Goat breeding sign data real-time monitoring method and system based on internet of things

Abstract

This invention belongs to the field of livestock breeding monitoring technology and discloses a method and system for real-time monitoring of vital sign data in goat farming based on the Internet of Things (IoT). The method includes: updating the location and number of base stations based on historical vital sign data and historical location data; receiving and performing in-depth analysis of uploaded vital sign data and location data in real time through the current base station to determine the daily steady-state health status of the goats; applying stress tests to the goats to determine their stress response health status and the status of smart devices; and generating a graded treatment strategy based on the stress response health status and the daily steady-state health status. This invention significantly improves the real-time performance, accuracy, and reliability of goat farming vital sign data monitoring through the coordinated operation of multiple links, including adaptive base station deployment, multimodal in-depth analysis, stress test linkage verification, and device status diagnosis.

Description

Technical Field

[0001] This invention relates to the field of livestock breeding monitoring technology, and more specifically, to a method and system for real-time monitoring of vital signs data in goat farming based on the Internet of Things. Background Technology

[0002] In recent years, with the rapid development of IoT technology, its application in livestock farming has become increasingly widespread. In goat farming, real-time monitoring of goat vital signs and behavioral activity helps to promptly detect health abnormalities, reduce the risk of disease transmission, and improve farming efficiency. Existing technologies include monitoring solutions that collect vital sign data such as heart rate, body temperature, and activity levels in goats using devices like smart collars, and then upload this data to base stations or servers via wireless communication.

[0003] However, existing technologies still have the following shortcomings: The deployment location and number of base stations are usually fixed or determined by human experience, making it difficult to adapt to the dynamic changes in the goat's activity area. With the changing seasons, adjustments to grazing areas, or changes in the behavior of the goats, the original base station layout may result in signal coverage blind spots or base station redundancy, leading to increased data packet loss rate, increased transmission delay, and even the inability to collect goat vital signs data in some areas.

[0004] Current technologies for assessing the health status of goats primarily rely on the analysis of single vital signs. This single-dimensional assessment is susceptible to interference from sensor noise, equipment malfunctions, or individual differences, resulting in high false alarm and false negative rates. Furthermore, existing technologies struggle to distinguish between goat health abnormalities and smart device malfunctions, potentially leading to fatigue among farm managers due to false alarms or delayed responses to actual diseases. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: A method for real-time monitoring of vital signs data in goat farming based on the Internet of Things, the method comprising: S1. Obtain the vital signs data and location data of the goat during the target historical period, and record them as historical vital signs data and historical location data, respectively; update the base station location and number of base stations based on the historical vital signs data and historical location data; S2. Real-time collection of vital signs and location data of goats based on wearable smart devices, and lightweight pre-analysis to obtain pre-analysis results; sorting and uploading the unuploaded vital signs and location data according to the pre-analysis results to obtain a data transmission sorting table; S3. Receive and perform in-depth analysis of the vital signs and location data uploaded by the current base station in real time to determine the daily homeostatic health status of the goats. S4. Apply stress test to goats at a target time using target stress test items, obtain vital sign data and location data of goats within a time interval centered on the target time and with a preset duration as the step size, determine the stress response health status and smart device status, and generate a graded treatment strategy based on the stress response health status and daily steady-state health status.

[0006] Further, S1 includes: S1.1 For each deployed base station, extract the location data of the goats received by the base station during the target historical period and record it as historical location data; for each base station, calculate the average coordinates of all location points of the historical vital signs data, and the average coordinates are the historical service centroid of the base station; S1.2. At a preset fixed time, for each base station, obtain the distance between its current physical location and the historical service center, and determine whether to update the base station location based on the distance; S1.3 Identify historical sheep flock gathering areas based on the historical positioning data; determine whether to add base stations based on the location of historical sheep flock gathering areas and the locations of existing base stations. S1.4 Count the total number of data packets received by each base station in the target historical period, determine the utilization rate of each base station based on the total number of data packets, determine whether there are redundant base stations based on the utilization rate, and delete redundant base stations if they exist.

[0007] Furthermore, the step of determining whether to update the base station location based on the distance includes: if the distance is greater than a preset movement trigger threshold, determining to update the base station location and generating a movement command to instruct the base station to move half the distance in the direction of the historical service center of gravity; after the movement is completed, re-evaluating the distance between the new location of the base station and the historical service center of gravity, and if it is still greater than the threshold, moving half again until it enters the movement trigger threshold to obtain the current base station location.

[0008] Furthermore, the step of identifying historical sheep flock gathering areas based on the historical location data, and determining whether to add base stations based on the location of historical sheep flock gathering areas and existing base stations, includes: Obtain a floor plan of the farm's activity area and divide the farm's activity area into a regular grid with a side length of L meters; Using historical location data, the total dwell time of goats in each grid was counted, and the average dwell time of all grids was calculated. If the total dwell time of a certain grid exceeds twice the average dwell time, the grid is determined to be a historical sheep gathering area. If the distance from the center of the grid to any deployed base station is greater than the Bluetooth reliable communication radius, the grid is determined to be a coverage gap hotspot, and a new base station is added at the center of the grid.

[0009] Further, S2 includes: S2.1 Based on the smart device worn on the goat, the goat's vital signs and location data are collected in real time at the first sampling frequency; S2.2 On the smart device, perform lightweight pre-analysis on the vital signs data and location data respectively to obtain pre-analysis results. The pre-analysis results include three levels: normal, attention, and abnormal, and corresponding feature identifiers. S2.3. Based on the levels in the pre-analysis results, dynamically sort the unuploaded vital sign data and location data to generate a data transmission sorting table; S2.4 The smart device uploads vital sign data and location data to the matched base station via Bluetooth channel in sequence according to the data transmission sorting table.

[0010] Further, S3 includes: S3.1 Receive vital sign data and location data uploaded from different smart devices through the current base station, and align and clean the data according to the timestamp and the individual goat identifier; S3.2. Perform in-depth analysis on the cleaned data to obtain temporal features; construct a multimodal temporal feature matrix based on the temporal features; input the multimodal temporal feature matrix into a pre-trained health status assessment model; output the health status classification results of goats based on the health status assessment model, which are denoted as daily steady-state health status, including four states: healthy, sub-healthy, suspected onset of disease, and acute onset of disease. S3.3 Generate corresponding processing suggestions based on the daily steady-state health status.

[0011] Further, S4 includes: S4.1. Every preset period, select a preset number of candidate stress test items from the stress test library; each candidate stress test item is associated with a historical stress response feature library; S4.2. Select the candidate stress test items from a preset perspective to obtain the target stress test items; S4.3. Apply stress test to goats at the target time using the target stress test items; while applying the stress test, mark the test start timestamp by synchronously triggering a signal; S4.4. Obtain the vital signs and location data of the goats within a time interval centered on the target time and with a preset duration as the step size, and determine the stress response health status based on the vital signs and location data of the goats within the time interval. S4.5. Establish a linkage verification matrix based on the stress response health status and the daily steady-state health status, perform consistency judgment based on the linkage verification matrix, and generate a graded treatment strategy; the rows of the linkage verification matrix correspond to the daily steady-state health status, and the columns correspond to the stress response health status.

