Non-inductive urination monitoring equipment for elderly men

Abstract

The invention relates to the technical field of physiological signal monitoring, in particular to elderly male non-inductive urination monitoring equipment which comprises a zipper action detection module, a urination place recognition module, a sound fragment extraction module, a behavior event analysis module and a physiological monitoring information generation module. According to the method, a behavior triggering time point is generated by detecting magnetic field change of a specific part, space coordinate data is collected to limit a urination place range, a multi-source data label is constructed by combining an environment sound feature matching result, and time information is associated to form a sortable event structure; behavior linkage data is formed by using a matching relationship among behavior time points, place numbers and sound segments, and the extracted linkage data has a clear event boundary and a discrimination basis, so that behavior monitoring is converted from single signal dependence to multi-dimensional signal fusion, the accuracy of urination behavior recognition and the structural integrity of event records are improved, and the accuracy of urination behavior recognition is improved. And the stability and applicability of behavior acquisition in a natural living state are enhanced.

Description

Technical Field

[0001] This invention relates to the field of physiological signal monitoring technology, and in particular to a non-invasive urination monitoring device for elderly men. Background Technology

[0002] The field of physiological signal monitoring technology involves technologies for the real-time acquisition, processing, and identification of various signals generated during human physiological activities through sensors and electronic devices. These primarily include electrocardiogram (ECG) monitoring, respiratory rate monitoring, body temperature monitoring, blood pressure and blood oxygen monitoring, electroencephalogram (EEG) monitoring, and monitoring of urinary system-related signals such as urination. This technology integrates microelectronics, biosensor technology, data communication technology, and signal recognition algorithms, aiming to achieve continuous, non-invasive, and automated monitoring of human vital signs. It has wide applications in medical care, elderly health management, remote monitoring, and auxiliary diagnosis and treatment of chronic diseases. Traditional non-invasive voiding monitoring devices for elderly men refer to devices used to monitor the urination behavior of elderly men in their natural daily life in a non-invasive and automated manner. They are mainly used for high-incidence situations of urination abnormalities caused by conditions such as benign prostatic hyperplasia. They are usually recorded by manually filling out a voiding diary, relying on the elderly man's memory and active recording of urination time, frequency, and duration, or by using large medical monitoring devices to monitor in real time through external detection elements such as catheters, pressure pads, or uroflowmeters. These methods are complicated to operate, have poor compliance, are uncomfortable to wear, or have obvious psychological resistance, and are not suitable for the high-frequency, continuous urination information collection in daily life.

[0003] Traditional voiding monitoring methods generally rely on subjective recording by users or data collection through external devices such as catheters and sensor pads. The operation process is complicated, the degree of cooperation is low, and data interruption often occurs due to incomplete recording or interference with behavioral privacy. The devices are conspicuous and have a noticeable foreign body sensation when worn, which can easily cause resistance and interruption of use, affecting the efficiency of continuous monitoring. They are also insufficient in extracting behavioral information in natural states, lack multi-dimensional information cross-validation methods, and cannot achieve effective verification and complete confirmation of behavioral recognition results. They are also difficult to cover the dynamic environmental characteristics when the behavior occurs, which limits the accurate judgment and systematic expression of behavioral events. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a non-intrusive urination monitoring device for elderly men. One aspect of the non-intrusive urination monitoring device for elderly men includes:

[0005] The zipper action detection module detects the change in the magnetic field between the zipper head and the magnetic plate, determines whether the preset action triggering conditions are met, records the time points when the conditions are met, and generates a set of action triggering time data. The urination location identification module activates the positioning component to collect continuous coordinates based on the action trigger time data set, filters the coordinate data located within the preset urination area boundary set, encodes the filtering results, and constructs a urination location marker set; The sound segment extraction module extracts environmental sound signals based on the action trigger time data set, compares the fitting degree of the sound segments with the frequency features of the preset urination process audio template, retains the sound segments that meet the preset threshold, and outputs a set of urination feature sound segments. The behavioral event analysis module calls the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data to construct a data group including location number, sound number, and time label, which is then reordered to form a behavioral monitoring data sequence. The physiological monitoring information generation module extracts the time, behavior type, and location number from the behavior monitoring data sequence, integrates them into structured record entries, summarizes all records, and generates an elderly male unconscious urination monitoring data record.

