intelligent dynamic control method for bladder irrigation based on multimodal physiological signal fusion. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0019] Figure 1 This is a flowchart illustrating a dynamic control method for intelligent bladder irrigation based on multimodal physiological signal fusion, as shown below. Figure 1 As shown, the method includes: Acquire bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals during bladder irrigation. The pulse sequence of the bladder wall mechanoreceptor signal is decoded to extract the pulse firing frequency and pulse interval distribution features. The core body temperature signal is trend-separated to extract the body temperature baseline drift and the amplitude of periodic body temperature fluctuations. The lower abdominal electrodermal signal is subjected to time-domain statistical analysis to extract the skin conductance level and the number of skin electrical responses. Each feature is mapped to a standardized feature space. The dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is calculated. Based on the physiological balance index, the patient's current physiological regulation mode is identified. For the identified physiological regulation mode, a matching temperature and viscosity joint regulation strategy is selected from a preset multi-mode regulation strategy library. During the execution of the temperature and viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored. By calculating the deviation between the actual response trajectory and the expected response trajectory, the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected. The flushing fluid is processed by the temperature control device and viscosity control device according to the corrected adjustment parameters and then infused into the bladder.

[0020] For example, pulse sequence decoding of bladder wall mechanoreceptor signals, trend separation of core body temperature signals, and time-domain statistical analysis of lower abdominal electrodermal signals include: The bladder wall mechanoreceptor signal is preprocessed with bandpass filtering to remove baseline drift and high-frequency noise interference. A threshold detection method is used to identify effective pulse events. The pulse firing frequency within a unit time window is statistically analyzed. The mean and standard deviation of the interval between adjacent pulses are calculated as pulse interval distribution characteristics. The core body temperature signal is filtered by a moving average filter to extract long-period trend components as the body temperature baseline. The body temperature fluctuation component is obtained by the difference between the core body temperature signal and the body temperature baseline. The body temperature fluctuation component is subjected to spectral analysis to identify the dominant period and extract the fluctuation amplitude corresponding to the dominant period. The average conductivity value of the lower abdominal skin electrical signal within a preset time period is calculated as the skin electrical conductivity level. The abrupt change point of the skin electrical signal is detected by the first-order difference method. The number of occurrences of the abrupt change point is counted as the number of skin electrical response. The peak amplitude and rise time of each skin electrical response are recorded. The extracted pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and number of skin conductance responses are used as the original feature set.

[0021] In the signal preprocessing stage, bladder wall mechanoreceptor signals acquired by implanted or surface sensors typically contain effective physiological information in the range of 0.5-150Hz, but are superimposed with low-frequency baseline drift (0.1-0.5Hz) caused by respiratory movements and high-frequency noise (above 150Hz) generated by electromyography interference. A third-order Butterworth bandpass filter is used to process the raw signal, setting the passband to 1-120Hz to filter out out-of-band interference components. The filtered signal undergoes full-wave rectification and envelope extraction. A dynamic threshold is set to 1.5 times the root mean square value of the signal. A valid pulse event is identified when the signal amplitude exceeds this threshold and the duration is greater than 8 milliseconds. A 5-second sliding time window is selected, and the pulse firing frequency is obtained by dividing the total number of pulses detected within the window by the time window length, expressed in pulses per second. For the extraction of pulse interval distribution characteristics, the peak times of continuously detected pulses are recorded, and the time difference between adjacent peaks is calculated to form an interval sequence. The arithmetic mean of this sequence reflects the rhythmicity of pulse firing, while the standard deviation quantifies the stability of the rhythm. The smaller the standard deviation, the more regular the activation state of the mechanoreceptor.

[0022] Trend separation of the core body temperature signal was achieved using a moving average filter with a window length of 300 seconds. This window length effectively suppresses the influence of short-cycle physiological activities such as respiration and heartbeat, preserving the slow trend of body temperature changes. The filter output is the baseline component of body temperature. The original body temperature signal is subtracted point by point from the corresponding baseline value to obtain the detrended body temperature fluctuation component. A fast Fourier transform is performed on this fluctuation component to convert the time-domain signal to the frequency domain for analysis. The frequency component with the highest energy is searched within the frequency range of 0.005-0.05Hz. The period corresponding to this frequency is defined as the dominant period, which usually reflects the characteristic oscillation period of the neural regulatory system. The amplitude spectrum value at the dominant frequency is extracted as the periodic fluctuation amplitude of body temperature. An increase in this parameter often indicates enhanced sympathetic nerve activity. The body temperature baseline drift is obtained by calculating the difference between the current baseline value and the initial baseline value. A continuously increasing drift may indicate an inflammatory response or changes in metabolic state.

