A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology

Abstract

This invention discloses a breath ammonia detection system for chronic kidney disease based on intelligent sensing technology, belonging to the field of breath ammonia detection technology for chronic kidney disease. The system includes: a breath sampling module, an ammonia sensor array module, a respiratory phase matching module, a feature compensation calculation module, and a state determination output module. The breath sampling module controls the patient to exhale deeply and generates an expiratory volume marker; the ammonia sensor array module starts at the beginning of the exhalation and dynamically adjusts the bias voltage of the electrochemical ammonia sensor during exhalation to obtain the original current response sequence; the respiratory phase matching module truncates the baseline based on time alignment and calculates the noise amplitude; the feature compensation calculation module performs peak value preservation and volume normalization on the out-of-limit response peak to obtain a characteristic value of the breath ammonia concentration per unit volume; the state determination output module compares this characteristic value with the reference interval for kidney function staging and outputs a risk level identifier. The system achieves highly sensitive, interference-resistant detection and staging determination of breath ammonia.

Description

Technical Field

[0001] This invention relates to the field of respiratory ammonia detection technology for chronic kidney disease, specifically a respiratory ammonia detection system for chronic kidney disease based on intelligent sensing technology. Background Technology

[0002] Clinical staging of chronic kidney disease relies on blood biochemical indicators such as serum creatinine and estimated glomerular filtration rate. These methods require invasive blood sampling and have long testing cycles, making them unsuitable for routine home monitoring and early screening. Exhaled ammonia, as an exhaled component of urea metabolism abnormalities in the body when kidney function declines, can reflect the state of kidney detoxification function through changes in its concentration, providing a feasible non-invasive auxiliary assessment method for chronic kidney disease.

[0003] Existing breath ammonia detection devices typically employ a single electrochemical sensor driven by a fixed bias voltage, uniformly amplifying and filtering the entire signal after sampling. This approach struggles to distinguish the temporal correlation between oral environment fluctuations, expiratory flow changes, and the target ammonia response during patient exhalation. The intrusion of airflow into the expiratory dead space interferes with the baseline signal. Furthermore, prolonged operation in high-humidity expiratory environments alters the charge transfer efficiency of the electrode surface, causing a slow drift in the output current and reducing the accuracy of identifying the effective ammonia response peak in the detection results. The results are often presented as the average concentration of the entire expiratory sample. This overall averaging strategy discards the dynamic changes in ammonia concentration at different stages of exhalation, failing to reconstruct the concentration differences between deep alveolar gas and airway exchange gases, resulting in a single basis for staging.

[0004] In actual breath ammonia detection, before exhalation begins, the sensor output exhibits a baseline current range related to residual gas in the oral cavity and environmental background. The amplitude of this baseline varies individually due to patient breathing habits and ambient humidity. Directly capturing a fixed time window as the baseline can cause effective signals from some patients to be submerged by baseline noise. Furthermore, in high-humidity exhalation atmospheres, the electrochemical ammonia sensor, under a fixed bias voltage mode, exhibits a competitive relationship between the ammonia oxidation reaction and the electrochemical reduction reaction of water molecules. The reduction current of water molecules becomes an interfering component, making it difficult to reliably extract the sensor's response characteristics to low-concentration ammonia from the current sequence. Summary of the Invention

[0005] This invention provides a breath ammonia detection system for chronic kidney disease based on intelligent sensing technology. It establishes time alignment between the expiratory volume and current response sequence by marking the start and end times of the expiratory action. Individualized current response baseline segments are dynamically extracted from the sequence, and noise amplitude is calculated. Simultaneously, the system dynamically scans the bias voltage during sampling to separate the response differences between ammonia oxidation current and water molecule reduction current. A volume-normalized characteristic value of the breath ammonia concentration per unit volume is extracted from the original sequence. Based on a comparison of the membership degree of this characteristic value with the reference interval for renal function staging, a risk level identifier is output.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a breath ammonia detection system for chronic kidney disease based on intelligent sensing technology. The system includes a breath sampling module, an ammonia sensor array module, a respiratory phase matching module, a feature compensation calculation module, and a state determination output module.

[0007] The exhalation sampling module controls a patient with chronic kidney disease to perform a deep exhalation of a set duration into a gas sampling container, simultaneously recording the start and end times of the exhalation and generating an expiratory volume marker. As a technical solution of this invention, the module detects the temperature change of the airflow in front of the patient's mouth and nose using an infrared pyroelectric sensor. When the slope of the temperature change meets a threshold condition, it is marked as the start time and timing begins. Simultaneously, a flow meter measures and integrates the expiratory flow rate in real time; when the integration result reaches the target volume threshold, it is automatically marked as the end time. The expiratory volume marker is calculated by multiplying the timing duration by the average flow rate, and the start and end timestamps are associated and stored, achieving precise quantification and marking of the exhalation process. Preferably, to avoid invalid exhalations, if no effective flow is detected within a set judgment time after the start time, the system automatically determines the exhalation is invalid and resets, waiting for the next action, improving the reliability and convenience of the test.

[0008] Upon receiving the signal indicating the start of the exhalation action, the ammonia sensor array module activates its internal heating circuit. It then uses pulse width modulation to control the heating power, shortening the heating time and heating the electrochemical ammonia sensor to a predetermined isothermal operating point. During the heating process, a transitional heating current sequence is recorded. After reaching the isothermal operating point, the module gradually increases the bias voltage applied to the sensor from its initial value, maintaining stability after each step and recording the final output current to generate a bias scanning current sequence. By concatenating the transitional heating current sequence and the bias scanning current sequence in chronological order and associating them with sampling time stamps, a time-varying original current response sequence is constructed. This scheme, through dynamic adjustment of the bias voltage, can elicit the response characteristics of different components in the exhaled sample on the sensor, providing rich information for subsequent feature extraction. Preferably, during the heating transition, if the rate of change of the output current for several consecutive sampling cycles is lower than a set threshold, the heating is terminated early, and the process jumps to the bias voltage scanning stage to shorten the testing time and improve detection efficiency.