[0006] As a further embodiment of the present invention, the action trigger time data set includes trigger time point, displacement behavior label, and magnetic field change parameters; the urination location mark set includes location number, spatial coordinate point, and boundary area identifier; the urination characteristic sound segment set includes frequency fitting value, sound segment number, and audio feature label; the behavior monitoring data sequence includes location mark index, sound segment index, and time series label; and the elderly male non-sensory urination monitoring data record includes event occurrence time, behavior event type, and location number information.

[0007] As a further aspect of the present invention, the recorded time point that meets the conditions refers to the specific time point recorded when the zipper action detection module identifies the change in the magnetic field between the zipper head and the magnetic plate as reaching the preset action triggering standard.

[0008] As a further aspect of the present invention, the retained sound segments that meet the preset threshold refer to audio segments extracted from environmental sound signals that have a fitting degree with the frequency characteristics of the urination template that reaches or exceeds the set threshold.

[0009] As a further aspect of the present invention, the zipper action detection module includes: The magnetic field signal acquisition submodule obtains the magnetic field reference value when the magnetic suction component and magnetic sheet at the zipper head position are stationary, calls the magnetic sensor to collect the magnetic field change value during the movement of the zipper head, arranges the sampled data in time sequence, and generates magnetic field time series signal data. The magnetic field disturbance identification submodule calculates the change in magnetic flux density between adjacent sampling points based on the magnetic field time series signal data. If the change value is greater than the preset magnetic flux disturbance threshold, it is marked as an effective disturbance interval, and a set of magnetic disturbance interval data is obtained. The action time point extraction submodule extracts magnetic flux density extreme points as candidate nodes based on the magnetic disturbance interval data set, and filters out invalid points according to the preset magnetic flux extreme value threshold and duration threshold, thereby generating an action trigger time data set.

[0010] As a further aspect of the present invention, the urination location identification module includes: Based on the action trigger time data set, the positioning data acquisition submodule activates the positioning component to continuously record spatial coordinate point data within the corresponding time period, and numbers and stores all acquired data in chronological order to generate a time-related coordinate set. The spatial boundary filtering submodule, based on the time-related coordinate set, calls the spatial boundary range parameter in the urination area boundary set, calculates the spatial inclusion relationship between each set of coordinate points and the corresponding boundary, filters the coordinate points located within the boundary, and obtains a subset of coordinates within the location. The marker encoding generation submodule matches the spatial number parameters of the coordinate points according to the time sequence of the coordinate points for the subset of coordinates in the location, calls the preset marker encoding rules to number and label the coordinate groups, establishes unique identification data corresponding to the spatial location, and generates a set of urination location markers.

[0011] As a further aspect of the present invention, the sound segment extraction module includes: The audio data extraction submodule, based on the action trigger time data set, locates the continuous sound wave signal segment corresponding to the time point, synchronously extracts the original sound sample from the ambient sound channel according to the time range, arranges it frame by frame and stores it as structured audio data, and obtains the time-aligned audio set. The frequency feature matching submodule calculates the frequency energy distribution vector for each audio segment based on the time-aligned audio set, and performs a cosine fitting degree operation on the vector with the preset standard frequency vector corresponding to the urination process audio template to obtain the frequency fitting degree value between all audio segments and the template, and generates an audio matching fitting degree sequence. The fitting sound filtering submodule calls the audio matching fitting degree sequence, compares the fitting degree value with the preset fitting degree threshold segment by segment, filters audio segments with a fitting degree higher than the fitting degree threshold, and reassembles them according to the original time label to output a set of urination characteristic sound segments.

[0012] As a further aspect of the present invention, the behavioral event analysis module includes: The behavior data aggregation submodule acquires data from the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data, extracts the corresponding location number, sound number, and time tag, and performs horizontal merging by behavior event to generate a behavior joint data structure. The time tag synchronization submodule performs a unified format conversion on the time tags of all behavior data items according to the behavior joint data structure, and establishes a sorting index based on the timestamp field to re-sort the original records and obtain a time series sorting index table. The monitoring sequence construction submodule calls the time series sorting index table, reorganizes the behavior joint data structure according to the index position, organizes the behavior event data in chronological order and adds a continuous label field to generate a behavior monitoring data sequence.