[0023] The lower abdominal skin electrical activity (SEA) signal reflects the sympathetic nervous system's regulation of sweat gland secretion. Signal acquisition was performed using silver-silver chloride electrodes with a 2 cm electrode spacing. The arithmetic mean of the conductance values ​​at all sampling points over a 60-second time period was calculated as the skin conductance level; an increase in this baseline value indicates enhanced overall activation of the sympathetic nervous system. To detect rapid neural response events, a first-order difference operation was performed on the SEA signal. A valid SEA response was defined as a difference value exceeding 0.05 microsiemens followed by a monotonically increasing trend within the next 2 seconds. The number of responses meeting the criteria within the preset observation period was counted; this frequency corresponds to the patient's stress or discomfort level. The rise time and peak amplitude of each SEA response were also recorded. Rapid responses with a rise time of less than 1 second are usually associated with acute stimulation, while the peak amplitude quantifies the response intensity.

[0024] The nine parameters extracted above—pulse firing frequency, mean pulse interval, standard deviation of pulse interval, body temperature baseline drift, body temperature fluctuation amplitude, skin conductance level, number of skin conductance responses, peak amplitude of response, and rise time—form the original feature set, providing a data foundation for subsequent feature fusion and physiological state assessment.

[0025] For example, each feature is mapped to a standardized feature space, and the dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is then calculated, including: The pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and skin conductance response frequency are normalized respectively, and the numerical range of each feature is mapped to a unified standardized feature space. A multi-channel time series matrix is ​​established, wherein the rows of the multi-channel time series matrix correspond to different standardized features, the columns correspond to continuous time sampling points, and each matrix element records the standardized value of a specific feature at a specific time. For any two feature channels in the multi-channel time series matrix, calculate the time delay cross-correlation coefficient between them. The time delay cross-correlation quantifies the influence of one feature on another feature under different time delays. A temporal correlation matrix is ​​constructed, wherein the rows and columns of the temporal correlation matrix correspond to each normalized feature, and the value of the matrix element is the maximum time delay cross-correlation coefficient between the corresponding feature pairs. The temporal correlation matrix characterizes the dynamic coupling strength and temporal dependency between multimodal signals. The temporal correlation matrix is ​​decomposed to extract principal components that represent the overall coordinated mode of multimodal signals, and the physiological balance index is calculated based on the principal components.

[0026] The collected pulse firing frequency, pulse interval distribution characteristics, baseline temperature drift, periodic temperature fluctuation amplitude, skin conductance level, and number of skin conductance responses were normalized. The normal physiological range for pulse firing frequency is 5 to 30 times per second. Normalization was performed by dividing by the maximum value of this range, 30, to map the values ​​to the 0-1 interval. The pulse interval distribution characteristics were characterized using standard deviation, with a physiological fluctuation range of 10 to 100 milliseconds. Z-score normalization was used, calculated by subtracting the mean from the current value and then dividing by the standard deviation. The observed baseline temperature drift ranged from -0.5 to +0.5 degrees Celsius. Normalization was performed by adding 0.5 and then dividing by 1.0. The typical range for periodic temperature fluctuation amplitude is 0.1 to 0.5 degrees Celsius. A min-maximum normalization method was used to linearly map the values ​​to the 0-1 interval. The measurement range of skin conductance was 2 to 20 microsiemens. After logarithmic transformation, normalization was performed; logarithmic transformation can compress features with large numerical ranges. The normal range for the number of skin conductance responses within a 10-minute observation window was 0 to 15, which was directly normalized by dividing by 15. After normalization, all features have the same numerical scale, eliminating the influence of differences in the original dimensions of different physiological signals on subsequent analysis.

[0027] A multi-channel time series matrix M was constructed, with dimensions of 6 rows and N columns. The 6 rows correspond to the 6 standardized features mentioned above, and the N columns correspond to consecutive time sampling points. The sampling frequency was set to 10 times per second, with each sampling point spaced 100 milliseconds apart. The matrix element M(i,j) represents the standardized value of the i-th feature at time j. To ensure the effectiveness of the time series analysis, a time window length of 60 seconds was chosen, corresponding to 600 sampling points. This window length is sufficient to capture the periodic variation patterns during bladder filling.