[0009] The respiratory phase matching module is used to obtain the time alignment between the expiratory volume marker and the original current response sequence. Specifically, the module extracts the timestamps of the start and end times of the expiratory action, finds time-aligned sampling points in the original current response sequence, and uses these as the baseline start point and baseline end point, respectively. A segment of the sequence preceding the baseline start point is extracted from the original sequence as the current response baseline segment, and the standard deviation of the current values ​​at all sampling points within this baseline segment is calculated. Three times this standard deviation is used as the average noise amplitude. Through precise phase matching and baseline extraction, the obtained average noise amplitude represents the system's background noise level, providing a dynamic and objective basis for identifying effective signal peaks.

[0010] The feature compensation calculation module receives the average noise amplitude and sets a multiple thereof as the peak detection threshold. It iterates through the interval from the baseline start point to the baseline end point in the original current response sequence, marking regions that continuously exceed the detection threshold as response peak regions, and extracts the maximum current value within each region as the peak value of that response peak. This peak value is then preserved by replacing all current values ​​within the entire response peak region with this peak value, generating a preserved peak sequence. Finally, each peak value in the preserved peak sequence is divided by the expiratory volume value in the expiratory volume marker to obtain the characteristic value of expiratory ammonia concentration per unit volume. This processing effectively eliminates noise interference and, through volume normalization, eliminates the influence of differences in vital capacity and expiratory effort among different patients, allowing the characteristic value to more accurately reflect the ammonia concentration level in exhalation.

[0011] The status determination output module compares the membership degrees of each peak value in the unit volume of exhaled ammonia concentration with pre-stored renal function staging reference intervals. By statistically analyzing the renal function staging levels to which each peak falls and calculating the percentage of peaks falling into each level out of the total number of peaks, the staging level with the highest percentage value is selected as the risk level identifier for the chronic kidney disease patient. This determination method, based on the frequency membership principle of multi-peak statistics, improves the robustness and accuracy of risk grading compared to comparisons of single values.

[0012] As a technical solution of the present invention, the system further includes an ammonia sensor array self-calibration module. Before each breath test, this module obtains an initial zero-point current value by introducing zero-level air into the sensor and measuring its zero-bias current; after the test, it measures the zero-point current value again to obtain the post-test zero-point current value, and the difference between the two is used as the calibration bias current value. This value is used to correct the characteristic value of the ammonia concentration per unit volume of exhaled air, thereby compensating for the zero-point drift generated by the sensor during the test, ensuring the stability and accuracy of long-term measurements.

[0013] As a further improvement of the present invention, the system also includes an exhaled humidity compensation module. This module simultaneously collects the relative humidity value within the gas collection container using a capacitive humidity sensor while the ammonia sensor array samples, generating a humidity time series aligned with the original current response sequence. For each response peak region, the average relative humidity value within its time window is calculated. When this average value exceeds a set humidity reference value, a humidity correction factor is generated. This correction factor is used to adjust the characteristic value of the exhaled ammonia concentration per unit volume, thereby effectively correcting the impact of environmental or exhaled humidity changes on the sensor response and improving the system's detection consistency and reliability under different environments.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: The respiratory phase matching module extracts the timestamps of the start and end times of the expiratory action from the expiratory volume markers. The sampling point with the smallest time difference from these two timestamps is then located in the original current response sequence to pinpoint the individualized baseline start and end points. A sequence segment between the baseline start point and a predetermined number of sampling points is extracted as the current response baseline segment, not limited by a fixed time window. The start and end positions of this baseline segment correspond to each patient's actual expiratory process. The standard deviation of the current values ​​at all sampling points within this baseline segment is calculated, and three times the standard deviation is used as the average noise amplitude. This noise amplitude reflects the true fluctuations of the patient's oral background signal and environmental interference during the current test. In the feature compensation calculation module, a peak detection threshold is set as a multiple of this average noise amplitude. The sequence segment from the baseline start point to the baseline end point is traversed, and the maximum current value within each response peak region is extracted and peak hold processing is performed to generate a held peak sequence. This ensures that the effective ammonia response peak can be reliably separated from the individualized noise background, avoiding the missed detection of effective signals from some patients with low ammonia concentrations under a fixed threshold.

[0015] Upon receiving the timestamp of the exhalation termination moment, the ammonia sensor array module activates its internal heating circuit to raise the operating temperature of the electrochemical ammonia sensor to its isothermal operating point. During the heating process, the output current value is recorded to generate a heating transition current sequence. After reaching the isothermal operating point, the bias voltage amplitude is gradually increased from the initial bias voltage value according to a preset scan step size, and the output current value at the end of each bias voltage step is recorded to generate a bias scan current sequence. The two sequences are then spliced ​​together over time to form the original current response sequence. Under different bias voltages, the difference between the oxidation potential of ammonia molecules and the reduction potential of water molecules on the sensor electrodes is amplified. The current sequence obtained during the bias scan contains information on the changes in the contribution of ammonia oxidation and water reduction reactions to the total current as a function of the bias voltage. The feature compensation calculation module performs peak preservation and volume normalization on response peaks exceeding a set multiple of the noise amplitude. The resulting characteristic value of the exhaled ammonia concentration per unit volume mainly characterizes the true response intensity of the ammonia component, reducing the contribution of interference current components introduced by the electrochemical reaction of water molecules under high humidity conditions to the ammonia concentration calculation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a schematic diagram of a breath ammonia detection system for chronic kidney disease based on intelligent sensing technology. Figure 2This is a flowchart of the temperature rise transition and bias voltage scan control of the ammonia sensor array module; Figure 3 This is a flowchart of the respiratory phase matching module; Figure 4 This is a flowchart of feature compensation and chronic kidney disease risk level determination. Detailed Implementation

[0018] 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, 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.