[0012] Furthermore, S4 also includes: Extract vital sign data within a preset length window before the stress test is applied, calculate the average heart rate and average activity intensity of the goats within this window, and record them as baseline values; extract vital sign data within a preset length window after the stress test is applied, calculate the maximum heart rate and maximum activity intensity within this window, and record them as peak stress response. If the change in the peak stress response relative to the baseline value simultaneously meets the preset conditions: the increase in heart rate exceeds the preset minimum sensitivity threshold, and the increase in activity intensity exceeds the preset minimum sensitivity threshold, then the smart device is preliminarily judged to be normal. If the change in the stress response peak does not reach the preset condition, the current daily homeostatic health status of the goat is queried; if the daily homeostatic health status is "healthy" or "sub-healthy", the smart device is determined to be faulty; if the daily homeostatic health status is "suspected illness" or "emergency illness", the smart device is determined to be suspicious.

[0013] Check the time series of positioning data within the time interval. If the continuous loss of positioning data exceeds the preset loss duration, or the calculated movement speed between adjacent positioning points exceeds the physiological limit speed of a goat, directly output the smart device status as faulty.

[0014] Furthermore, the method for determining the target time includes: Obtain the location of each current base station and its corresponding base station coverage radius; draw a circular coverage boundary with the current base station location as the center and the base station coverage radius as the radius; for the overlapping area of ​​adjacent current base stations, use the perpendicular bisector of the line connecting the two current base stations as the coverage boundary dividing line to form the theoretical coverage polygon of each base station; The smart device collects the Received Signal Strength Indicator (RSSI) of the current base station with which it has established a Bluetooth connection in real time at the second sampling frequency; when the RSSI value is detected to continuously drop from above the preset high threshold to below the preset low threshold and the duration of the drop reaches or exceeds the preset duration threshold, it is determined that the goat is currently moving from the strong coverage area of ​​the base station towards the coverage boundary and crossing the boundary, and the moment of crossing the boundary is determined. The moment of crossing the boundary is recorded as the candidate target moment. At the same time, the crossing direction is determined based on the base station connection status before and after the RSSI value decreases: if the base station is connected before crossing and is still connected after crossing and the RSSI value is continuously lower than the preset low threshold, it is determined to be the direction of leaving the coverage area; if the base station is connected before crossing and is switched to base station B after crossing, it is determined to be the cross-base station transfer direction. Prioritize the boundary crossing time in the direction of leaving the coverage area as the final target time; if there is no boundary crossing event in the direction of leaving the coverage area within the current preset period, select the boundary crossing time in the direction of cross-base station transfer; if neither exists, keep the target time of the previous period or use the default time.

[0015] Another objective of this invention is to provide a real-time monitoring system for vital sign data in goat farming based on the Internet of Things, the system comprising: The base station information update module is used to acquire the vital signs data and location data of the goat during the target's historical period, which are respectively recorded as historical vital signs data and historical location data; and to update the base station location and number of base stations based on the historical vital signs data and historical location data. The data acquisition and uploading module is used to collect the vital signs and location data of goats in real time based on the wearable smart device, and perform lightweight pre-analysis to obtain the pre-analysis results; according to the pre-analysis results, the unuploaded vital signs and location data are sorted and uploaded to obtain a data transmission sorting table; The preliminary health assessment module is used to receive and perform in-depth analysis of the vital signs and location data uploaded by the base station in real time to determine the daily homeostatic health status of the goats. The stress test and treatment module is used to apply stress tests to goats at a target time using target stress test items, obtain the goats' vital signs and location data within a time interval centered on the target time and with a preset duration as the step size, determine the stress response health status and the status of the smart device, and generate a graded treatment strategy based on the stress response health status and the daily steady-state health status.

[0016] The technical effects and advantages of the present invention, which is a method and system for real-time monitoring of vital signs data in goat farming based on the Internet of Things, are as follows: This invention achieves adaptive optimization of base station location and quantity. By acquiring historical vital sign data and location data of goats within a target historical period, this invention calculates the historical service center of gravity for each base station and determines whether to update the base station location based on the distance between the current physical location and the historical service center of gravity. It gradually moves the base station closer to the actual activity center of the goats by moving it half a distance at a time. Simultaneously, it identifies historical goat herd aggregation areas through grid-based statistical analysis, adds base stations to coverage hotspots, and identifies and removes redundant base stations based on the total number of data packets. This mechanism allows base station deployment to dynamically adapt to changes in goat activity patterns, effectively eliminating signal coverage blind spots, reducing data packet loss rates, and avoiding over-deployment of base stations, thus saving hardware costs and energy consumption.

[0017] This invention enables lightweight pre-analysis and prioritized uploading of data on smart devices. On the smart device side, it performs sliding window statistics and behavior recognition on vital sign and location data, classifying the pre-analysis results into three levels: normal, attentive, and abnormal. A data transmission sorting table is then generated based on the priority of abnormal (highest), attentive (second highest), and normal (lowest). The smart device uploads data sequentially according to this sorting table, ensuring that abnormal data is transmitted to the base station with minimal latency. Simultaneously, normal data is uploaded in batches with compression, effectively reducing device power consumption and communication bandwidth usage.

[0018] This invention enables in-depth health status assessment based on multimodal temporal features. It performs in-depth analysis on cleaned vital sign data, extracting temporal features such as heart rate variability, body temperature change rate, rumination interval, and activity intensity integral. A multimodal temporal feature matrix is ​​constructed and input into a pre-trained long short-term memory network or random forest model, outputting daily steady-state health status, including categories such as healthy, sub-healthy, suspected illness, and acute illness. This allows for a refined assessment of goat health status.

[0019] This invention introduces a periodic stress test and linkage verification mechanism. At preset intervals, the invention selects target stress test items from a stress test library and applies the stress test to goats at the target time. It acquires vital sign and location data during the pre-stress baseline period, stress response period, and stress recovery period to determine the stress response health status, including four states: healthy stress response, decreased stress recovery ability, sensory or neurological dysfunction, and oversensitivity to stress. Based on this, a linkage verification matrix is ​​established between daily homeostatic health status and stress response health status, and a graded treatment strategy is generated through consistency discrimination. This mechanism can effectively distinguish between real diseases and equipment malfunctions, reduce false alarm rates, and detect hidden health risks early, achieving adaptive closed-loop optimization of health status assessment.

[0020] This invention provides a method for determining the target time of a goat crossing a base station signal boundary. The invention utilizes a smart device to collect real-time received signal strength indicators (RSIs) between the goat and the base station. When the signal strength continuously decreases from above a preset high threshold to below a preset low threshold, and the duration of this decrease exceeds a preset duration threshold, it is determined that the goat is crossing the base station coverage boundary, and the moment of boundary crossing is taken as the target time. This method reuses existing base station deployment data, requires no additional sensors, achieves individualized adaptive test scheduling, and the determination of the stress test timing has high repeatability and biological rationality.

[0021] This invention enables synchronous diagnosis of the status of intelligent devices. During stress testing, by analyzing the change in the peak stress response compared to the baseline value, and combining this with the daily steady-state health status, the invention determines whether the sensors of the intelligent device are functioning properly. Simultaneously, by examining the time series of positioning data, it determines whether there are any anomalies in the positioning module. This mechanism can promptly detect loose equipment fixation, sensor failure, or positioning module malfunction, prompting livestock managers to perform equipment maintenance and avoiding misjudgments of health status due to equipment problems.