[0028] For any two feature channels in a multi-channel time series matrix, the time-delay cross-correlation coefficient is calculated. Feature channels i and j are selected, and the search range of the time delay τ is set to -5 seconds to +5 seconds, with a step size of 0.1 seconds. For each delay value τ, the time series of feature j is shifted by τ seconds, and then the Pearson correlation coefficient between the shifted feature j and feature i is calculated. The cross-correlation function curve is obtained by iterating through all delay values. The time-delay cross-correlation coefficient reflects the degree of influence of changes in feature j on feature i after a specific time delay. For example, changes in the mechanoreceptor signal of the bladder wall may trigger a response in the electrodermal signal after 1 to 2 seconds; the cross-correlation coefficient corresponding to this time delay reaches its peak, characterizing the temporal characteristics of nerve conduction and autonomic nervous system response.

[0029] Construct a temporal correlation matrix R, which is a 6x6 symmetric matrix. Each matrix element R(i,j) is defined as the maximum absolute value of the cross-correlation coefficients of all time delays between feature i and feature j. The diagonal elements R(i,j) are set to 1, indicating that a feature is perfectly correlated with itself. The off-diagonal elements R(i,j) range from 0 to 1; values ​​closer to 1 indicate stronger dynamic coupling between the two features. The temporal correlation matrix fully characterizes the network of temporal dependencies between the six normalized features. The asymmetry of the matrix stems from the causal temporal relationships between different features; that is, the delay time at which feature i affects feature j may differ from the delay time at which feature j affects feature i.

[0030] Singular value decomposition is performed on the time series correlation matrix, decomposing matrix R into... The form is given by, where U and V are orthogonal matrices. The matrix is ​​a diagonal matrix. The elements of the diagonal matrix are arranged in descending order, and the eigenvectors corresponding to the three largest singular values ​​constitute the principal component space. Projecting the original six-dimensional eigenvectors onto this three-dimensional principal component space yields a dimensionality-reduced representation of the co-modal patterns. The physiological balance index is defined as the product of the variance contribution rate of the first principal component and the mean of all elements in the temporal correlation matrix, with a value ranging from 0 to 1. When the physiological balance index is higher than 0.75, it indicates that the multimodal signals exhibit a highly coordinated steady-state pattern; when the index is lower than 0.45, it indicates an imbalance trend in the bladder-nervous-thermal system, requiring active intervention to regulate flushing parameters.

[0031] For example, matrix decomposition is performed on the temporal correlation matrix to extract principal components representing the overall coordinated mode of multimodal signals, and a physiological balance index is calculated based on the principal components, including: Singular value decomposition is performed on the time-series correlation matrix to obtain a left singular vector matrix, a singular value diagonal matrix, and a right singular vector matrix. The singular value sequence is extracted from the singular value diagonal matrix, and the decreasing slope of adjacent singular values ​​in the singular value sequence is calculated. Identify the inflection points in the descending slope sequence where the absolute value of the slope is less than a preset slope threshold. Determine the singular values ​​before the inflection point and their corresponding left singular vectors as effective principal components. The number of effective principal components is directly determined by the index of the inflection point. The real-time values ​​of each standardized feature at the current moment are extracted from the multi-channel time series matrix. The real-time values ​​are then multiplied by matrix-vector multiplication with each column vector in the left singular vector matrix corresponding to the effective principal components to obtain the component values ​​of each effective principal component space dimension at the current moment. A piecewise nonlinear mapping function is constructed, and the component values ​​of each effective principal component space dimension are respectively input into the piecewise nonlinear mapping function. When the component value is located in the preset central interval, a linear mapping is applied, and when the component value is not in the central interval, a saturation suppression mapping is applied. The output values ​​of each dimension after mapping are weighted and summed according to the square root of the corresponding singular value. The weighted summation result is normalized to a preset numerical range as the physiological balance index.

[0032] When performing singular value decomposition on the temporal correlation matrix, the matrix is ​​represented as a product of three matrices. The column vectors of the left singular vector matrix correspond to the orthogonal basis of the feature space; the diagonal elements of the singular value diagonal matrix reflect the contribution of each basis vector to the signal variation; and the column vectors of the right singular vector matrix represent the mode components of the time dimension. All singular values ​​are extracted from the singular value diagonal matrix in descending order, forming a monotonically decreasing numerical sequence. The difference between the i-th and (i+1)-th singular values ​​in this sequence is calculated and divided by the index difference to obtain the descent slope at each adjacent position.