[0019] See Figure 1 This invention provides a breath ammonia detection system for chronic kidney disease based on intelligent sensing technology. The system includes a breath sampling module, an ammonia sensor array module, a respiratory phase matching module, a feature compensation calculation module, and a state determination output module. The breath sampling module controls a chronic kidney disease patient to perform a deep exhalation of a predetermined duration into a gas sampling container, simultaneously recording the start and end times of the exhalation and generating an exhalation volume marker. The ammonia sensor array module activates its internal heating circuit at the end of the exhalation, continuously sampling the exhalation sample in the gas sampling container and dynamically adjusting the bias voltage amplitude applied to the electrochemical ammonia sensor during sampling to obtain a raw current response sequence that changes over time. The respiratory phase matching module obtains the time alignment relationship between the exhalation volume marker and the raw current response sequence, and based on the time alignment relationship, extracts the current response baseline segment corresponding to the start time of the exhalation, calculating the average noise amplitude of the baseline segment. The feature compensation calculation module performs peak preservation processing on response peaks in the original current response sequence that exceed a set multiple of the average noise amplitude, and then performs expiratory volume normalization transformation on the preserved peak sequence to obtain the characteristic value of expiratory ammonia concentration per unit volume. The state determination output module compares the characteristic value of expiratory ammonia concentration per unit volume with the pre-stored renal function staging reference intervals to output the risk level identifier of the chronic kidney disease patient.

[0020] Example 1: In practice, the breath sampling module continuously detects temperature changes caused by airflow in front of the patient's mouth and nose using an integrated infrared pyroelectric sensor. The infrared pyroelectric sensor outputs a temperature value sequence at a fixed sampling frequency. The signal processor within the breath sampling module performs first-order difference operations on this temperature value sequence to acquire the airflow temperature change curve in real time and continuously calculates the slope value between several consecutive sampling points on the curve. When the calculated slope value exceeds a pre-set slope threshold stored in the non-volatile memory of the breath sampling module, the module marks the current moment as the start of the exhalation action and simultaneously triggers a timer to begin accumulating the count in milliseconds. The set slope threshold is determined and written into the non-volatile memory after statistical analysis of the temperature curve slopes of multiple subjects' natural exhalation actions during the system calibration phase.

[0021] A flow meter, operating synchronously with an infrared pyroelectric sensor, is installed in the airway between the gas collection container and the patient's mouth. The flow meter measures the expiratory flow rate in real time and transmits this value as a digital signal to the expiratory flow collection module. Upon receiving the expiratory flow rate, the module initiates a time integration operation. The integration uses a rectangular summation method; each integration cycle multiplies the current expiratory flow rate by the cycle duration and adds the result to the integration variable. When the integration result variable reaches a preset target volume threshold, the module automatically marks the current moment as the end of the expiratory action and simultaneously sends a stop signal to the timer. It then reads the accumulated timing value from the timer since the start of the expiratory action and records this accumulated value as the set duration. The target volume threshold is preset based on the lung capacity reference range for patients with chronic kidney disease and is loaded from the configuration file into the expiratory flow collection module's memory during system initialization.

[0022] After obtaining the set duration, the exhalation sampling module calculates the arithmetic mean of the expiratory flow rates measured by the flow meter during the time interval from the start to the end of the exhalation, and uses this arithmetic mean as the average flow rate. Next, the exhalation sampling module multiplies the accumulated timing value by the average flow rate, and uses the product as the expiratory volume marker. The exhalation sampling module then associates and stores the timestamps of the start and end of the exhalation with the expiratory volume marker in an internal data buffer in a structured data record format, forming an expiratory volume marker data packet.

[0023] After detecting the airflow temperature change curve using an infrared pyroelectric sensor and marking the start time of the exhalation action, the exhalation acquisition module starts a countdown timer with a set judgment duration. If the expiratory flow rate measured by the flow meter fails to exceed the start flow threshold before the set judgment duration expires, the exhalation acquisition module determines this exhalation action as invalid. After the invalid exhalation determination takes effect, the exhalation acquisition module clears the timer's accumulated value, removes the stored timestamp of the exhalation action start time, and re-enters the waiting state to continue detecting the start time of the next possible exhalation action using the infrared pyroelectric sensor. The set judgment duration is 500ms, and the start flow threshold is 0.1L / min. These two values ​​are predetermined based on the minimum flow characteristics required for an effective deep exhalation in respiratory physiology and are stored as constant values ​​in the firmware of the exhalation acquisition module.

[0024] Example 2: In specific implementation, please refer to Figure 2 The ammonia sensor array module contains a digital signal processor (DSP) and a timer. It receives the timestamp of the start of the exhalation action from the exhalation sampling module via the system bus. After parsing the timestamp, the DSP sets the sampling start flag in its internal status register to valid and sends a pulse width modulation (PWM) signal to the drive port of the internal heating circuit. Rapid heating is achieved by controlling the duty cycle. The heating resistance wire in the heating circuit begins heating the gas flowing through the electrochemical ammonia sensor's chamber, ensuring that the isothermal operating point is reached before the end of the exhalation action. During the stable isothermal operating point period, the analog-to-digital converter (ADC) inside the ammonia sensor array module samples the output current of the electrochemical ammonia sensor every sampling cycle. The sampling cycle is controlled by a timer, and its duration is equal to a fixed value after dividing the system master clock frequency. The ADC converts the analog current value obtained from each sampling into a digital quantity. The DSP reads this digital quantity and appends the count value of the current sampling time, storing it as a record in the designated heating transition current sequence data area in the internal buffer. For each record stored, the DSP simultaneously calculates the rate of change of the output current value over three consecutive sampling cycles. The rate of change is calculated as follows: The difference between the output current of the current sampling period and the previous sampling period is taken as the first difference; the difference between the output current of the previous sampling period and the sampling period before that is taken as the second difference. The absolute values ​​of the first and second differences are calculated respectively. The maximum value of the two is divided by the sampling period duration to obtain the rate of change value. The digital signal processor compares the calculated rate of change value with a set rate of change threshold stored in a register. The set rate of change threshold is obtained through experimental calibration based on the output current fluctuation characteristics of the electrochemical ammonia sensor during the temperature stabilization phase; its value is... If the rate of change calculated for three consecutive sampling cycles is less than the set rate of change threshold, the ammonia sensor array module determines that the electrochemical ammonia sensor is close to thermal equilibrium, and the heating transition phase can be ended prematurely. The digital signal processor then terminates the high-level output of the heating circuit drive port, stops recording the heating transition current sequence, and switches the system state to the bias voltage scanning phase.