[0013] As a further aspect of the present invention, the physiological monitoring information generation module includes: The behavior field extraction submodule obtains the event content in the behavior monitoring data sequence, extracts the time tag, behavior event type field and location number field corresponding to each record, performs field format regularization and field value standardization processing, and generates an event field structure set; The record entry construction submodule indexes and combines the content of each group of fields according to the event field structure set, and concatenates the time tag, event type and location number into a single structured urination record format according to the specified structure, and organizes them in order to generate a set of urination record entries; The data archiving output submodule calls the set of voiding record entries, rearranges the record content according to the time index, writes all entries into the archiving buffer, sets the archiving file structure and performs the archiving operation to obtain the non-invasive voiding monitoring data records of elderly men.

[0014] As a further aspect of the present invention, the field format regularization and field value standardization processing refers to unifying the format style of extracted field contents such as time, event type and location number, and converting field values ​​into a unified format expression method that conforms to preset standards.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By detecting changes in the magnetic field at specific locations to generate behavioral trigger time points, collecting spatial coordinate data to define the scope of urination locations, and combining environmental sound feature matching results to construct multi-source data labels, and associating time information to form a sortable event structure, behavioral linkage data is constructed by utilizing the matching relationship between behavioral time points, location numbers, and sound segments. The extracted linkage data has clear event boundaries and discrimination criteria, realizing the transformation of behavior monitoring from single signal dependence to multi-dimensional signal fusion, improving the accuracy of urination behavior recognition and the structural integrity of event records, and enhancing the stability and applicability of behavior collection in natural living conditions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the overall process of the present invention; Figure 2 This is a schematic diagram of the framework of the present invention; Figure 3 This is a flowchart of the zipper action detection module in this invention; Figure 4 This is a flowchart of the urination site identification module in this invention; Figure 5 This is a flowchart of the sound segment extraction module in this invention; Figure 6 This is a flowchart of the behavioral event analysis module in this invention; Figure 7 This is a flowchart of the physiological monitoring information generation module in this invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] This invention provides a non-intrusive urination monitoring device for elderly men, such as... Figure 1-2 The schematic diagram shown is of a non-invasive voiding monitoring device for elderly men, including: The zipper action detection module detects the change in the magnetic field generated by the magnetic component at the zipper head position, determines whether the displacement behavior between the zipper head and the magnetic piece meets the preset action triggering conditions, records the corresponding time point when the conditions are met, and outputs a set of action triggering time data. The urination site identification module is based on the action trigger time data set, activates the positioning component to continuously collect spatial coordinates, filters spatial coordinate data within the boundary set of the urination area, marks and encodes the data that meets the conditions, and generates a urination site mark set. The sound segment extraction module extracts continuous environmental sound signals within the corresponding time period based on the action trigger time data set, compares the frequency feature fitting degree between each sound signal and the urination process audio template, retains the sound segments whose fitting degree meets the preset threshold, and outputs a set of urination feature sound segments. The behavioral event analysis module constructs joint behavioral data that includes urination location number, sound segment number and time label based on the set of urination location markers, the set of urination characteristic sound segments and the set of action trigger time data, and rearranges them in chronological order to form a behavioral monitoring data sequence. The physiological monitoring information generation module extracts the occurrence time, behavior type and associated location number of each event from the behavioral monitoring data sequence, integrates them into urination record entries with event structure fields, and archives all record entries to generate unconscious urination monitoring data records for elderly men.

[0024] The action trigger time data set includes trigger time point, displacement behavior label, and magnetic field change parameters; the urination location mark set includes location number, spatial coordinate point, and boundary area identifier; the urination characteristic sound segment set includes frequency fitting value, sound segment number, and audio feature label; the behavior monitoring data sequence includes location mark index, sound segment index, and time series label; and the elderly male non-sensory urination monitoring data record includes event occurrence time, behavior event type, and location number information.