[0022] In summary, this invention significantly improves the real-time performance, accuracy, and reliability of goat farming vital sign data monitoring through the coordinated efforts of multiple stages, including adaptive base station deployment, priority data uploading, multimodal deep analysis, stress test linkage verification, and equipment status diagnosis. Attached Figure Description

[0023] Figure 1 The flowchart illustrates the method for real-time monitoring of vital signs data in goat farming based on the Internet of Things provided by this invention.

[0024] Figure 2 The structural block diagram of the Internet of Things-based real-time monitoring system for goat breeding vital signs provided by the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 This invention provides a method for real-time monitoring of vital signs data in goat farming based on the Internet of Things (IoT). The method includes: S1. Acquire the vital signs and location data of the goat within the target historical period, denoted as historical vital signs data and historical location data, respectively. Update the base station location and number of base stations based on the historical vital signs and location data to obtain the current base station location and current number of base stations. The base station is used to receive the vital signs and location data of the goat collected by a smart device, such as a smart collar, worn on the goat. The base station is set within the goat's activity area. The smart device and the base station transmit data via Bluetooth. When the distance between the smart device and the base station is within a preset transmission distance range, a successful match is achieved, and data transmission can begin. The target historical period can be the previous week or the previous month. Vital signs data include the goat's heart rate, body temperature, and rumination frequency. Location data includes the goat's latitude and longitude coordinates within its activity area, the timestamp of the collection time, and the duration of stay at that location. The duration of stay is calculated using the time difference between consecutive location points. A base station refers to a fixed or mobile signal receiving device set within the goat's activity area, used to receive the vital signs and location data transmitted and uploaded via Bluetooth by the smart device worn on the goat. The base station possesses data storage, forwarding, and edge computing capabilities, and can aggregate data to a central server via wired or wireless networks. The preset transmission distance range refers to the maximum straight-line distance between the smart device and the base station that allows for a successful Bluetooth connection and data transmission; a typical value for this distance is fifty meters. When the actual distance between the smart device and the base station is less than or equal to fifty meters, successful pairing allows for data transmission; when the actual distance exceeds fifty meters, the signal strength is insufficient to maintain reliable communication, and transmission may be interrupted or fail.

[0027] S1.1 For each deployed base station, extract the location data (GPS coordinates) of the goats received by the base station during the target historical period, and record it as historical location data; for each base station, calculate the average coordinates of all location points of the historical vital signs data, that is, sum all longitude values ​​and divide by the total number of historical location data to obtain the average longitude, and similarly calculate the average latitude. The average coordinates are the historical service center of gravity of the base station. The historical service center of gravity represents the location where the goats most frequently appeared and successfully completed data upload within the coverage area of ​​the base station during the target historical period.

[0028] S1.2 At a preset fixed time (e.g., when sheep enter the pen to rest on Sunday night), for each base station, obtain the distance between its current physical location and the historical service center of gravity. Determine whether to update the base station location based on the distance. If the distance is greater than a preset movement trigger threshold (e.g., 10 meters), determine to update the base station location and generate a movement command, instructing the base station to move half the distance in the direction of the historical service center of gravity (e.g., if the distance is 20 meters, move 10 meters). After the movement is completed, re-evaluate the distance between the new location of the base station and the historical service center of gravity. If it is still greater than the threshold, move half again until it enters the movement trigger threshold to obtain the current base station location.

[0029] S1.3 Identify historical sheep flock gathering areas based on the historical positioning data; determine whether to add a base station based on the location of the historical sheep flock gathering areas and the already deployed base stations; specifically including: obtaining a plan view of the farm's activity area, dividing the farm's activity area into regular (square) grids with a side length of L meters (L can be 10); using historical positioning data to count the total dwell time of goats in each grid, and calculating the average dwell time of all grids; if the total dwell time of a certain grid exceeds twice the average dwell time, the grid is determined to be a historical sheep flock gathering area, and the distance from the center of the grid to any already deployed base station is greater than the Bluetooth reliable communication radius (e.g., 50 meters), then the grid is determined to be a coverage missing hotspot, and a new base station is added at the center of the grid to eliminate the coverage blind spot and ensure that the vital signs data of highly concentrated areas can be completely received.

[0030] S1.4. Count the total number of data packets received by each base station during the target historical period. Based on the total number of data packets, determine the utilization rate of each base station. Based on the utilization rate, determine whether there are redundant base stations. If redundant base stations exist, delete them. Specifically, this includes: counting the total number of data packets received by each base station during the target historical period, and calculating the average number of data packets received by all base stations; if the total number of data packets received by a base station is less than 30% of the average number of data packets received (i.e., it is idle or inefficient for a long time), and the distance between the base station and its nearest neighbor base station is less than 30 meters (there is coverage overlap and redundancy), then the base station is determined to be a redundant base station. After issuing a warning notification and confirming that there are no abnormalities, the system removes the redundant base station, and its original coverage task is taken over by the neighboring base station, thereby maintaining the economy and effectiveness of the number of base stations. Based on the above steps, the current base station location and the current number of base stations can be obtained. A data packet refers to a complete data unit that a smart device encapsulates and sends to a base station according to a preset data format during a Bluetooth transmission session. Each data packet contains one or more goat vital sign data records, location data records, a timestamp of the data collection time, a goat individual identifier, and a cyclic redundancy check code used to verify data integrity.

[0031] S2. Real-time collection of vital signs and location data of goats using wearable smart devices, and lightweight pre-analysis to obtain pre-analysis results; sorting and uploading the unuploaded vital signs and location data according to the pre-analysis results to obtain a data transmission sorting table.

[0032] S2.1 Based on the smart device worn on the goat, the goat's vital signs data and location data are collected in real time at a first sampling frequency. The smart device has built-in acceleration sensor, temperature sensor, heart rate sensor and low power Bluetooth module, etc. The first sampling frequency refers to the time density of the original collection of the goat's vital signs data and location data by the built-in sensors of the smart device. The specific value can be once per second, which can be adjusted according to the actual situation.