[0025] After the electrochemical ammonia sensor reaches its isothermal operating point, if no premature termination occurs, it automatically enters the bias voltage scanning phase after the temperature feedback loop of the heating circuit provides a temperature stabilization indication signal. During the bias voltage scanning phase, the bias voltage generation circuit inside the ammonia sensor array module is controlled by a digital signal processor (DSP) via a digital-to-analog converter (DAC). The DSP uses the initial bias voltage value as the starting point, which is set to... The bias voltage amplitude applied between the working electrode and the reference electrode of the electrochemical ammonia sensor is gradually increased according to a preset scan step size. The scan step size is determined by the total range and number of steps of the bias voltage scan, stored in the configuration register, and set to 0.05V. After each step increase, the digital signal processor starts a hold timer to maintain the output voltage of the bias voltage generation circuit at a stable bias voltage value for a set hold time, which is set to 150ms. At the end of the hold time, the analog-to-digital converter samples the output current of the electrochemical ammonia sensor, the digital signal processor records the sampled output current value, and stores this output current value along with the currently applied bias voltage value in the designated bias scan current sequence data area in the internal buffer.

[0026] The bias voltage scanning phase ends when the bias voltage amplitude reaches the preset maximum bias voltage value. The maximum bias voltage value is determined based on the safe operating voltage range of the electrochemical ammonia sensor. In a specific configuration, the maximum bias voltage value is... Subsequently, the digital signal processor (DSP) reads records from the heating transition current sequence data area and the bias scan current sequence data area from its internal buffer, and concatenates them according to the chronological order in which they were generated. Each record in the heating transition current sequence data area contains a sampling time stamp and an output current value during the heating period, while each record in the bias scan current sequence data area contains a sampling time stamp and an output current value during the bias voltage scan. The DSP merges the two sets of records to generate a complete set of records arranged in ascending chronological order. It then extracts the sampling time stamp and the corresponding output current value for each sampling point in this set, which together constitute the original current response sequence that varies over time. The original current response sequence is stored as an array in the shared memory of the ammonia sensor array module, accessible to the respiratory phase matching module via the system bus.

[0027] Example 3: In specific implementation, please refer to Figure 3 The respiratory phase matching module reads the expiratory volume marker data packet from the data buffer of the expiratory acquisition module via the system bus. The respiratory phase matching module has an internal timestamp parsing unit that decomposes the expiratory volume marker data packet, extracts the timestamps for the start and end times of the expiratory action, and stores these two timestamps into the first register group and the second register group, respectively.

[0028] Meanwhile, the respiratory phase matching module reads the raw current response sequence generated by the ammonia sensor array module through a shared memory access interface. The raw current response sequence is an array of structures, with each element containing two fields: a sampling timestamp and an output current value. The sequence retrieval unit within the respiratory phase matching module obtains the timestamp value of the start of the exhalation action from the first register group. Using this timestamp value as the target search value, it iterates through the sampling timestamp field of all elements in the raw current response sequence. The sequence retrieval unit calculates the absolute value of the time difference between each sampling timestamp field and the target search value, and continuously tracks the minimum absolute value of the time difference within the currently traversed range and its corresponding array index. When the traversal is complete, the sequence retrieval unit determines the array element corresponding to the minimum absolute value of the time difference as the baseline starting point, and stores this array index in the baseline starting point index register.

[0029] Optionally, the sequence retrieval unit searches for the sampling point with the smallest time difference from the timestamp of the exhalation termination moment in the original current response sequence using the same method. The sequence retrieval unit obtains the timestamp value of the exhalation termination moment from the second register group as the target search value, traverses the original current response sequence again, calculates the absolute value of the time difference between each sampling time tag field and the target search value, determines the array element corresponding to the smallest absolute value of the time difference as the baseline termination point, and stores the array index in the baseline termination point index register.

[0030] After locating the baseline start point and baseline end point, the baseline segment extraction unit within the respiratory phase matching module reads a set quantity value from the configuration register. This set quantity value determines how many sampling points are selected before the baseline start point to participate in the baseline noise calculation. Its value is determined based on the ratio of the stabilization time of the electrochemical ammonia sensor's output signal to the sampling period in an ammonia-free environment; in this embodiment, the set quantity value is 200. The baseline segment extraction unit checks whether the index value stored in the baseline start point index register is greater than or equal to the set quantity value. If the index value is greater than or equal to the set quantity value, the baseline segment extraction unit extracts a continuous sequence segment from the original current response sequence, starting from an element whose index value is the baseline start point index value minus the set quantity value, and ending at an element whose index value is the baseline start point index value minus one. This extracted sequence segment is used as the current response baseline segment. If the baseline start point index value is less than the set quantity value, meaning that a set number of sampling points before the baseline start point exceed the starting boundary of the original current response sequence, the baseline segment extraction unit extracts a sequence segment from the first sampling point of the original current response sequence to the baseline start point, using this extracted sequence segment as the current response baseline segment.