[0025] Specifically, such as Figure 2 , 3 As shown, the zipper action detection module includes: The magnetic field signal acquisition submodule obtains the magnetic field reference value when the magnetic suction component and magnetic sheet at the zipper head position are stationary, calls the magnetic sensor to collect the magnetic field change value during the movement of the zipper head, arranges the sampled data in time sequence, and generates magnetic field time series signal data. The front-end sensing unit first initializes and configures the built-in microelectromechanical Hall effect sensor, setting it to continuously monitor the magnetic flux density in three-dimensional space at a sampling frequency of 100Hz. It establishes a communication connection with the sensor through an internal high-speed bus. During the initial stabilization phase after startup, it executes a static benchmark calibration program lasting 2 seconds. During this period, it continuously reads 200 sets of triaxial magnetic field data in the zipper's closed and stationary state. It performs arithmetic mean calculation on the X-axis, Y-axis, and Z-axis readings, and solidifies the three mean values ​​into a magnetic field benchmark value vector. For example, the benchmark value vector determined by the calculation is 15 microtesla for the X-axis, 22 microtesla for the Y-axis, and 38 microtesla for the Z-axis. Then, it enters the dynamic real-time monitoring mode, continuously capturing the magnetic field fluctuations caused by the displacement of the magnetic suction component relative to the magnetic sheet during the vertical movement of the zipper head. Each frame of data collected includes a high-precision timestamp and the triaxial magnetic induction intensity reading at that moment. These data are written into a first-in-first-out circular buffer in chronological order to construct a continuous magnetic field time series signal data.

[0026] The magnetic field disturbance identification submodule calculates the change in magnetic flux density between adjacent sampling points based on magnetic field time series signal data. If the change value is greater than the preset magnetic flux disturbance threshold, it is marked as an effective disturbance interval, and a set of magnetic disturbance interval data is obtained. To extract effective action signals from complex environmental magnetic fields, the magnetic field time series signal data undergoes in-depth processing. An internally preset sliding time window with a length of 5 sampling points is used. Vector magnitude calculation is performed on the triaxial data of each sampling point within the window, specifically calculating the squares of the X, Y, and Z axis readings. These three squares are then summed and their square roots are taken to obtain the total magnetic flux density magnitude at that moment. Next, an adjacent difference operation is performed: the magnetic flux density magnitude at the current sampling moment is subtracted from the magnitude at the previous sampling moment, and the absolute value is taken to obtain the change in magnetic flux density. A preset magnetic flux disturbance threshold of 8 microtesla is then applied, and the real-time calculated change value is compared point-by-point with this threshold. If the change value is greater than or equal to 8 microtesla for 5 consecutive sampling points... If the value is in microtesla, it indicates that there is a significant magnetic field disturbance within the current time period. The start and end times of the data segment are marked, and the interval is extracted as the effective disturbance interval. The data is then aggregated to generate a set of magnetic disturbance interval data. For example, if the magnitude calculation result is 88 microtesla at a certain moment and 80 microtesla at the previous moment, the absolute value of the difference between the two is 8 microtesla. If the subsequent data continues to meet the condition, the disturbance is confirmed to have been captured.

[0027] The action time point extraction submodule extracts magnetic flux density extreme points as candidate nodes based on the magnetic disturbance interval data set, and filters out invalid points according to the preset magnetic flux extreme value threshold and duration threshold, and generates an action trigger time data set. Based on the acquired magnetic disturbance interval data set, the processing logic further performs feature verification and extreme value localization to pinpoint the time of action occurrence. It traverses the magnetic flux density modulus sequence within each magnetic disturbance interval and uses a numerical comparison algorithm to find the maximum value point in the sequence as a candidate node. To eliminate false alarms caused by daily human activities or environmental electromagnetic interference, a magnetic flux extreme value threshold of 45 microtesla and a duration threshold of 0.6 seconds are introduced as dual judgment criteria. First, it checks whether the modulus value of the candidate node exceeds 45 microtesla. Second, it calculates whether the duration of the disturbance interval exceeds 0.6 seconds. Only when both conditions are met simultaneously is the candidate node confirmed as the action trigger point, and finally, a precise action trigger time data set is output.

[0028] Specifically, such as Figure 2 , 4 As shown, the urination location recognition module includes: The positioning data acquisition submodule activates the positioning component to continuously record spatial coordinate point data within the corresponding time period based on the action trigger time data set, and numbers and stores all acquired data in chronological order to generate a time-related coordinate set. The system is activated directly by time commands from the action trigger time data set. Once a confirmed action trigger time is received, the positioning tag maintains a low-power standby mode during system operation and periodically broadcasts its spatial location at a frequency of 20Hz. After receiving the signal, the base station calculates the distance from the tag to each base station using a time-of-flight ranging algorithm and uses trilateration to calculate the tag's three-dimensional coordinates in the indoor coordinate system. The acquisition window is set from 5 seconds before the action trigger to 30 seconds after the trigger. All coordinate points within this time period are continuously recorded, and each acquired coordinate data is time-synchronized. The coordinate points are strictly bound to the acquisition time and assigned a unique serial number according to the time sequence. This generates a set of associated coordinates containing time and spatial information. For example, the coordinate data recorded 2 seconds after the action trigger is 200 cm on the X-axis and 350 cm on the Y-axis. This data is immediately stored in the set.