[0033] S2.2 On the smart device, perform lightweight pre-analysis on the vital signs data and location data respectively to obtain pre-analysis results. The pre-analysis results include three levels: normal, attention, and abnormal, and corresponding feature identifiers. The specific steps for performing lightweight pre-analysis to obtain pre-analysis results are as follows: A sliding window statistical analysis is performed on the vital signs data. Using consecutive vital signs data points from the past 30 seconds as a time window, the average heart rate and body temperature within that window are calculated and recorded as the current data. The current data is compared with preset threshold values. If the current data exceeds the upper or lower limit of the threshold by less than 10%, it is marked as a "pay attention" level; if the exceedance is greater than or equal to 10%, it is marked as an "abnormal" level; otherwise, it is marked as a "normal" level. The preset threshold values ​​refer to a set of pre-defined numerical boundaries used to determine whether the vital signs data of goats are within the normal range. These thresholds are statistically derived from historical vital sign data of healthy goats of the same breed and age under the same environmental conditions, and can be dynamically adjusted according to seasonal changes and the stage of breeding. Specifically, they include heart rate thresholds and body temperature thresholds. The heart rate thresholds include an upper and lower limit of the normal heart rate range. For adult goats, the upper limit of the normal heart rate range is set at 100 beats per minute, and the lower limit is set at 60 beats per minute. When the current average heart rate collected by the intelligent device is higher than 100 beats per minute, it is determined to exceed the upper limit; when it is lower than 60 beats per minute, it is determined to exceed the lower limit. The body temperature thresholds include an upper and lower limit of the normal body temperature range. For adult goats, the upper limit of the normal body temperature range is set at 40.0 degrees Celsius, and the lower limit is set at 38.5 degrees Celsius. When the current average body temperature collected by the intelligent device is higher than 40.0 degrees Celsius, it is determined to exceed the upper limit; when it is lower than 38.5 degrees Celsius, it is determined to exceed the lower limit. Behavioral recognition is performed on the location data. The displacement velocity of the goat is calculated based on the latitude and longitude difference between consecutive location points, and the unit is converted to meters per second. At the same time, the rate of change of rotation angle is calculated based on the directional changes of three adjacent location points. If the displacement velocity is less than 0.1 meters per second for 30 seconds and the rate of change of rotation angle is less than 5 degrees per second for 30 seconds, it is judged as abnormal stationary behavior, and the corresponding data window is marked as attention level; if the displacement velocity exceeds 3 meters per second, it is judged as violent movement behavior, and the corresponding data window is marked as abnormal level; all other cases are marked as normal level. If any data window in the vital signs data and location data is marked as abnormal, the final pre-analysis result for that time window will be abnormal; if neither is marked as abnormal but at least one is marked as attention-level, the final result will be attention-level; the final result will be normal only if both are normal.

[0034] S2.3. Based on the levels in the pre-analysis results, dynamically sort the unuploaded vital sign data and location data, set the data marked as abnormal to the highest upload priority, set the data of the abnormal level to the medium priority, and set the data of the normal level to the low priority; generate a data transmission sorting table according to the priority from high to low. S2.4 The smart device uploads its vital signs data and location data to the matched base station via Bluetooth channel according to the data transmission sorting table. Matching refers to the discovery, connection, and authentication process specified in the Bluetooth protocol between the smart device and the base station, enabling them to establish a data communication link. Specifically, this includes the following steps: The smart device periodically broadcasts its own device identifier and the goat's individual identifier; after receiving the broadcast signal within its signal coverage area, the base station verifies whether the smart device is in the allowed access device whitelist; after successful verification, the base station sends a connection request to the smart device; after the smart device confirms the request, both parties establish a Bluetooth asynchronous connectionless link or a Bluetooth Low Energy connection; after the connection is established, the smart device can upload data to the base station according to the data transmission sorting table.

[0035] S3. Receive and perform in-depth analysis of the vital signs and location data uploaded by the current base station in real time to determine the daily homeostatic health status of the goats.

[0036] S3.1 Receive vital sign data and location data uploaded from different smart devices through the current base station, and align and clean the data according to the timestamp and the individual goat identifier, and remove duplicate, missing or abnormal data records; S3.2. Perform in-depth analysis on the cleaned data to obtain temporal features; construct a multimodal temporal feature matrix based on the temporal features, including heart rate variability, body temperature change rate, rumination interval, and activity intensity integral; input the multimodal temporal feature matrix into a pre-trained health status assessment model, which is constructed based on a long short-term memory network or a random forest algorithm; output the health status classification results of goats based on the health status assessment model, denoted as daily steady-state health status, including four states: healthy, sub-healthy, suspected illness, and acute illness. In-depth analysis of the cleaned data yielded the following time-series characteristics: Heart rate variability was calculated from the heart rate data in the cleaned vital signs data: the time interval sequence between two consecutive heartbeats was extracted, converted to milliseconds, and the standard deviation of this sequence within a 60-second time window was calculated. This standard deviation is the heart rate variability. A decreased heart rate variability usually indicates that the goat is under stress or in a state of disease. The body temperature change rate was calculated from the vital signs data after cleaning: using a 60-second time window, the difference between the body temperature values ​​at the beginning and end of the window was divided by 60 seconds to obtain the body temperature change rate, in degrees Celsius per second. A consistently positive body temperature change rate exceeding 0.05 degrees Celsius per second indicates a fever trend, while a consistently negative rate below -0.03 degrees Celsius per second indicates metabolic abnormalities or hypothermia risk. The rumination interval was calculated from the rumination data in the post-wash vital signs data: the start and end times of each rumination cycle were identified by the periodic chewing vibration pattern detected by the throat vibration sensor in the smart device. The difference between the start times of two adjacent rumination cycles was the rumination interval, in minutes. A significantly prolonged rumination interval suggests digestive system abnormalities or systemic diseases. Activity intensity integrals were calculated from the post-cleaning vital signs data: the raw acceleration values ​​collected by the triaxial accelerometer in the smart device were squared and then squared to obtain the resultant acceleration modulus. This modulus was then integrated within a 60-second time window to obtain the activity intensity integral, also known as the activity intensity value, with units of gravitational acceleration multiplied by seconds. A significantly lower activity intensity integral than the individual's historical baseline suggests depression, while a significantly higher integral suggests restlessness or pain.

[0037] The steps for constructing a multimodal temporal feature matrix based on the aforementioned temporal features include: Determine the time window length: A 60-second time window unit can be set. Within each time window, a set of time-series feature values ​​is calculated, including four features: heart rate variability, body temperature variability, rumination interval, and activity intensity integral. Determine the number of time windows: Take sixty consecutive time windows before the current moment, covering the historical data of the past sixty minutes, as the data source for constructing the matrix. Arrange the heart rate variability, body temperature variability, rumination interval, and activity intensity integral calculated within each time window in a fixed order to form a row vector containing the four feature values. The sixty time windows form a total of sixty row vectors. Stack the sixty row vectors vertically in chronological order from oldest to newest to form a two-dimensional matrix of sixty rows by four columns. This matrix is ​​the multimodal time-series feature matrix. Each row in the matrix corresponds to a time window, and each column corresponds to a type of vital sign feature.

[0038] The health status assessment model specifically refers to a pre-trained supervised learning model used to map the input multimodal temporal feature matrix to the health status classification result of the goat. The model structure is constructed using a Long Short-Term Memory network or a random forest algorithm.

[0039] If a Long Short-Term Memory (LSTM) network is used, the model structure includes an input layer with an input dimension of 60 x 4; an LSM hidden layer containing 128 memory units to learn long-term dependencies in temporal features; a dropout layer with a dropout rate of 0.5% to prevent overfitting; a fully connected layer that maps the output of the LSM layer to the output vectors of four nodes; and a normalized exponential activation layer that transforms the output vectors into probability distributions for the four health state categories. If a Random Forest algorithm is used, the model structure includes 100 decision trees, each with a maximum depth of 20 layers. When a node splits, two features are randomly selected for optimal splitting, and the health state classification is finally determined by the voting results of all decision trees.