[0033] The descending slope sequence is iterated, and the absolute value of each slope is compared with a preset slope threshold. The slope threshold is determined statistically based on experimental data and is usually set to 15% of the initial absolute value of the slope. When the absolute value of the slope at a certain position first falls below this threshold, that position is the inflection point. The signal components corresponding to all singular values ​​before the inflection point contain the main physiological information, while the singular values ​​at and after the inflection point mainly correspond to measurement noise and minor perturbations. If the inflection point position is denoted as k, then the first k singular values ​​and the first k columns of the left singular vector matrix constitute the effective principal component set.

[0034] At the current time t, standardized characteristic values ​​are read from three channels: bladder wall mechanoreceptors, core body temperature, and lower abdominal skin conductance. These values ​​include pulse firing frequency, pulse interval distribution characteristics, baseline temperature drift, amplitude of periodic temperature fluctuations, skin conductance level, and number of skin conductance responses. These six values ​​are arranged into a column vector in a fixed order. Matrix-vector multiplication is then performed on this column vector with columns 1 through k of the left singular vector matrix to obtain k scalar results. Each scalar represents the projection component of the current physiological state onto the corresponding principal component dimension.

[0035] A piecewise nonlinear mapping function is constructed to process each projection component. A pre-defined central interval boundary value is set from -0.8 to +0.8, corresponding to the normal fluctuation range of physiological equilibrium. When the value of a projection component falls within this interval, the mapping function directly outputs the original value using an identity transformation. When the value of a projection component exceeds +0.8, the mapping function transforms into a saturated function form, with the output value gradually approaching the upper limit of +1.2 as the input value increases. When the value of a projection component is below -0.8, the mapping function also adopts a saturated form, with the output value approaching the lower limit of -1.2. The mapping curve in the saturated region uses a transformation of the hyperbolic tangent function to achieve a smooth transition.

[0036] The k output values ​​after piecewise mapping are multiplied by the square root of their respective singular values. The square root of the singular value represents the magnitude of the standard deviation of each principal component and, as a weighting coefficient, can reasonably reflect the differences in the contribution of different dimensions to the overall state. The k weighted values ​​are summed to obtain the comprehensive evaluation value. This comprehensive evaluation value is then mapped to a numerical range of 0 to 100 through a linear transformation. The linear coefficients of the mapping formula are determined based on the maximum and minimum values ​​of the comprehensive evaluation value in the training dataset. The mapped value is the physiological balance index. The closer the index value is to 50, the more ideal the bladder-nervous-thermal system is in balance. The degree of deviation from 50 reflects the severity of the system imbalance, and the direction of the value indicates the type of imbalance.

[0037] For example, identifying the patient's current physiological regulation pattern based on the physiological balance index and selecting a matching temperature-viscosity combined regulation strategy from a pre-set multi-mode regulation strategy library includes: A classification rule for physiological regulation modes is pre-established. The classification rule for physiological regulation modes defines the numerical range and rate of change characteristics of physiological balance index corresponding to different physiological regulation modes. The physiological regulation modes include thermal balance maintenance mode, stress compensation mode and imbalance early warning mode. Obtain the rate of change of the physiological balance index at the current moment within a preset time window; The physiological balance index is compared with the numerical range defined in the physiological regulation mode classification rules, and the rate of change is matched with the rate of change feature defined in the physiological regulation mode classification rules. Based on the comparison and matching results, the patient's current physiological regulation mode is identified. Access the multi-mode regulation strategy library, which stores temperature and viscosity joint regulation strategies corresponding to each physiological regulation mode. Each temperature and viscosity joint regulation strategy specifies the adjustment range of the flushing fluid temperature, the adjustment range of the flushing fluid viscosity, and the timing relationship between the two. Based on the identified physiological regulation pattern, a temperature-viscosity joint regulation strategy that matches the physiological regulation pattern is retrieved from the multi-mode regulation strategy library and selected as the current regulation strategy to be executed.

[0038] Before activating the bladder irrigation dynamic control system, physiological regulation mode classification rules were pre-established in the system storage module. These rules were obtained through statistical analysis of data from over 100 clinical cases. For the thermal balance maintenance mode, the physiological balance index was defined as ranging from 0.65 to 0.85, with a corresponding rate of change characteristic of less than 0.02 / minute absolute value within a unit time window, indicating that the patient's bladder-nervous-thermal system was in a steady-state regulation phase. For the stress compensation mode, the physiological balance index was defined as ranging from 0.40 to 0.65 or 0.85 to 0.95, with a rate of change characteristic of 0.02 to 0.08 / minute, indicating that the patient's physiological regulation mechanism was in an active compensation phase. For the imbalance warning mode, the physiological balance index was defined as below 0.40 or above 0.95, with a rate of change characteristic exceeding 0.08 / minute, suggesting that the patient was at risk of regulatory decompensation.