[0029] The spatial boundary filtering submodule uses the time-related coordinate set to call the spatial boundary range parameter in the urination area boundary set, calculates the spatial inclusion relationship between each set of coordinate points and the corresponding boundary, filters the coordinate points located within the boundary, and obtains the coordinate subset within the site. To determine whether a user is located within a designated urination area, a set of urination area boundaries defined by the vertex coordinates of a two-dimensional polygon is pre-stored. Taking the master bedroom bathroom as an example, its boundary is defined as a rectangular area enclosed by four vertex coordinates (100, 100), (300, 100), (300, 400), and (100, 400), in centimeters. A ray casting algorithm is executed on each coordinate point in the time-related coordinate set, drawing a virtual ray from the coordinate point to be measured in the positive X-axis direction. The total number of intersections between the ray and each side of the polygon is calculated. If the total number of intersections is odd, the coordinate point is determined to be inside the polygon; if it is even, it is determined to be outside. All coordinate points are traversed, and only those points determined to be within any valid urination area boundary are retained, while invalid points located in the corridor or bedroom area are filtered out, thereby obtaining a subset of coordinates within the location.

[0030] The marker coding generation submodule targets a subset of coordinates within a location, matches the spatial number parameters of the coordinate points according to their time sequence, calls preset marker coding rules to number and label the coordinate groups, establishes unique identifier data corresponding to the spatial location, and generates a set of markers for urination locations. The selected subset of coordinates within the location is semantically mapped and encoded. An internal spatial number mapping table is maintained, for example, the master bathroom corresponds to code 01 and the guest bathroom corresponds to code 02. Each data item in the subset of coordinates within the location is read, its boundary area is identified and the corresponding area code is extracted. Then, the preset marking and encoding rules are called to combine and concatenate the area code, the original serial number and the status code within the area. For example, if the serial number of a user at a certain point in area 01 is 001, a unique identifier string in the format "01-001-status code" will be generated. The generated identifier codes are arranged in chronological order to construct unique identifier data corresponding to the spatial location, that is, the set of urination location markers.

[0031] Specifically, such as Figure 2 , 5 As shown, the audio segment extraction module includes: The audio data extraction submodule is based on the action trigger time data set, locates the continuous sound wave signal segment corresponding to the time point, synchronously extracts the original sound sample from the ambient sound channel according to the time range, arranges it frame by frame and stores it as structured audio data, and obtains the time-aligned audio set. Using action trigger time as an index, key segments are extracted from a continuous environmental recording stream. A high-sensitivity microphone array records ambient sound around the clock at a sampling rate of 48,000 Hz, and the data is temporarily stored in a cyclically overwritten memory block. When an action trigger time point is received, the start time of the capture window is calculated to be the trigger time minus 2 seconds, and the end time is the trigger time plus 15 seconds, thus locking a time range of 17 seconds. Based on these two time anchor points, the corresponding pointer positions are located in the memory block, and the pulse code modulation audio data within this time period is completely copied. Then, the original audio is processed into frames, with a frame length of 25 milliseconds and a frame shift of 10 milliseconds. Each frame of data is multiplied by a Hamming window function to eliminate signal abrupt changes at frame edges. Finally, the processed frame data is arranged in chronological order to obtain a time-aligned audio set.