[0040] The specific training method for the model includes the following steps: A training dataset was collected, extracting a large number of multimodal temporal feature matrix samples from goats already labeled with healthy status. Each sample corresponds to a 60-second time window, and professional veterinarians labeled the goats with health status tags based on actual clinical examination results. The tags included four categories: healthy, sub-healthy, suspected disease, and acute disease. The training data was preprocessed: each feature was normalized, mapping heart rate variability, body temperature variability, rumination interval, and activity intensity integral to a zero-to-one interval to eliminate the influence of different feature dimensions. A 10-fold cross-validation method was used to divide the training and validation sets. The training data was randomly divided into ten equal parts, with nine parts used as the training set and one as the validation set alternately, repeated ten times. The training set was input into a Long Short-Term Memory (LSTM) network or a random forest model. Standard training strategies for multi-class classification tasks were used to update the model parameters. The model's classification accuracy was calculated using the validation set, and the model with the highest validation accuracy was saved as the final model. The final model was evaluated using an independent test set not used in training. The average classification accuracy of the model across the four health states was calculated, requiring an average accuracy of at least 85% for deployment.

[0041] S3.3. Generate corresponding processing suggestions based on the daily steady-state health status: if it is a healthy state, continue routine monitoring; if it is a sub-healthy state, increase the sampling frequency and mark it as a concern; if it is a suspected disease or an emergency disease state, generate early warning information and push it to the breeding management terminal. The early warning information includes the time, location and vital signs data of the abnormal event.

[0042] S4. Apply stress test to goats at a target time using target stress test items, obtain vital sign data and location data of goats within a time interval centered on the target time and with a preset duration as the step size, determine the stress response health status and smart device status, and generate a graded treatment strategy based on the stress response health status and daily steady-state health status.

[0043] S4.1. Every preset period, a preset number of candidate stress test items are randomly or polled from the stress test library. The candidate stress test items include playing wolf howling audio, simulating eagle howling, and suddenly releasing vibration signals. Each candidate stress test item is associated with a historical stress response feature library.

[0044] The preset period refers to the fixed time interval between two adjacent stress tests. This interval can be 72 hours, meaning that a stress test is performed on each goat or each sampled group every three days. This period length ensures timely capture of dynamic changes in the goats' health status while avoiding excessively frequent stress stimulation that could cause unnecessary physiological burden on the goats.

[0045] A stress test library is a set of stress test items pre-stored in a server or base station. Each stress test item corresponds to a triggerable stressor. Items in the stress test library include playing wolf howling audio, simulating eagle howling, and suddenly releasing vibration signals.

[0046] The preset quantity refers to the number of candidate stress test items selected from the stress test library each time, and can specifically be three. That is, at the beginning of each 72-hour cycle, the system randomly or in a round-robin manner selects three different stimulus items from the stress test library as candidates for subsequent screening processes.

[0047] The historical stress response feature database refers to a dataset created separately for each candidate stress test item, recording the average response characteristics of a goat herd to that item in multiple past stress tests. The historical stress response feature database for each item includes the average and standard deviation of the increase in heart rate, the average and standard deviation of the increase in activity intensity, and the average and standard deviation of the recovery time constant for all historical tests under that item, as well as records of abnormal events triggered by that item, such as stampedes or equipment detachment. This feature database is used for subsequent item screening to eliminate stimuli that may cause excessive stress or safety accidents.

[0048] S4.2. Select the candidate stress test items from a preset perspective to obtain the target stress test items.

[0049] Preset perspectives refer to evaluation criteria that screen and rank candidate stress test items from multiple dimensions, specifically including the following three perspectives: The first perspective is the stress safety index. Based on the historical stress response feature database corresponding to this item, the proportion of historical test heart rate increases exceeding twice the group average increase, the proportion of activity intensity increases exceeding twice the group average increase, and the number of abnormal events triggered under this item are calculated. The stress safety index is obtained by weighting and summing these three indicators, with weights set to 0.3, 0.3, and 0.4 respectively. A higher stress safety index indicates that the item is more likely to trigger excessive stress response or safety accidents, and should be prioritized for exclusion; abnormal events include stampedes, fence collisions, etc. The second perspective is stress response discrimination. Based on the historical stress response feature database corresponding to the item, the difference in heart rate increase between healthy goats and sub-healthy goats under this item is calculated. The larger the difference, the more effectively the item can distinguish goats in different health states, and the higher the discrimination, the more priority is given to retention. The third perspective is item freshness. Record the date when each item was last selected as the target stress test item. The further back in time the item is from the current date, the higher its freshness. Prioritize the selection of items with high freshness to avoid repeated use of the same stimulus, which could lead to habituation in the goats and thus loss of stress response. Calculate the comprehensive scores of the three candidate items from the three perspectives mentioned above, sort them from low to high according to the stress safety index, and remove items whose safety index exceeds the preset safety threshold (e.g., 0.7); sort the remaining items from high to low according to the stress response discrimination, and select the item with the highest freshness as the final target stress test item.

[0050] S4.3. Apply stress test to goats at the target time using the target stress test items; while applying the stress test, mark the test start timestamp by synchronous triggering signal.

[0051] The stress test is administered using different devices depending on the type of stress test item. For sound-based stress tests, such as playing wolf howls and simulated eagle cries, the sound is played through speakers built into base stations deployed around the goats' activity area. The base station receives instructions from a central server and triggers the playback at the target time, with a playback duration of three to five seconds and a volume set to 90 decibels to ensure clear sound for all goats within the base station's coverage area. For vibration-based stress tests, such as the sudden release of vibration signals, vibration motors pre-embedded in the ground or fence of the activity area are used. These motors receive instructions via a wireless controller, with the vibration frequency set to 50 Hz and lasting for three seconds.

[0052] S4.4. Obtain the vital signs and location data of the goats within a time interval centered on the target time and with a preset duration as the step size, and determine the stress response health status based on the vital signs and location data of the goats within the time interval.

[0053] The specific steps for determining the stress response health status of goats based on their vital signs and location data within a time interval are as follows: Extract vital sign and location data within a four-minute time interval centered on the target time, extending two minutes before and after it. Divide this interval into three sub-stages: the pre-stress baseline period (two minutes to thirty seconds before the target time); the stress response period (from the target time to sixty seconds after the target time); and the stress recovery period (sixty seconds to two minutes after the target time). Calculate the average heart rate and average activity intensity during the baseline period before stress, and record them as the baseline heart rate value and baseline activity intensity value; calculate the maximum heart rate and maximum activity intensity during the stress response period, and record them as the stress heart rate peak and stress activity intensity peak; calculate the time taken for the heart rate to drop from the peak value to within 10% of the baseline value during the stress recovery period, and record it as the recovery time constant. If the peak stress heart rate increases by more than 5% compared to the baseline heart rate, and the peak stress activity intensity increases by more than 10% compared to the baseline activity intensity, then the goat is deemed to have a clear physiological response to this stress test, and the degree of engagement is assessed; otherwise, it is determined that the stress response is missing, and the stress response health status is directly output as sensory or neurological dysfunction. The recovery time constant is compared with the normal recovery time constant range of a healthy goat population of the same breed and age. In this embodiment, the lower limit of the normal recovery time constant range is set at 30 seconds and the upper limit at 90 seconds, which can be adjusted according to the actual situation. If the recovery time constant is greater than 90 seconds, it is determined that the stress recovery ability is reduced; if the recovery time constant is between 30 and 90 seconds, it is determined that the stress recovery ability is normal; if the recovery time constant is less than 30 seconds, it suggests that there may be excessive excitement or excessive activation of the sympathetic nervous system. The overall assessment of stress response health status is as follows: If the stress response is absent, the output status is sensory or neurological dysfunction. If the stress response is present and the recovery time constant is normal, the output status is healthy stress response. If the stress response is present but the recovery time constant is greater than 90 seconds, the output status is decreased stress recovery ability. If the stress response is present but the recovery time constant is less than 30 seconds, the output status is oversensitivity to stress.