[0031] After obtaining the current response baseline, the noise calculation unit within the breathing phase matching module performs statistical calculations on the output current values ​​of all sampling points within the current response baseline. The noise calculation unit sums all output current values ​​within the current response baseline and divides the sum by the number of sampling points within the current response baseline to obtain the arithmetic mean. The noise calculation unit uses the arithmetic mean to calculate the standard deviation of the output current values ​​within the current response baseline. The standard deviation is calculated using the following formula: in, This represents the standard deviation of the output current value within the baseline range of the current response. This indicates the total number of sampling points within the current response baseline segment. Indicates the first segment of the current response baseline. The output current value at each sampling point This represents the arithmetic mean of the output current values ​​at all sampling points within the current response baseline segment. The noise calculation unit multiplies the calculated standard deviation by three, and the resulting product is used as the average noise amplitude. The basis for multiplying the standard deviation by three to use as the average noise amplitude is that, for the baseline current noise of an electrochemical ammonia sensor under target gas-free conditions, its instantaneous value follows a normal distribution, and the probability of the instantaneous value exceeding three times the standard deviation is less than 0.3%. Using this threshold to define the upper limit of noise can balance detection sensitivity and false trigger suppression.

[0032] The noise calculation unit combines the array indices of the baseline start point, the array indices of the baseline end point, and the average noise amplitude into a data structure, which serves as the phase matching parameter set. The communication unit within the breathing phase matching module transmits the phase matching parameter set to the receiving buffer of the feature compensation calculation module via the system bus, completing the transmission of phase matching data.

[0033] Example 4: In specific implementation, please refer to Figure 4 The feature compensation calculation module has an internal phase matching parameter set receiving buffer. It listens for data packets from the respiratory phase matching module via the system bus. When a phase matching parameter set arrives, it stores it in the receiving buffer. The parsing unit within the feature compensation calculation module reads the phase matching parameter set from the receiving buffer and extracts three parameters: the array index of the baseline start point, the array index of the baseline end point, and the average noise amplitude. These parameters are then stored in the start index register, end index register, and noise amplitude register within the feature compensation calculation module, respectively.

[0034] The peak detection threshold calculator within the feature compensation calculation module reads the average noise amplitude from the noise amplitude register and a set factor from the configuration register. The value of the set factor must balance the ability to capture low-concentration ammonia response peaks with the ability to suppress baseline noise fluctuations; in this embodiment, the set factor is set to 3. The average noise amplitude is multiplied by the set factor, and the resulting product is used as the peak detection threshold. The peak detection threshold is latched into the threshold register for use in subsequent comparison operations.

[0035] The sequence scanning unit within the feature compensation calculation module obtains the baseline start point array index from the start index register and the baseline end point array index from the end index register, and reads the original current response sequence through the shared memory access interface. The sequence scanning unit uses the baseline start point array index as the scan starting point and the baseline end point array index as the scan ending point, traversing the original current response sequence element by element. During the traversal, the sequence scanning unit compares the output current value of each sampling point with the peak detection threshold in the threshold register. When the output current value of a sampling point is greater than the peak detection threshold, and the output current value of the preceding sampling point is less than or equal to the peak detection threshold, the sequence scanning unit marks the array index of that sampling point as the region start index of a response peak region. Starting from the region start index, the sequence scanning unit continues traversing backwards until it detects that the output current value of a sampling point has fallen back to less than or equal to the peak detection threshold, and the output current value of the preceding sampling point is greater than the peak detection threshold. At this point, the array index preceding the array index of that sampling point is marked as the region end index of the response peak region. The sequence scanning unit treats the continuous interval formed by the region start index and the region end index as a response peak region and records it in the response peak region list.

[0036] The peak extraction unit within the feature compensation calculation module iterates through each response peak region in the response peak region list. For each response peak region, it reads the output current values ​​of all sampling points within the range from the region's start index to its end index. By comparing these values ​​one by one, it finds the maximum output current value and uses this maximum output current value as the peak-to-peak value of that response peak region. The peak extraction unit stores the peak-to-peak value of each response peak region and its corresponding response peak region index into the peak list data structure.

[0037] After obtaining the peak list, the peak hold processor within the feature compensation calculation module performs peak hold processing on each response peak region. For each response peak region recorded in the response peak region list, the peak hold processor creates a replacement sequence of output current values ​​with the same number of sampling points as that response peak region. Each element in the replacement sequence is assigned the peak-to-peak value corresponding to that response peak region. The peak hold processor replaces the original output current values ​​within that response peak region in the original current response sequence element by element with the replacement sequence, ensuring that the output current value of each sampling point within the response peak region is uniformly the peak-to-peak value of that response peak region. For sampling points in the original current response sequence that do not fall within any response peak region, their output current values ​​remain unchanged. After all response peak regions have been replaced, the peak hold processor generates the held peak sequence and stores it in the first sequence buffer within the feature compensation calculation module.

[0038] The peak time position statistics unit within the feature compensation calculation module traverses the preserved peak sequence, identifies each continuous segment where peak replacement has occurred, extracts the starting position array index of each continuous segment, and uses each extracted array index as a peak time position. The peak time position statistics unit arranges all peak time positions in ascending order of value, forming a one-dimensional array as the peak time position vector, and stores the peak time position vector in the position vector buffer within the feature compensation calculation module.

[0039] The normalization unit within the feature compensation calculation module reads the expiratory volume marker data packet stored in the expiratory volume acquisition module via the system bus and parses the expiratory volume value from it. The normalization unit reads the held peak sequence from the first sequence buffer and performs a division operation on the output current value corresponding to each sampling point in the held peak sequence. The dividend is the output current value, and the divisor is the expiratory volume value. After the division operation, the normalization unit generates a new sequence where the value corresponding to each sampling point represents the output current value per unit volume. This sequence is the characteristic value of the expiratory ammonia concentration per unit volume. The normalization unit stores the characteristic value of the expiratory ammonia concentration per unit volume in the output buffer within the feature compensation calculation module.