[0039] The system continuously acquires the current physiological balance index from continuous sampling values ​​over the past 5-minute time window. Using a differential calculation method, the difference between the latest physiological balance index and the value 5 minutes ago is divided by the time interval of 300 seconds to obtain the rate of change in units of per minute. For example, if the current physiological balance index is 0.72 and the value 5 minutes ago was 0.68, the calculated rate of change is positive 0.008 / minute.

[0040] When executing the pattern recognition logic, the current physiological balance index of 0.72 is first compared with a preset range of values. If the value falls within the range of 0.65 to 0.85, it is initially classified as a candidate for the thermal balance maintenance mode. Then, the rate of change (0.008 / min) is compared with the defined threshold of 0.02 / min for this mode, confirming that the rate of change is less than the threshold and meets the characteristics of the thermal balance maintenance mode. Through dual verification of value range comparison and rate of change feature matching, the system ultimately identifies that the patient is currently in the thermal balance maintenance mode.

[0041] The multi-mode regulation strategy library is stored using a relational database structure. For the thermal equilibrium maintenance mode, the corresponding temperature and viscosity joint regulation strategy is set as follows: the flushing fluid temperature is adjusted by ±0.5 degrees Celsius, and the viscosity by ±8%. The timing of the adjustments is specified as follows: temperature fine-tuning is performed first, followed by viscosity adjustment after a 30-second interval, ensuring that the temperature effect is initially manifested before viscosity intervention. For the stress compensation mode, the regulation strategy is set as follows: the temperature adjustment range is increased to ±1.5 degrees Celsius, and the viscosity adjustment range is increased to ±15%. The timing of the adjustments is changed to simultaneous activation of temperature and viscosity, accelerating the recovery of physiological balance by increasing the intensity of stimulation. For the imbalance warning mode, the regulation strategy is set as follows: when the temperature adjustment reaches ±2.5 degrees Celsius and the viscosity adjustment reaches ±25%, an alarm mechanism is triggered simultaneously. This strategy prioritizes rapid temperature adjustment, followed by a significant viscosity intervention after a 15-second delay.

[0042] Based on the identified thermal equilibrium maintenance mode, the system sends a query command to the multi-mode regulation strategy library, carrying the mode identifier as the search keyword. The database performs an index lookup, locates the strategy record corresponding to the mode identifier, and reads the time-series parameters stored in the record: temperature regulation range ±0.5 degrees Celsius, viscosity regulation range ±8%, and a 30-second interval between temperature and viscosity regulation. The system extracts this complete set of strategy parameters and loads it into the execution buffer, serving as the source of operation commands for subsequent temperature and viscosity control devices, thus achieving automated closed-loop control from mode recognition to strategy selection.

[0043] For example, during the execution of the temperature-viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored, and the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected by calculating the deviation between the actual response trajectory and the expected response trajectory, including: The current physiological balance index is obtained as the starting value. Based on the difference between the target physiological balance index and the starting value, the expected physiological balance index value at each sampling time is calculated in a linearly increasing manner over time. The expected physiological balance index values ​​are stored in chronological order as an array of expected response trajectories. The measured values ​​of the current physiological balance index are collected at fixed intervals. The measured values ​​at each sampling time are stored in chronological order as an actual response trajectory array. The timestamps of the actual response trajectory array correspond completely with those of the expected response trajectory array. Extract the expected value corresponding to the current time from the expected response trajectory array, calculate the difference between the current measured value and the expected value to obtain the deviation at the current time, count the deviation values ​​of the three most recent consecutive sampling times, and calculate the arithmetic mean of the three deviation values ​​as the average deviation; Determine whether the sign of the current deviation is consistent with that of the average deviation. If the signs are consistent and the absolute value of the average deviation is greater than a preset deviation threshold, divide the average deviation by the difference between the target physiological balance index and the initial value to obtain the normalized deviation ratio. Multiply the temperature adjustment range by a first correction factor to update the new temperature adjustment range. Multiply the viscosity adjustment range by a second correction factor to update the new viscosity adjustment range. The product of the first correction factor and the second correction factor is a fixed constant. The first correction factor increases monotonically as the normalized deviation ratio increases.

[0044] During the regulation strategy execution phase, dynamic optimization of regulation parameters is achieved through a real-time trajectory tracking mechanism. When the regulation strategy is initiated, the current physiological balance index is first collected as the starting value. Assuming the current value is 0.42 and the target physiological balance index is set to 0.75, the difference is 0.33. Based on the preset total regulation duration, for example, set to 600 seconds, the difference is evenly distributed along the time axis, and the expected values ​​at each sampling time are calculated.