[0054] S4.5. Establish a linkage verification matrix based on the stress response health status and daily homeostasis health status, perform consistency judgment based on the linkage verification matrix, and generate a graded treatment strategy; the rows of the linkage verification matrix correspond to the daily homeostasis health status, including healthy, sub-healthy, suspected onset, and emergency onset, etc., and the columns of the linkage verification matrix correspond to the stress response health status, including stress response health, decreased stress recovery ability, sensory or neurological dysfunction, and oversensitivity to stress response, etc.

[0055] The steps for consistency determination based on the linkage verification matrix include: The current daily homeostatic health status and stress response health status of the goat are located in the row and column of the linkage verification matrix, respectively, and the judgment rules for that cell are read. The specific rules are as follows: If both the daily homeostatic health status and the stress response health status are healthy, then both are considered consistent and normal. If the daily homeostatic health status is healthy but the stress response health status shows decreased stress recovery ability, then both are considered inconsistent, suggesting a possible hidden health risk. If the daily homeostatic health status is healthy but the stress response health status shows sensory or neurological dysfunction, then both are considered inconsistent, suggesting a possible hidden neurological dysfunction risk. If the daily homeostatic health status is healthy but the stress response health status shows hypersensitivity to stress, then both are considered inconsistent, suggesting that the goat may have an anxiety predisposition or sympathetic overactivation, belonging to an individual at risk of subclinical behavioral abnormalities. If the daily homeostatic health status is healthy... If the goat's condition is sub-healthy and its stress response health status is also stress response health, then the two are considered to be basically consistent, and the sub-healthy state is cross-validated. If the daily homeostatic health status is sub-healthy and the stress response health status is a decline in stress recovery ability, then the two are considered to be consistent and point to a tendency towards weakness in sub-health. If the daily homeostatic health status is sub-healthy and the stress response health status is sensory or neurological dysfunction, then the two are considered inconsistent, suggesting that the daily sub-health may be due to equipment problems rather than a real disease. If the daily homeostatic health status is sub-healthy and the stress response health status is hypersensitive to stress, then the two are considered to be consistent and mutually reinforcing, suggesting that the goat is in a state of high stress susceptibility, and environmental enrichment and stocking density adjustments should be given special attention. If the daily homeostatic health status is suspected of being diseased and the stress response health status is stress response health, then the two are considered inconsistent, and the daily homeostatic health status is primarily suspected as a false positive, requiring investigation of equipment or transmission malfunctions; if the daily homeostatic health status is suspected of being diseased and the stress response health status is a decline in stress recovery ability, then the two are considered consistent, confirming the existence of a disease state; if the daily homeostatic health status is suspected of being diseased and the stress response health status is sensory or neurological dysfunction, then the two are considered consistent, confirming the existence of a disease state that may involve the nervous system; if the daily homeostatic health status is suspected of being diseased and the stress response health status is hypersensitive to stress, then the two are considered consistent, suggesting that the disease state may be accompanied by autonomic nervous system dysfunction; if the daily homeostatic health status is... If the health status is an emergency onset and the stress response health status is also a stress response health status, then the two are considered severely inconsistent. System malfunctions should be prioritized and on-site manual verification should be arranged. If the daily homeostatic health status is an emergency onset and the stress response health status is a decreased stress recovery ability, then the two are considered consistent, confirming an emergency onset. If the daily homeostatic health status is an emergency onset and the stress response health status is a sensory or neurological dysfunction, then the two are considered consistent, confirming an emergency onset that may involve the nervous system. If the daily homeostatic health status is an emergency onset and the stress response health status is an oversensitivity to stress, then the two are considered consistent, confirming an emergency onset accompanied by severe autonomic nervous system dysfunction, and the treatment priority should be raised by one level.

[0056] A tiered handling strategy is generated based on the consistency judgment results, which is divided into four levels: The Level 1 intervention strategy applies to situations where both conditions are consistent, both are normal or sub-healthy, and are mutually matched. Specifically, this includes situations where the daily homeostatic health status is healthy and the stress response health status is stress response health, and situations where the daily homeostatic health status is sub-healthy and the stress response health status is stress response health. Specific measures include: maintaining the routine monitoring plan, extending the stress testing cycle from 72 hours to 120 hours, and keeping the daily sampling frequency unchanged.

[0057] The secondary treatment strategy applies to situations where the two conditions are inconsistent but do not indicate an emergency illness. Specifically, this includes situations where the goat's daily homeostatic health status is healthy but its stress response health status shows decreased stress recovery ability; situations where the goat's daily homeostatic health status is healthy but its stress response health status shows sensory or neurological dysfunction; situations where the goat's daily homeostatic health status is healthy but its stress response health status shows hypersensitivity to stress; and situations where the goat's daily homeostatic health status is sub-healthy but its stress response health status shows sensory or neurological dysfunction. The specific measures are: increasing the sampling frequency of daily vital signs data from once per second to twice per second, maintaining the same stress testing cycle, and simultaneously generating an artificially assisted observation task on the livestock management terminal, prompting the farmer to conduct behavioral observations of the goat during the next natural observation period.

[0058] The three-tiered management strategy applies to situations where both conditions are consistent and indicate a medium-to-high risk level. Specifically, this includes situations where the goat's daily homeostatic health status is sub-healthy and its stress response health status is characterized by decreased stress recovery ability; where the daily homeostatic health status is sub-healthy and its stress response health status is characterized by hypersensitivity to stress; where the daily homeostatic health status is suspected of being diseased and its stress response health status is characterized by decreased stress recovery ability; where the daily homeostatic health status is suspected of being diseased and its stress response health status is characterized by sensory or neurological dysfunction; and where the daily homeostatic health status is suspected of being diseased and its stress response health status is characterized by hypersensitivity to stress. Specific measures include: immediately generating an early warning message and pushing it to the breeding management terminal and the veterinarian's mobile terminal; marking the goat as a key monitoring target; and conducting additional stress tests every six hours for the next 24 hours to track changes in its condition.

[0059] The Level 4 response strategy applies to situations where the two conditions contradict each other and the goat's normal homeostatic health status is an emergency illness, or situations where the two conditions are consistent and the goat's normal health status is an emergency illness. Specifically, this includes situations where the goat's normal homeostatic health status is an emergency illness and its stress response health status is also stress response health (contradictory); situations where the goat's normal homeostatic health status is an emergency illness and its stress response health status is a decreased stress recovery ability (consistent); situations where the goat's normal homeostatic health status is an emergency illness and its stress response health status is sensory or neurological dysfunction (consistent); and situations where the goat's normal homeostatic health status is an emergency illness and its stress response health status is hypersensitive to stress (consistent). The specific measures are as follows: First, dispatch inspectors to verify the goat's actual physical condition and the wearing of the smart device on-site; if it is confirmed on-site that the goat is indeed in an emergency illness state, immediately initiate the veterinary intervention process; if it is confirmed on-site that the goat's physical condition is normal, determine it as a system malfunction, trigger the self-check process of the base station and smart device, and suspend the goat's automatic early warning function until repair is completed.