[0032] The frequency feature matching submodule calculates the frequency energy distribution vector for each audio segment based on the time-aligned audio set, and performs a cosine fit operation on the vector with the preset standard frequency vector corresponding to the urination process audio template to obtain the frequency fit value between all audio segments and the template, and generates an audio matching fit sequence. Next, a deep analysis of each frame of audio data is performed. First, a Fast Fourier Transform is performed on the time-domain signal of each frame to convert it into a frequency-domain signal and the power spectrum is calculated. Then, the power spectrum is filtered using a Mel filter bank to simulate the auditory characteristics of the human ear. Then, the first 13 coefficients are extracted through Discrete Cosine Transform to form a Mel frequency cepstral coefficient feature vector. The template contains a pre-stored frequency feature template of standard urination sound. This template is based on the mean of a 13-dimensional feature vector obtained from a large number of samples. Cosine similarity calculation is performed. For the feature vector and template vector of the frame to be tested, the dot product of the two vectors is first calculated, and then the magnitude of the two vectors is calculated separately. Finally, the dot product result is divided by the product of the two magnitudes to obtain a similarity value between 0 and 1. This operation is repeated for each frame in the set to generate an audio matching fit sequence. For example, if the dot product is 50 and the magnitude product is 50, the fit is 1.

[0033] The fitting sound filtering submodule calls the audio matching fit degree sequence, compares the fit degree value with the preset fit degree threshold segment by segment, filters audio segments with a fit degree higher than the fit degree threshold, and reassembles them according to the original time label to output a set of urination feature sound segments. The system determines whether a real urination sound exists based on the fit sequence. A preset fit threshold of 0.82 is used, which is determined by analyzing a large number of urination audio samples containing background noise. The system iterates through the audio matching fit sequence and compares each fit value with 0.82. If the fit value of a frame is greater than or equal to 0.82, it is marked as a high-confidence frame. To prevent occasional noise interference, a smoothing filter is applied. Only when at least 10 consecutive high-confidence frames appear is the time segment considered a valid urination sound segment. These consecutive segments that meet the conditions are extracted and recombined according to the original time labels to output a set of urination feature sound segments.

[0034] Specifically, such as Figure 2 , 6 As shown, the behavioral event analysis module includes: The behavior data aggregation submodule acquires data from the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data, extracts the corresponding location number, sound number, and time tag, and performs horizontal merging by behavior event to generate a behavior joint data structure. It is responsible for aggregating data streams scattered across different dimensions into a unified behavior record. It reads the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data from memory. The aggregation logic uses the action trigger time as the core primary key and sets a time association window with a duration of 30 seconds before and after. For each action trigger record, it traverses the other two sets and searches for data items that fall within the time window. Specifically, it checks whether there are valid location markers and high-fit sound segments within 30 seconds before and after the action. If the three have an overlap or containment relationship on the time axis, they are regarded as components of the same behavior event. The location number, sound feature number, and time label are extracted and merged horizontally to generate a behavior joint data structure containing multi-dimensional attributes.

[0035] The time tag synchronization submodule performs a unified format conversion on the time tags of all behavioral data items based on the behavioral federated data structure, and builds a sorting index based on the timestamp field to re-sort the original records, thereby obtaining a time series sorting index table. The merged data undergoes time-series calibration and index construction. Due to different data sources, the time formats may be inconsistent. First, the timestamps of all records are uniformly converted into a standard universal timestamp format. Then, a sorted index is built based on the size of the timestamp values. The merge sort algorithm is used to traverse and compare all the data structures, placing records with smaller timestamps first and records with larger timestamps last. After multiple rounds of comparison and exchange, a time-series sorted index table is generated that is strictly arranged in chronological order, providing a foundation for subsequent continuous analysis.

[0036] The monitoring sequence construction submodule calls the time series sorting index table, reorganizes the behavior joint data structure according to the index position, organizes the behavior event data in chronological order and adds a continuous label field to generate a behavior monitoring data sequence; Based on the sorting index table, independent urination processes are identified and marked. Each behavior record is read sequentially and a session number field is introduced, with an initial value of 1. The time difference between the current record and the previous record is calculated and compared with a preset event segmentation threshold of 300 seconds. This is based on the fact that the interval between a single urination process of the elderly usually does not exceed 5 minutes. If the time difference is less than 300 seconds, it is determined that the two records belong to different stages of the same urination behavior and are assigned the same session number. If the time difference is greater than 300 seconds, the session number value is incremented by 1, indicating the start of a new urination behavior. The data with session numbers are organized in chronological order to generate a behavior monitoring data sequence. For example, if the first record is 1000 seconds and the second is 1100 seconds, the difference of 100 seconds is less than 300 seconds, so they are merged into the same session.