[0040] The status determination output module reads the characteristic value of expiratory ammonia concentration per unit volume from the output buffer of the feature compensation calculation module via the system bus. The memory controller inside the status determination output module reads the pre-stored renal function staging reference interval table from its internal non-volatile memory. The renal function staging reference interval table is a mapping table where each entry contains a renal function stage number, the corresponding lower limit and upper limit of expiratory ammonia concentration. For example: CKD stage 1: [80, 150] μg / L; CKD stage 2: (150, 250] μg / L; CKD stage 3: (250, 400] μg / L; CKD stage 4: (400, 600] μg / L; CKD stage 5: >600 μg / L. These values ​​are based on clinical statistical data and can be adjusted according to population characteristics. The renal function stage number uses integer encoding, and the lower and upper limits of expiratory ammonia concentration are both expressed as integers. The real value is expressed in units of 1.

[0041] The interval comparator in the state determination output module extracts each peak value in the characteristic value of ammonia concentration per unit volume of exhaled air, and selects the peak value with the largest value as the representative concentration value.

[0042] To convert the current value to the actual concentration, use the formula... Perform calculations, where This represents the ammonia concentration (μg / L). The sensitivity coefficient (calibrated using a standard gas, for example) ), In response to the peak current value (nA). The value represents the expiratory volume (L). The interval comparator compares each extracted peak value sequentially with the lower and upper limits of expiratory ammonia concentration for each entry in the renal function staging reference interval table. When a peak value is greater than or equal to the lower limit of expiratory ammonia concentration for a certain entry, but less than the upper limit, the interval comparator determines that the peak value falls into the renal function staging level corresponding to that entry. The interval comparator records the renal function staging level number to which each peak value falls and stores the renal function staging level numbers to which all peak values ​​belong in a statistical array.

[0043] While the status determination output module outputs the risk level identifier in binary code to the display terminal, the warning text generator within the status determination output module queries a text mapping table stored in internal non-volatile memory based on the risk level identifier. Each kidney function stage number in the text mapping table corresponds to a text warning message. The warning text generator extracts the text warning message that matches the current risk level identifier and encapsulates this text warning message along with the risk level identifier into a warning data message. The wireless communication control unit within the status determination output module calls the driver interface of the wireless communication module to send the warning data message to the monitoring terminal via the wireless communication module at a preset target address. The wireless communication module uses a wireless LAN protocol stack conforming to the IEEE 802.11 standard for data transmission. After receiving and parsing the warning data message, the monitoring terminal displays the risk level identifier of the chronic kidney disease patient and the corresponding text warning message to the monitoring personnel.

[0044] Example 5: In practical implementation, this system also includes an ammonia sensor array self-calibration module and an exhaled humidity compensation module. The two modules interact with the ammonia sensor array module and the feature compensation calculation module through the system bus, respectively.

[0045] Regarding the ammonia sensor array self-calibration module, it performs a zero-point drift calibration of the electrochemical ammonia sensor before each breath test. The input interface of the ammonia sensor array self-calibration module is connected to the system startup signal line. When the module receives the system startup signal, its internal control state machine enters the calibration process. The control state machine sends an opening command to the drive circuit of the zero-level air valve. After the zero-level air valve opens, zero-level air is introduced into the gas chamber inlet of the electrochemical ammonia sensor at a constant flow rate. Zero-level air refers to air with an ammonia concentration below [a certain value] after catalytic oxidation and adsorption filtration treatment. The duration of continuous ventilation for clean air is controlled by a purification duration timer within the control state machine. The timer's value is determined based on the ratio of the air chamber volume to the zero-order air velocity. In a specific configuration, the purification duration is set to... Second.

[0046] Throughout the entire purification process involving continuous zero-level air supply, the bias voltage control unit within the ammonia sensor array self-calibration module maintains the bias voltage of the electrochemical ammonia sensor at zero volts. The analog-to-digital converter within the ammonia sensor array self-calibration module samples the output current of the electrochemical ammonia sensor at regular intervals. These intervals are generated by a timer frequency division; in a specific configuration, the sampling interval is set to... Milliseconds. The analog-to-digital converter converts the analog current value obtained from each sampling into a digital quantity. The control state machine stores the digital quantity into the zero-bias current sequence buffer in the order of sampling time to obtain the zero-bias current sequence.

[0047] The moving weighted average calculator within the ammonia sensor array self-calibration module performs a moving weighted average operation on the zero-bias current sequence. The moving weighted average calculator maintains a sliding window with a length set to... In a specific configuration Values For the zero bias current sequence, the first... Each sampling point and subsequent consecutive sampling points A window consisting of sampling points is used to calculate the moving weighted average using an exponentially decaying weight distribution method. The weighting coefficients are... Set as The weights of adjacent sampling points are calculated as follows: Decreasing, of which This indicates the offset number of the sampling point relative to the starting position of the window. Each time the sliding window moves forward one sampling point, the moving weighted average calculator calculates the moving weighted average value within the current window and compares the difference between the newly calculated moving weighted average value and the previously calculated moving weighted average value. When the change in the moving weighted average value after a consecutive set number of times is less than the set stability threshold, the control state machine uses the last calculated moving weighted average value as the initial zero-point current value. The set number of times is taken as [value missing] in a specific configuration. The set stability threshold takes the value in a specific configuration. The control state machine records the initial zero-point current value in the first storage area of ​​the calibration result register, and at the same time sends a closing command to the drive circuit of the zero-level air valve to stop the supply of zero-level air.

[0048] After completing the current breath test, the system triggers the ammonia sensor array self-calibration module again to perform zero-point drift calibration. The control state machine of the ammonia sensor array self-calibration module sends an opening command to the drive circuit of the zero-level air valve again. Zero-level air is once again introduced into the inlet of the gas chamber of the electrochemical ammonia sensor at a constant flow rate. The continuous ventilation duration is the same as the purification duration before the test, and the bias voltage is once again kept at zero volts. The output current of the electrochemical ammonia sensor is collected again at the same sampling interval as before to obtain the zero-bias current sequence after the test. The moving weighted average calculator performs a moving weighted average calculation on the zero-bias current sequence after the test again with the same sliding window length, weighting coefficient, and stability threshold conditions. When the change in the number of consecutive set numbers is less than the set stability threshold, the moving weighted average value is taken as the zero-point current value after the test. The control state machine records the zero-point current value after the test in the second storage area of ​​the calibration result register.