[0045] During actual execution, the system continuously collects measured values ​​of the physiological balance index at 30-second intervals. The measured value at the first sampling time might be 0.430, and at the second sampling time, it might be 0.448. These measured values ​​are stored in the actual response trajectory array in chronological order, ensuring that their timestamps are perfectly aligned with the expected response trajectory array. When the third sampling time is reached, the system extracts the expected value corresponding to that time, 0.4695, while the measured value is 0.462. The current deviation is calculated as 0.462 - 0.4695 = -0.0075, indicating that the actual response is slightly lower than expected.

[0046] To avoid the impact of random fluctuations caused by a single deviation measurement, the system backtracks to the deviation values ​​at the three most recent sampling times. Assuming the previous two deviations were -0.0065 and -0.0070 respectively, the arithmetic mean of the three is (-0.0065 - 0.0070 - 0.0075) / 3 = -0.0070. At this point, the system checks for sign consistency. Since both the current deviation and the average deviation are negative and have the same sign, and if the absolute value of the average deviation (0.0070) is greater than the preset threshold of 0.005, a correction mechanism is triggered.

[0047] During the correction calculation, the average deviation is first divided by the initial deviation to obtain the normalized deviation ratio. Based on the direction of the deviation, if the actual response lags behind expectations and requires increased adjustment, the first correction factor is set to... ,in With an amplification factor of 8, the first correction factor is 1.1696. Since the product of the two correction factors must remain constant at 1.15, the second correction factor is calculated as 1.15 / 1.1696 = 0.9832. The current temperature adjustment range of 3.5℃ is multiplied by 1.1696 to update it to 4.09℃, and the viscosity adjustment range is adjusted from 1.8 mPa·s to 1.77 mPa·s. This temperature-responsive-prioritized strategy accelerates the physiological equilibrium index towards the target value.

[0048] In subsequent sampling cycles, the corrected parameters remain in effect until the next deviation assessment. If the deviation sign reverses after three consecutive sampling points, or the average absolute value of the deviation falls below the threshold, the system pauses parameter correction and maintains the current settings. This closed-loop control method achieves adaptive compensation for individual patient differences and dynamic changes in the flushing process by continuously comparing the expected trajectory with the actual trajectory, ensuring that the adjustment process always proceeds along the predetermined physiological recovery path.

[0049] A second aspect of the present invention provides an intelligent bladder irrigation dynamic control system based on multimodal physiological signal fusion, comprising: The signal acquisition unit is used to acquire the bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals of the patient during bladder irrigation. The feature extraction unit is used to decode the pulse sequence of the bladder wall mechanoreceptor signal, extract the pulse firing frequency and pulse interval distribution features, perform trend separation on the core body temperature signal, extract the body temperature baseline drift and the amplitude of body temperature periodic fluctuations, perform time-domain statistical analysis on the lower abdominal skin electrical signal, extract skin conductance level and skin electrical response frequency, map each feature to a standardized feature space, capture the dynamic coupling relationship of multimodal signals by constructing a temporal correlation matrix between features, and calculate the physiological balance index characterizing the overall state of the bladder-nerve-body temperature system. The strategy control unit is used to identify the patient's current physiological regulation mode based on the physiological balance index, select a matching temperature and viscosity joint regulation strategy from a preset multi-mode regulation strategy library for the identified physiological regulation mode, continuously monitor the response trajectory of the physiological balance index during the execution of the temperature and viscosity joint regulation strategy, and dynamically correct the temperature regulation amplitude and the viscosity regulation amplitude by calculating the deviation between the actual response trajectory and the expected response trajectory. The output unit is used to drive the temperature control device and viscosity control device to process the flushing fluid according to the corrected adjustment parameters and then infuse it into the bladder.

[0050] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0051] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0052] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent dynamic control of bladder irrigation based on multimodal physiological signal fusion, characterized in that, include: Acquire bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals during bladder irrigation. The pulse sequence of the bladder wall mechanoreceptor signal is decoded to extract the pulse firing frequency and pulse interval distribution features. The core body temperature signal is trend-separated to extract the body temperature baseline drift and the amplitude of periodic body temperature fluctuations. The lower abdominal electrodermal signal is subjected to time-domain statistical analysis to extract the skin conductance level and the number of skin electrical responses. Each feature is mapped to a standardized feature space. The dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is calculated. Based on the physiological balance index, the patient's current physiological regulation mode is identified. For the identified physiological regulation mode, a matching temperature and viscosity joint regulation strategy is selected from a preset multi-mode regulation strategy library. During the execution of the temperature and viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored. By calculating the deviation between the actual response trajectory and the expected response trajectory, the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected. The flushing fluid is processed by the temperature control device and viscosity control device according to the corrected adjustment parameters and then infused into the bladder.