[0060] In this embodiment, S4 further includes: Extract vital sign data within a preset length window (e.g., 30 seconds before the start of the test) before the stress test is applied, calculate the average heart rate and average activity intensity of the goats within this window, and record them as the baseline values; extract vital sign data within a preset length window (e.g., 30 seconds after the start of the test) after the stress test is applied, calculate the maximum heart rate and maximum activity intensity within this window, and record them as the stress response peak. If the change in the peak stress response relative to the baseline value simultaneously meets the preset conditions: the increase in heart rate exceeds the preset minimum sensitivity threshold (e.g., 5%), and the increase in activity intensity exceeds the preset minimum sensitivity threshold (e.g., 10%), then it is determined that the vital signs sensors (heart rate sensor, accelerometer) of the smart device can respond normally to external stimuli, and the device status is initially judged to be normal. If the change in the stress response peak does not meet the preset conditions, the current daily homeostatic health status of the goat is queried. If the daily homeostatic health status is "healthy" or "sub-healthy", the smart device is determined to have a sensor failure or a severe decrease in sensitivity, and the smart device status is faulty. If the daily homeostatic health status is "suspected illness" or "emergency illness", the goat itself may be experiencing stress sluggishness due to illness. In this case, the smart device is not determined to be faulty, but its status is determined to be suspicious.

[0061] The system checks the time series of location data within a time interval. If the continuous loss of location data exceeds the preset loss duration (which can be 10 seconds, set according to requirements), or if the calculated movement speed between adjacent location points exceeds the physiological limit speed of a goat (e.g., 15 meters / second), it determines that the positioning module or GPS antenna of the smart device is abnormal and directly outputs the smart device status as faulty.

[0062] In this embodiment, the method for determining the target time includes: Obtain the location of each current base station and its corresponding base station coverage radius; draw a circular coverage boundary with the current base station location as the center and the base station coverage radius as the radius; for the overlapping area of ​​adjacent current base stations, use the perpendicular bisector of the line connecting the two current base stations as the coverage boundary dividing line to form the theoretical coverage polygon of each base station; The smart device worn on the goat collects the Received Signal Strength Indicator (RSSI) of the current base station with which it has established a Bluetooth connection in real time at a second sampling frequency, and caches the RSSI sequence within the last 10 seconds. When the RSSI value is detected to continuously drop from above the preset high threshold (-60dBm) to below the preset low threshold (-75dBm) and the duration of the drop reaches or exceeds the preset duration threshold (5 seconds), it is determined that the goat is currently moving from the strong coverage area of ​​the base station towards the coverage boundary and crossing the boundary, and the moment of crossing the boundary is determined. The strong coverage area of ​​the base station refers to the spatial area near the base station within the Bluetooth reliable communication radius centered on the base station location. The calculation method for the boundary crossing time can be as follows: set a monitoring threshold, for example, a preset high threshold of -60dBm and a preset low threshold of -75dBm; record the moment t1 when the RSSI value is last higher than the high threshold and the moment t2 when it is first lower than the low threshold; take the arithmetic mean of t1 and t2 as the boundary crossing time t0, t0=(t1+t2) / 2, and require t2-t1≤5 seconds, otherwise it is determined as an invalid crossing; the second sampling frequency refers to the time density of the smart device's collection of the received signal strength indication value, which can be taken as twice per second, that is, the received signal strength indication value of the currently connected base station is collected once every 0.5 seconds. This frequency is higher than the first sampling frequency. The moment of crossing the boundary is recorded as the candidate target moment. At the same time, the crossing direction is determined based on the base station connection status before and after the RSSI value decreases: if the base station is connected before crossing and is still connected after crossing and the RSSI value is continuously lower than the preset low threshold, it is determined to be the direction of leaving the coverage area; if the base station is connected before crossing and is switched to base station B after crossing, it is determined to be the cross-base station transfer direction. Prioritize the crossing time of the boundary in the direction of leaving the coverage area as the final target time; if there is no crossing event in the direction of leaving the coverage area in the current preset period, select the crossing time of the boundary in the direction of cross-base station transfer; if neither exists, keep the target time of the previous period or use the default time (e.g., push it 90 minutes after sunrise).

[0063] Please see Figure 2 Another objective of this invention is to provide a real-time monitoring system for vital sign data in goat farming based on the Internet of Things, the system comprising: The base station information update module is used to acquire the vital signs data and location data of the goat during the target's historical period, which are respectively recorded as historical vital signs data and historical location data; and to update the base station location and number of base stations based on the historical vital signs data and historical location data. The data acquisition and uploading module is used to collect the vital signs and location data of goats in real time based on the wearable smart device, and perform lightweight pre-analysis to obtain the pre-analysis results; according to the pre-analysis results, the unuploaded vital signs and location data are sorted and uploaded to obtain a data transmission sorting table; The preliminary health assessment module is used to receive and perform in-depth analysis of the vital signs and location data uploaded by the base station in real time to determine the daily homeostatic health status of the goats. The stress test and treatment module is used to apply stress tests to goats at a target time using target stress test items, obtain the goats' vital signs and location data within a time interval centered on the target time and with a preset duration as the step size, determine the stress response health status and the status of the smart device, and generate a graded treatment strategy based on the stress response health status and the daily steady-state health status.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0065] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations using a large amount of collected data, and are the closest to the real situation. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0066] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for real-time monitoring of goat farming vital data based on the Internet of Things, characterized by, The method includes: S1. Obtain the vital signs data and location data of the goat during the target historical period, and record them as historical vital signs data and historical location data, respectively; update the base station location and number of base stations based on the historical vital signs data and historical location data to obtain the current base station location and current number of base stations; S2. Real-time collection of vital signs and location data of goats based on wearable smart devices, and lightweight pre-analysis to obtain pre-analysis results; sorting and uploading the unuploaded vital signs and location data according to the pre-analysis results to obtain a data transmission sorting table; S3. Receive and perform in-depth analysis of the vital signs and location data uploaded by the current base station in real time to determine the daily homeostatic health status of the goats. S4. Apply stress test to goats at a target time using target stress test items, obtain vital sign data and location data of goats within a time interval centered on the target time and with a preset duration as the step size, determine stress response health status and smart device status, and generate a graded treatment strategy based on the stress response health status and daily steady-state health status. S1 includes: S1.1 For each deployed base station, extract the location data of the goats received by the base station during the target historical period and record it as historical location data; for each base station, calculate the average coordinates of all location points of the historical vital signs data, and the average coordinates are the historical service centroid of the base station; S1.

2. At a preset fixed time, for each base station, obtain the distance between its current physical location and the historical service center, and determine whether to update the base station location based on the distance; S1.3 Identify historical sheep flock gathering areas based on the historical positioning data; determine whether to add base stations based on the location of historical sheep flock gathering areas and the locations of existing base stations. S1.4 Count the total number of data packets received by each base station in the target historical period, determine the utilization rate of each base station based on the total number of data packets, determine whether there are redundant base stations based on the utilization rate, and delete redundant base stations if there are redundant base stations. S4 includes: S4.