[0037] Specifically, such as Figure 2 , 7 As shown, the physiological monitoring information generation module includes: The behavior field extraction submodule obtains the event content from the behavior monitoring data sequence, extracts the time tag, behavior event type field and location number field corresponding to each record, performs field format regularization and field value standardization processing, and generates an event field structure set; The behavior monitoring data sequence is parsed to extract business fields with medical reference value. Each group of data with the same session number is traversed, and aggregation calculation and field extraction are performed. First, the timestamp of the first record in the session is extracted and converted into a readable format of hours, minutes, and seconds as the start time. Second, all records in the session containing sound segments are counted, and the duration of each sound segment is accumulated to obtain the total urination duration. Then, the location number field is extracted and the built-in dictionary table is queried to convert it into a specific location name, such as the master bedroom bathroom. Finally, the behavior type field is set to urination, and the extracted values ​​are formatted to generate a standardized event field structure set.

[0038] The record entry construction submodule indexes and combines the content of each group of fields according to the event field structure set, and concatenates the time tag, event type and location number into a single structured voiding record format according to the specified structure, and organizes them in order to generate a set of voiding record entries; Standardized fields are assembled into final data entries. According to a preset data protocol, fields such as start time, total urination duration, location, and behavior type are encapsulated into key-value pairs. To ensure the uniqueness and traceability of the data, a globally unique identifier is generated for each encapsulated record. At the same time, a CRC32 checksum algorithm is executed on the record content to convert all characters into a byte stream, calculate a 32-bit integer checksum, and append it to the end of the record. This process generates a complete set of urination record entries. For example, an entry may contain information such as time 10:30:15, duration 25 seconds, location master bedroom bathroom, and checksum.

[0039] The data archiving output submodule calls the set of voiding record entries, rearranges the record content by time index, writes all entries to the archiving buffer, sets the archiving file structure and performs archiving operations to obtain the voiding monitoring data records of elderly men without feeling the impact. It is responsible for persistently storing the processed records to non-volatile storage media, grouping the set of voiding record entries by date, and generating a separate CSV archive file each day. First, a buffer is allocated in memory, and the records to be written are arranged in chronological order. Then, the file interface is called to create or open the archive file for the day, and the column header is written to the first line. Next, the data in the buffer is converted into comma-separated strings line by line and written to the file. Every 50 records written, a forced refresh operation is performed to synchronize the memory data to the hard disk to prevent data loss in case of power failure. Finally, the archiving operation is completed, and the data records of the non-invasive voiding monitoring of elderly men are available for doctors or family members to query.

[0040] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A non-intrusive urination monitoring device for elderly men, characterized in that, include: The zipper action detection module detects the change in the magnetic field between the zipper head and the magnetic plate, determines whether the preset action triggering conditions are met, records the time points when the conditions are met, and generates a set of action triggering time data. The urination location identification module activates the positioning component to collect continuous coordinates based on the action trigger time data set, filters the coordinate data located within the preset urination area boundary set, encodes the filtering results, and constructs a urination location marker set; The sound segment extraction module extracts environmental sound signals based on the action trigger time data set, compares the fitting degree of the sound segments with the frequency features of the preset urination process audio template, retains the sound segments that meet the preset threshold, and outputs a set of urination feature sound segments. The behavioral event analysis module calls the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data to construct a data group including location number, sound number, and time label, which is then reordered to form a behavioral monitoring data sequence. The physiological monitoring information generation module extracts the time, behavior type, and location number from the behavior monitoring data sequence, integrates them into structured record entries, summarizes all records, and generates an elderly male unconscious urination monitoring data record.

2. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that: The action trigger time data set includes trigger time point, displacement behavior label, and magnetic field change parameters; the urination location mark set includes location number, spatial coordinate point, and boundary area identifier; the urination characteristic sound segment set includes frequency fitting value, sound segment number, and audio feature label; the behavior monitoring data sequence includes location mark index, sound segment index, and time series label; and the elderly male non-sensory urination monitoring data record includes event occurrence time, behavior event type, and location number information.

3. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that: The recorded time point that meets the conditions refers to the time point when the zipper action detection module detects that the change in the magnetic field between the zipper head and the magnetic plate reaches the preset action trigger standard.

4. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that: The retained sound segments that meet the preset threshold refer to audio segments extracted from environmental sound signals that have a fitting degree with the frequency characteristics of the urination template that reaches or exceeds the preset threshold.

5. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that, The zipper action detection module includes: The magnetic field signal acquisition submodule obtains the magnetic field reference value when the magnetic suction component and magnetic sheet at the zipper head position are stationary, calls the magnetic sensor to collect the magnetic field change value during the movement of the zipper head, arranges the sampled data in time sequence, and generates magnetic field time series signal data. The magnetic field disturbance identification submodule calculates the change in magnetic flux density between adjacent sampling points based on the magnetic field time series signal data. If the change value is greater than the preset magnetic flux disturbance threshold, it is marked as an effective disturbance interval, and a set of magnetic disturbance interval data is obtained. The action time point extraction submodule extracts magnetic flux density extreme points as candidate nodes based on the magnetic disturbance interval data set, and filters out invalid points according to the preset magnetic flux extreme value threshold and duration threshold, thereby generating an action trigger time data set.

6. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that, The urination site identification module includes: Based on the action trigger time data set, the positioning data acquisition submodule activates the positioning component to continuously record spatial coordinate point data within the corresponding time period, and numbers and stores all acquired data in chronological order to generate a time-related coordinate set. The spatial boundary filtering submodule, based on the time-related coordinate set, calls the spatial boundary range parameter in the urination area boundary set, calculates the spatial inclusion relationship between each set of coordinate points and the corresponding boundary, filters the coordinate points located within the boundary, and obtains a subset of coordinates within the location. The marker encoding generation submodule matches the spatial number parameters of the coordinate points according to the time sequence of the coordinate points for the subset of coordinates in the location, calls the preset marker encoding rules to number and label the coordinate groups, establishes unique identification data corresponding to the spatial location, and generates a set of urination location markers.

7. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that, The sound segment extraction module includes: The audio data extraction submodule, based on the action trigger time data set, locates the continuous sound wave signal segment corresponding to the time point, synchronously extracts the original sound sample from the ambient sound channel according to the time range, arranges it frame by frame and stores it as structured audio data, and obtains the time-aligned audio set. The frequency feature matching submodule calculates the frequency energy distribution vector for each audio segment based on the time-aligned audio set, and performs a cosine fitting degree operation on the vector with the preset standard frequency vector corresponding to the urination process audio template to obtain the frequency fitting degree value between all audio segments and the template, and generates an audio matching fitting degree sequence. The fitting sound filtering submodule calls the audio matching fitting degree sequence, compares the fitting degree value with the preset fitting degree threshold segment by segment, filters audio segments with a fitting degree higher than the fitting degree threshold, and reassembles them according to the original time label to output a set of urination characteristic sound segments.

8. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that, The behavioral event analysis module includes: The behavior data aggregation submodule acquires data from the set of urination location markers, the set of urination characteristic sound segments, and the set of action trigger time data, extracts the corresponding location number, sound number, and time tag, and performs horizontal merging by behavior event to generate a behavior joint data structure. The time tag synchronization submodule performs a unified format conversion on the time tags of all behavior data items according to the behavior joint data structure, and establishes a sorting index based on the timestamp field to re-sort the original records and obtain a time series sorting index table. The monitoring sequence construction submodule calls the time series sorting index table, reorganizes the behavior joint data structure according to the index position, organizes the behavior event data in chronological order and adds a continuous label field to generate a behavior monitoring data sequence.

9. The non-intrusive urination monitoring device for elderly men according to claim 1, characterized in that, The physiological monitoring information generation module includes: The behavior field extraction submodule obtains the event content in the behavior monitoring data sequence, extracts the time tag, behavior event type field and location number field corresponding to each record, performs field format regularization and field value standardization processing, and generates an event field structure set; The record entry construction submodule indexes and combines the content of each group of fields according to the event field structure set, and concatenates the time tag, event type and location number into a single structured urination record format according to the specified structure, and organizes them in order to generate a set of urination record entries; The data archiving output submodule calls the set of voiding record entries, rearranges the record content according to the time index, writes all entries into the archiving buffer, sets the archiving file structure and performs the archiving operation to obtain the non-invasive voiding monitoring data records of elderly men.

10. The non-intrusive urination monitoring device for elderly men according to claim 9, characterized in that: The field format regularization and field value standardization process refers to unifying the format style of extracted fields such as time, event type, and location number, and converting field values ​​into a unified format expression that conforms to preset standards.

Images