2. The method according to claim 1, characterized in that, Pulse sequence decoding of bladder wall mechanoreceptor signals, trend separation of core body temperature signals, and time-domain statistical analysis of lower abdominal electrodermal signals were performed, including: The bladder wall mechanoreceptor signal is preprocessed with bandpass filtering to remove baseline drift and high-frequency noise interference. A threshold detection method is used to identify effective pulse events. The pulse firing frequency within a unit time window is statistically analyzed. The mean and standard deviation of the interval between adjacent pulses are calculated as pulse interval distribution characteristics. The core body temperature signal is filtered by a moving average filter to extract long-period trend components as the body temperature baseline. The body temperature fluctuation component is obtained by the difference between the core body temperature signal and the body temperature baseline. The body temperature fluctuation component is subjected to spectral analysis to identify the dominant period and extract the fluctuation amplitude corresponding to the dominant period. The average conductivity value of the lower abdominal skin electrical signal within a preset time period is calculated as the skin electrical conductivity level. The abrupt change point of the skin electrical signal is detected by the first-order difference method. The number of occurrences of the abrupt change point is counted as the number of skin electrical response. The peak amplitude and rise time of each skin electrical response are recorded. The extracted pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and number of skin conductance responses are used as the original feature set.

3. The method according to claim 2, characterized in that, Each feature is mapped to a standardized feature space. The dynamic coupling relationship of multimodal signals is captured by constructing a temporal correlation matrix between features. The physiological balance index characterizing the overall state of the bladder-nerve-temperature system is calculated, including: The pulse firing frequency, pulse interval distribution characteristics, body temperature baseline drift, body temperature periodic fluctuation amplitude, skin conductance level, and skin conductance response frequency are normalized respectively, and the numerical range of each feature is mapped to a unified standardized feature space. A multi-channel time series matrix is ​​established, wherein the rows of the multi-channel time series matrix correspond to different standardized features, the columns correspond to continuous time sampling points, and each matrix element records the standardized value of a specific feature at a specific time. For any two feature channels in the multi-channel time series matrix, calculate the time delay cross-correlation coefficient between them. The time delay cross-correlation quantifies the influence of one feature on another feature under different time delays. A temporal correlation matrix is ​​constructed, wherein the rows and columns of the temporal correlation matrix correspond to each normalized feature, and the value of the matrix element is the maximum time delay cross-correlation coefficient between the corresponding feature pairs. The temporal correlation matrix characterizes the dynamic coupling strength and temporal dependency between multimodal signals. The temporal correlation matrix is ​​decomposed to extract principal components that represent the overall coordinated mode of multimodal signals, and the physiological balance index is calculated based on the principal components.

4. The method according to claim 3, characterized in that, Matrix decomposition is performed on the temporal correlation matrix to extract principal components representing the overall coordinated mode of multimodal signals. Based on these principal components, a physiological balance index is calculated, including: Singular value decomposition is performed on the time-series correlation matrix to obtain a left singular vector matrix, a singular value diagonal matrix, and a right singular vector matrix. The singular value sequence is extracted from the singular value diagonal matrix, and the decreasing slope of adjacent singular values ​​in the singular value sequence is calculated. Identify the inflection points in the descending slope sequence where the absolute value of the slope is less than a preset slope threshold. Determine the singular values ​​before the inflection point and their corresponding left singular vectors as effective principal components. The number of effective principal components is directly determined by the index of the inflection point. The real-time values ​​of each standardized feature at the current moment are extracted from the multi-channel time series matrix. The real-time values ​​are then multiplied by matrix-vector multiplication with each column vector in the left singular vector matrix corresponding to the effective principal components to obtain the component values ​​of each effective principal component space dimension at the current moment. A piecewise nonlinear mapping function is constructed, and the component values ​​of each effective principal component space dimension are respectively input into the piecewise nonlinear mapping function. When the component value is located in the preset central interval, a linear mapping is applied, and when the component value is not in the central interval, a saturation suppression mapping is applied. The output values ​​of each dimension after mapping are weighted and summed according to the square root of the corresponding singular value. The weighted summation result is normalized to a preset numerical range as the physiological balance index.