1. Every preset period, select a preset number of candidate stress test items from the stress test library; each candidate stress test item is associated with a historical stress response feature library; S4.

2. Select the candidate stress test items from a preset perspective to obtain the target stress test items; S4.

3. Apply stress test to goats at the target time using the target stress test items; while applying the stress test, mark the test start timestamp by synchronously triggering a signal; S4.

4. Obtain the vital signs and location data of the goats within a time interval centered on the target time and with a preset duration as the step size, and determine the stress response health status based on the vital signs and location data of the goats within the time interval. S4.

5. Establish a linkage verification matrix based on the stress response health status and the daily steady-state health status, perform consistency judgment based on the linkage verification matrix, and generate a graded treatment strategy; the rows of the linkage verification matrix correspond to the daily steady-state health status, and the columns correspond to the stress response health status. The method for determining the target time includes: Obtain the location of each current base station and its corresponding base station coverage radius; draw a circular coverage boundary with the current base station location as the center and the base station coverage radius as the radius. The received signal strength indicator (RSSI) of the current base station with which it has established a Bluetooth connection is collected in real time by a smart device at a second sampling frequency; when the RSSI value is detected to continuously drop from above a preset high threshold to below a preset low threshold and the duration of the drop reaches or exceeds a preset duration threshold, it is determined that the goat is currently moving from the strong coverage area of ​​the base station towards the coverage boundary and crossing the boundary, and the moment of crossing the boundary is determined. The moment of crossing the boundary is recorded as the candidate target moment. At the same time, the crossing direction is determined based on the base station connection status before and after the RSSI value decreases: if the base station is connected before crossing and is still connected after crossing and the RSSI value is continuously lower than the preset low threshold, it is determined to be the direction of leaving the coverage area; if the base station is connected before crossing and is switched to base station B after crossing, it is determined to be the cross-base station transfer direction. Prioritize the boundary crossing time in the direction of leaving the coverage area as the final target time; if there is no boundary crossing event in the direction of leaving the coverage area within the current preset period, then select the boundary crossing time in the cross-base station transfer direction.

2. The real-time monitoring method of goat farming vital data based on Internet of Things according to claim 1, characterized in that, The process of determining whether to update the base station location based on the distance includes: if the distance is greater than a preset movement trigger threshold, determining to update the base station location and generating a movement command to instruct the base station to move half the distance in the direction of the historical service center of gravity; after the movement is completed, re-evaluating the distance between the new location of the base station and the historical service center of gravity, and if it is still greater than the threshold, moving half again until it enters the movement trigger threshold to obtain the current base station location.

3. The real-time monitoring method of goat farming vital data based on Internet of Things as claimed in claim 1, wherein, The historical sheep gathering areas are identified based on the historical location data; Determining whether to add a base station based on historical sheep gathering areas and the location of existing base stations includes: Obtain a floor plan of the farm's activity area and divide the farm's activity area into a regular grid with a side length of L meters; Using historical location data, the total dwell time of goats in each grid was counted, and the average dwell time of all grids was calculated. If the total dwell time of a certain grid exceeds twice the average dwell time, the grid is determined to be a historical sheep gathering area. If the distance from the center of the grid to any deployed base station is greater than the Bluetooth reliable communication radius, the grid is determined to be a coverage gap hotspot, and a new base station is added at the center of the grid.

4. The real-time monitoring method of goat farming vital data based on Internet of Things as claimed in claim 1, wherein, S2 includes: S2.1 Based on the smart device worn on the goat, the goat's vital signs and location data are collected in real time at the first sampling frequency; S2.2 On the smart device, perform lightweight pre-analysis on the vital signs data and location data respectively to obtain pre-analysis results. The pre-analysis results include three levels: normal, attention, and abnormal, and corresponding feature identifiers. S2.

3. Based on the levels in the pre-analysis results, dynamically sort the unuploaded vital sign data and location data to generate a data transmission sorting table; S2.4 The smart device uploads vital sign data and location data to the matched base station via Bluetooth channel in sequence according to the data transmission sorting table.

5. The method for real-time monitoring of vital signs data in goat farming based on the Internet of Things according to claim 1, characterized in that, S3 includes: S3.1 Receive vital sign data and location data uploaded from different smart devices through the current base station, and align and clean the data according to the timestamp and the individual goat identifier; S3.

2. Perform in-depth analysis on the cleaned data to obtain temporal features; construct a multimodal temporal feature matrix based on the temporal features; input the multimodal temporal feature matrix into a pre-trained health status assessment model; output the health status classification results of goats based on the health status assessment model, denoted as daily steady-state health status, including healthy, sub-healthy, suspected onset of disease, and acute onset of disease.

6. The method for real-time monitoring of vital signs data in goat farming based on the Internet of Things according to claim 1, characterized in that, S4 further includes: Extract vital sign data within a preset length window before the stress test is applied, calculate the average heart rate and average activity intensity of the goats within this window, and record them as baseline values; extract vital sign data within a preset length window after the stress test is applied, calculate the maximum heart rate and maximum activity intensity within this window, and record them as peak stress response. If the change in the peak stress response relative to the baseline value simultaneously meets the preset conditions: the increase in heart rate exceeds the preset minimum sensitivity threshold, and the increase in activity intensity exceeds the preset minimum sensitivity threshold, then the smart device is preliminarily judged to be normal. If the change in the stress response peak does not reach the preset condition, the current daily homeostatic health status of the goat is queried; if the daily homeostatic health status is "healthy" or "sub-healthy", the smart device is determined to be faulty; if the daily homeostatic health status is "suspected illness" or "emergency illness", the smart device is determined to be suspicious. If the location data is continuously lost for a period of time exceeding the preset loss duration, or if the calculated movement speed between adjacent location points exceeds the physiological limit speed of a goat, the smart device status will be directly output as faulty.

7. A real-time monitoring system for vital signs data in goat farming based on the Internet of Things (IoT), applied to the real-time monitoring method for vital signs data in goat farming based on the IoT as described in any one of claims 1-6, characterized in that, The IoT-based real-time monitoring system for goat farming vital signs includes: The base station information update module is used to acquire the vital signs data and location data of the goat during the target's historical period, which are respectively recorded as historical vital signs data and historical location data; and to update the base station location and number of base stations based on the historical vital signs data and historical location data. The data acquisition and uploading module is used to collect the vital signs and location data of goats in real time based on the wearable smart device, and perform lightweight pre-analysis to obtain the pre-analysis results; according to the pre-analysis results, the unuploaded vital signs and location data are sorted and uploaded to obtain a data transmission sorting table; The preliminary health assessment module is used to receive and perform in-depth analysis of the vital signs and location data uploaded by the base station in real time to determine the daily homeostatic health status of the goats. The stress test and treatment module is used to apply stress tests to goats at a target time using target stress test items, obtain the goats' vital signs and location data within a time interval centered on the target time and with a preset duration as the step size, determine the stress response health status and the status of the smart device, and generate a graded treatment strategy based on the stress response health status and the daily steady-state health status.

Images