[0049] The difference calculator in the ammonia sensor array self-calibration module reads the initial zero-point current value from the first storage area of ​​the calibration result register, and reads the zero-point current value after the test from the second storage area. It then calculates the difference between the zero-point current value after the test and the initial zero-point current value. The formula for calculating the difference is as follows: in, This indicates the calibration bias current value. This indicates the zero-point current value after the test. This represents the initial zero-point current value; all three variables are in microamperes. The difference calculator will... The calibration bias current value is used as the calibration bias current value. The communication interface within the ammonia sensor array self-calibration module encapsulates the calibration bias current value into a data message and sends it to the calibration data receiving port of the feature compensation calculation module via the system bus. When generating the characteristic value of exhaled ammonia concentration per unit volume, the feature compensation calculation module subtracts the calibration bias current value from the value of each sampling point in the held peak sequence to complete the zero-point drift correction of the characteristic value of exhaled ammonia concentration per unit volume.

[0050] Regarding the exhalation humidity compensation module, the internal capacitive humidity sensor is activated simultaneously with the ammonia sensor array module's continuous sampling. The activation signal for the capacitive humidity sensor is triggered by a sampling start flag sent by the ammonia sensor array module via the system bus. The capacitive humidity sensor acquires the relative humidity value within the gas collection container every sampling cycle. This sampling cycle is synchronized with the sampling cycle of the analog-to-digital converter in the ammonia sensor array module, obtained by frequency division from the same clock source. The capacitive humidity sensor outputs each acquired relative humidity value as a percentage. The humidity sequence buffer within the exhalation humidity compensation module stores the relative humidity values ​​in the humidity time sequence buffer according to the acquisition time, generating a humidity time sequence aligned with the original current response sequence.

[0051] In some embodiments, the exhaled humidity compensation module obtains a list of response peak regions generated by the feature compensation calculation module from the system bus. The list records the start and end indices of each response peak region. The humidity subsequence extraction unit within the exhaled humidity compensation module traverses the list of response peak regions. For each response peak region, using the start and end indices as time window boundaries, it extracts the relative humidity values ​​corresponding to all sampling points located between the start and end indices from the humidity time series. The extracted set of relative humidity values ​​is then used as the humidity value subsequence for that response peak region.

[0052] The average humidity calculator within the exhalation humidity compensation module accumulates all relative humidity values ​​in each subsequence of humidity values ​​for each response peak region. The accumulated result is then divided by the number of relative humidity values ​​in the subsequence to obtain the arithmetic mean. This arithmetic mean is used as the average relative humidity value for that response peak region. The average humidity calculator stores the average relative humidity value for each response peak region in an average humidity list.

[0053] The humidity reference value comparator in the exhalation humidity compensation module reads the set humidity reference value from the internal non-volatile memory. The set humidity reference value is a constant value, and in a specific configuration, it takes the value of... The humidity reference value comparator iterates through the average relative humidity values ​​of each response peak region in the average humidity list, comparing each average relative humidity value with the humidity reference value. When the average relative humidity value is greater than the set humidity reference value, the humidity correction factor calculator calculates the ratio of the average relative humidity value to the humidity reference value, and uses the reciprocal of the ratio as the humidity correction factor. When the average relative humidity value is less than or equal to the set humidity reference value, the humidity correction factor calculator sets the humidity correction factor to a unit value, with the unit value being [value to be filled in]. The humidity correction factor calculator stores the humidity correction factor corresponding to each response peak region into a humidity correction factor list.

[0054] The communication unit within the exhaled humidity compensation module sends a list of humidity correction factors to the feature compensation calculation module via the system bus. Upon receiving the humidity correction factor list, the feature compensation calculation module, during the generation of the exhaled ammonia concentration feature value per unit volume, multiplies the feature value per unit volume of the exhaled ammonia concentration for each response peak region by the humidity correction factor corresponding to that response peak region, thus completing the humidity compensation adjustment for the exhaled ammonia concentration feature value per unit volume.

[0055] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology, characterized in that, The system includes: The exhalation collection module is used to control patients with chronic kidney disease to perform a deep exhalation action of a set duration into the gas collection container, and simultaneously record the start and end times of the exhalation action to generate an exhalation volume marker. The ammonia sensing array module is used to activate the internal heating circuit at the start of the exhalation action, continuously sample the exhalation sample in the gas collection container, and dynamically adjust the bias voltage amplitude applied to the electrochemical ammonia sensor during the sampling process to obtain the original current response sequence that changes over time. The respiratory phase matching module is used to obtain the time alignment relationship between the expiratory volume marker and the original current response sequence, and to extract the current response baseline segment corresponding to the start time of the expiratory action based on the time alignment relationship, and calculate the average noise amplitude of the baseline segment. The feature compensation calculation module is used to perform peak retention processing on response peaks in the original current response sequence that exceed the set multiple of the average noise amplitude, and to perform expiratory volume normalization transformation on the retained peak sequence to obtain the characteristic value of expiratory ammonia concentration per unit volume. The status determination output module is used to compare the characteristic value of the ammonia concentration per unit volume of exhaled air with the pre-stored reference intervals for renal function staging, and output the risk level identifier of patients with chronic kidney disease.

2. The breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 1, characterized in that, The process by which the exhalation collection module controls a patient with chronic kidney disease to perform a deep exhalation of a set duration into a gas collection container, and simultaneously records the start and end times of the exhalation, generating an expiratory volume marker, includes: The infrared pyroelectric sensor in the exhalation acquisition module detects the airflow temperature change curve in front of the patient's mouth and nose. When the slope of the airflow temperature change curve exceeds the set slope threshold, it is marked as the start time of the exhalation action, and a timer is started to accumulate the time. The exhalation flow rate is measured in real time by the flow meter in the exhalation acquisition module. The exhalation flow rate is integrated over time. When the integration result reaches the preset target volume threshold, it is automatically marked as the end time of the exhalation action. The timer is stopped and the accumulated time value is read as the set duration. The expiratory volume marker is calculated by multiplying the accumulated time value by the average flow rate of the flow meter, and the start and end times of the expiratory action are associated with the expiratory volume marker as timestamps and stored.