5. The method according to claim 1, characterized in that, Based on the physiological balance index, the patient's current physiological regulation mode is identified, and a matching temperature-viscosity combined regulation strategy is selected from a pre-set multi-mode regulation strategy library, including: A classification rule for physiological regulation modes is pre-established. The classification rule for physiological regulation modes defines the numerical range and rate of change characteristics of physiological balance index corresponding to different physiological regulation modes. The physiological regulation modes include thermal balance maintenance mode, stress compensation mode and imbalance early warning mode. Obtain the rate of change of the physiological balance index at the current moment within a preset time window; The physiological balance index is compared with the numerical range defined in the physiological regulation mode classification rules, and the rate of change is matched with the rate of change feature defined in the physiological regulation mode classification rules. Based on the comparison and matching results, the patient's current physiological regulation mode is identified. Access the multi-mode regulation strategy library, which stores temperature and viscosity joint regulation strategies corresponding to each physiological regulation mode. Each temperature and viscosity joint regulation strategy specifies the adjustment range of the flushing fluid temperature, the adjustment range of the flushing fluid viscosity, and the timing relationship between the two. Based on the identified physiological regulation pattern, a temperature-viscosity joint regulation strategy that matches the physiological regulation pattern is retrieved from the multi-mode regulation strategy library and selected as the current regulation strategy to be executed.

6. The method according to claim 1, characterized in that, During the execution of the temperature-viscosity joint regulation strategy, the response trajectory of the physiological balance index is continuously monitored. By calculating the deviation between the actual response trajectory and the expected response trajectory, the temperature regulation amplitude and the viscosity regulation amplitude are dynamically corrected, including: The current physiological balance index is obtained as the starting value. Based on the difference between the target physiological balance index and the starting value, the expected physiological balance index value at each sampling time is calculated in a linearly increasing manner over time. The expected physiological balance index values ​​are stored in chronological order as an array of expected response trajectories. The measured values ​​of the current physiological balance index are collected at fixed intervals. The measured values ​​at each sampling time are stored in chronological order as an actual response trajectory array. The timestamps of the actual response trajectory array correspond completely with those of the expected response trajectory array. Extract the expected value corresponding to the current time from the expected response trajectory array, calculate the difference between the current measured value and the expected value to obtain the deviation at the current time, count the deviation values ​​of the three most recent consecutive sampling times, and calculate the arithmetic mean of the three deviation values ​​as the average deviation; Determine whether the sign of the current deviation is consistent with that of the average deviation. If the signs are consistent and the absolute value of the average deviation is greater than a preset deviation threshold, divide the average deviation by the difference between the target physiological balance index and the initial value to obtain the normalized deviation ratio. Multiply the temperature adjustment range by a first correction factor to update the new temperature adjustment range. Multiply the viscosity adjustment range by a second correction factor to update the new viscosity adjustment range. The product of the first correction factor and the second correction factor is a fixed constant. The first correction factor increases monotonically as the normalized deviation ratio increases.

7. A smart bladder irrigation dynamic control system based on multimodal physiological signal fusion, used to implement the method as described in any one of claims 1-6, characterized in that, include: The signal acquisition unit is used to acquire the bladder wall mechanoreceptor signals, core body temperature signals, and lower abdominal skin electrical signals of the patient during bladder irrigation. The feature extraction unit is used to decode the pulse sequence of the bladder wall mechanoreceptor signal, extract the pulse firing frequency and pulse interval distribution features, perform trend separation on the core body temperature signal, extract the body temperature baseline drift and the amplitude of body temperature periodic fluctuations, perform time-domain statistical analysis on the lower abdominal skin electrical signal, extract skin conductance level and skin electrical response frequency, map each feature to a standardized feature space, capture the dynamic coupling relationship of multimodal signals by constructing a temporal correlation matrix between features, and calculate the physiological balance index characterizing the overall state of the bladder-nerve-body temperature system. The strategy control unit is used to identify the patient's current physiological regulation mode based on the physiological balance index, select a matching temperature and viscosity joint regulation strategy from a preset multi-mode regulation strategy library for the identified physiological regulation mode, continuously monitor the response trajectory of the physiological balance index during the execution of the temperature and viscosity joint regulation strategy, and dynamically correct the temperature regulation amplitude and the viscosity regulation amplitude by calculating the deviation between the actual response trajectory and the expected response trajectory. The output unit is used to drive the temperature control device and viscosity control device to process the flushing fluid according to the corrected adjustment parameters and then infuse it into the bladder.

8. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 6.

Images