3. The breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 2, characterized in that, When the exhalation acquisition module detects the airflow temperature change curve through the infrared pyroelectric sensor, if the flow rate value of the flow meter does not exceed the start flow threshold within the set judgment time after the start time of the exhalation action, it is determined as an invalid exhalation and the timer is reset to wait for the start time of the next exhalation action.

4. The breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 2, characterized in that, The ammonia sensing array module activates the internal heating circuit at the moment of termination of exhalation, continuously samples the exhaled breath in the gas collection container, and dynamically adjusts the bias voltage amplitude applied to the electrochemical ammonia sensor during the sampling process to obtain the original current response sequence changing over time. This process includes: After receiving the timestamp of the start of the exhalation action, the ammonia sensor array module sends a high-level enable signal to the internal heating circuit. It uses pulse width modulation to drive the heating circuit to rapidly heat up the temperature, so that the heating circuit raises the operating temperature of the electrochemical ammonia sensor to the set constant temperature operating point. During the heating process, the output current value of the electrochemical ammonia sensor is recorded once every sampling cycle to generate a heating transition current sequence. After the electrochemical ammonia sensor reaches its constant operating temperature, the bias voltage amplitude is gradually increased from the initial bias voltage value according to the preset scanning step size. After each step increase, the bias voltage is kept stable for a period of time, and the output current value at the end of the holding time is recorded to generate a bias scanning current sequence. The heating transition current sequence and the bias scanning current sequence are spliced ​​together in chronological order, and the sampling time label of each sampling point in the spliced ​​sequence is extracted to form the original current response sequence that changes with time.

5. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 4, characterized in that, If the rate of change of the output current value in three consecutive sampling cycles is less than the set rate of change threshold during the sampling process of the heating transition current sequence, the ammonia sensing array module will terminate the heating transition process in advance and jump to the bias voltage scanning stage.

6. The breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 4, characterized in that, The process by which the respiratory phase matching module acquires the time alignment relationship between the expiratory volume marker and the original current response sequence, and extracts the current response baseline segment corresponding to the start of the expiratory action based on the time alignment relationship, and calculates the average noise amplitude of the baseline segment includes: Extract the timestamps of the start and end times of the exhalation action stored in the exhalation volume marker. At the same time, find the sampling point with the smallest time difference from the timestamp of the start time of the exhalation action in the original current response sequence as the baseline start point and find the sampling point with the smallest time difference from the timestamp of the end time of the exhalation action as the baseline end point. Extract a sequence segment from the original current response sequence that is between a predetermined number of sampling points before the baseline start point and the baseline start point as the current response baseline segment. Calculate the arithmetic mean and standard deviation of the current values ​​of all sampling points within the current response baseline segment, and take three times the standard deviation as the average noise amplitude. The baseline start point, baseline end point, and average noise amplitude are packaged into a phase matching parameter set, and the phase matching parameter set is sent to the feature compensation calculation module.

7. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 6, characterized in that, When the respiratory phase matching module truncates the baseline segment of the current response, if a set number of sampling points before the baseline start point exceed the starting boundary of the original current response sequence, it will truncate from the first sampling point of the original current response sequence to the baseline start point.

8. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 6, characterized in that, The feature compensation calculation module performs peak preservation processing on response peaks in the original current response sequence that exceed a set multiple of the average noise amplitude, and performs expiratory volume normalization transformation on the preserved peak sequence to obtain the characteristic value of expiratory ammonia concentration per unit volume. This process includes: Receive the average noise amplitude in the phase matching parameter set, set the peak detection threshold to a set multiple of the average noise amplitude, traverse the sequence segments from the baseline start point to the baseline end point in the original current response sequence, mark the sequence regions that continuously exceed the peak detection threshold as a response peak region, and extract the maximum current value in each response peak region as the peak value of the response peak. The peak value of each response peak is preserved by replacing the current value of all sampling points in each response peak region with the peak value of that response peak, generating a preserved peak sequence, and counting the time position of each peak in the preserved peak sequence to form a peak time position vector. Based on the expiratory volume value in the expiratory volume marker, each peak in the preserved peak sequence is divided by the expiratory volume value to obtain the peak sequence corresponding to the unit volume as the characteristic value of the expiratory ammonia concentration per unit volume.

9. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 8, characterized in that, The process by which the state determination output module compares the characteristic value of the ammonia concentration per unit volume of exhaled air with the pre-stored reference intervals for renal function staging, and outputs the risk level identifier for patients with chronic kidney disease, includes: Read the pre-stored reference interval table for renal function staging from the internal memory. The reference interval table for renal function staging contains multiple renal function staging levels and the lower limit and upper limit of exhaled ammonia concentration corresponding to each renal function staging level. Each peak value of the characteristic value of expiratory ammonia concentration per unit volume is compared with the lower limit and upper limit of expiratory ammonia concentration in the reference interval table for renal function staging to determine the renal function stage to which the peak value falls, and the largest peak value in the characteristic value of expiratory ammonia concentration per unit volume is extracted as the representative concentration value. The representative concentration value is compared with the reference range for renal function staging. The staging level to which the representative concentration value belongs is used as the risk level identifier for patients with chronic kidney disease, and the risk level identifier is output to the display terminal in binary code form.

10. A breath ammonia detection system for chronic kidney disease based on intelligent sensing technology according to claim 9, characterized in that, The status determination output module outputs the risk level identifier to the display terminal in binary code form, and at the same time generates a text warning message corresponding to the risk level identifier, and sends it to the monitoring terminal through the wireless communication module.

Images