Pet identity recognition and positioning system and method based on RFID and GPS fusion

Abstract

The application discloses an RFID+GPS fusion pet identity recognition positioning system and method, and particularly relates to the technical field of pet monitoring, and comprises the following steps: including establishing a motion intensity reference model and constructing a standard intensity grade matrix; respectively acquiring the response characteristics of a radio frequency identification module and a global positioning system module to a motion state, generating sensitivity scores and response accuracy scores; under a unified time axis, fusing the two scores to generate applicability scores; based on the score results, dynamically regulating and controlling the sampling frequencies of the two modules, realizing the synchronous acquisition and accurate matching of pet identity information and position information; through the construction of a fusion score mechanism and a dynamic sampling frequency control strategy, the application realizes the coordinated recognition of the radio frequency identification module and the global positioning system module under different motion states, improves the synchronism of identity information and position information, the adaptability of module response and the stability of the system in intense motion, and enhances the continuous tracking and accurate positioning capabilities.

Description

Technical Field

[0001] This invention relates to the field of pet monitoring technology, and more specifically, to a pet identification and positioning system and method integrating RFID and GPS. Background Technology

[0002] In recent years, with the advancement of urbanization and the expansion of the pet-owning population, problems such as lost pets, misidentification, and illegal transfer have gradually become prominent. Pet identification management and spatial behavior monitoring have become important research directions in smart cities and intelligent pet ownership scenarios. Radio Frequency Identification (RFID) technology, with its low power consumption, low cost, and ability to uniquely identify objects, has been widely used for pet identification; while Global Positioning System (GPS) technology, with its advantages of achieving wide-area, real-time, and continuous positioning, plays an irreplaceable role in pet tracking.

[0003] However, existing solutions that simply combine RFID and GPS exhibit significant shortcomings in dynamic and complex scenarios. First, the RFID module is prone to signal instability in high-speed movement or obstructed environments. Second, the GPS module experiences unpredictable fluctuations in its positioning refresh rate and error under increased motion or environmental disturbances. Furthermore, the two modules have inherent differences in sensing rhythm, sampling sequence, and data update mechanisms, making it difficult to synchronize identity and location information during high-speed movement. This is particularly problematic in areas with high concentrations of pets, easily leading to data mismatches, identification errors, and path confusion. Therefore, this paper proposes an RFID+GPS integrated pet identification and positioning system and method to address these issues. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] The RFID+GPS integrated pet identification and location method includes the following steps:

[0006] Establish an exercise intensity reference model, extract acceleration features corresponding to typical exercise behaviors based on historical movement trajectory data of different types of pets, and construct a standard intensity level matrix for quantifying exercise status;

[0007] The response characteristics of the RFID module to motion state are obtained, the frequency of signal response intensity change of the RFID module during the recognition process is collected, and the sensitivity score of the RFID module corresponding to different motion intensities is generated by combining the stability and reading rate of signal reading.

[0008] The response characteristics of the GPS module to motion states are obtained. The positioning refresh frequency, displacement distance change value and positioning error value of the GPS module under various motion states are collected to generate the response accuracy score of the GPS module corresponding to different motion intensities.

[0009] Based on a unified timeline, the recognition capabilities of the two modules are compared. The sensitivity score of the RFID module and the response accuracy score of the GPS module are fused to obtain the applicability score of the two modules for the current motion intensity.

[0010] A dynamic sampling frequency control mechanism is constructed based on the applicability score results. The signal sampling frequencies of the RFID module and the GPS module are set according to the current exercise intensity level, and the signal sampling frequencies of the RFID module and the GPS module are adjusted in real time to complete the synchronous collection and matching of pet identity information and location information.

[0011] In a preferred embodiment, the step of constructing a standard intensity level matrix for quantifying motion states includes:

[0012] Based on behavioral data samples of pets of different breeds, sizes and age groups, time-series trajectory data of their activities in the natural environment were collected, and continuous acceleration signals were collected using an inertial measurement unit.

[0013] The collected acceleration data is normalized, and behavioral feature indicators containing multiple preset motion behavior dimensions are extracted. The corresponding typical motion behaviors are then identified based on the feature change patterns.

[0014] Based on the behavioral classification results, corresponding exercise intensity grading rules are constructed, and the grade boundaries are set according to the combination pattern of acceleration change amplitude, movement frequency and duration to form a standard intensity grade matrix.

[0015] In a preferred embodiment, the step of obtaining the response characteristics of the radio frequency identification module to motion states includes:

[0016] Under different levels of exercise intensity, the continuous reading data of the RFID module on the pet's identification tag was recorded to obtain a continuous time series of signal strength changes;

[0017] The number of intervals, average success rate, and maximum loss interval duration during signal reading are obtained to evaluate the stability and persistence of the signal under various levels of motion, forming a corresponding stability index matrix.

[0018] By combining the reading rate and stability index matrix, the recognition performance of the RFID module under different motion intensity levels is quantitatively scored, a set of sensitivity score datasets for module applicability evaluation is generated and input into the module applicability scoring model, and the sensitivity scores of the RFID module corresponding to different motion intensities are output.

[0019] In a preferred embodiment, during signal acquisition, the continuous signal intensity change time series is marked as follows:

[0020] The time points when the signal strength is higher than the set threshold are marked as valid read points, and the time points when the signal strength is lower than the set threshold or there is no signal are marked as invalid read points. The number of times the valid read interval is interrupted is counted as the number of intervals. The ratio of the number of valid read points to the total number of read points per unit time is calculated to obtain the average read success rate. The longest duration of a continuous absence of valid read points is taken as the maximum loss interval duration.

[0021] In a preferred embodiment, during the process of quantifying the recognition performance of the RFID module under different motion intensity levels, an applicability scoring model is constructed using a sequence structure modeling approach:

[0022] The original read rate value, intermittent count value, average read success rate value and maximum loss interval duration value under each motion level are encoded into numerical vectors of the same length in sequence.

[0023] The input vectors corresponding to each set of motion levels are projected onto the two-dimensional numerical plane in a fixed order to form multiple sets of state paths.

[0024] Curve fitting is performed on each state path, the discrete gradient value sequence of the curve is calculated and its variance is statistically analyzed, which serves as the response fluctuation index of the RFID module under each motion intensity level.

[0025] The response volatility index is input into the set logarithmic mapping function, and the corresponding sensitivity score is calculated with the response volatility index as the independent variable. The logarithmic mapping function is defined as S(x)=ln(1+x), where x is the response volatility index and S(x) is the sensitivity score.

[0026] In a preferred embodiment, the step of obtaining the response characteristics of the global positioning system module to motion states includes:

[0027] Under the set multi-level motion intensity, the location information data continuously output by the global positioning system module is collected, and the time intervals in the collected data are statistically analyzed to calculate the positioning refresh frequency under each level and construct a frequency time series.

[0028] Based on the coordinate changes of adjacent points in the frequency time series, the displacement distance change value per unit time is calculated point by point, and combined with the total path length and duration, the average displacement change rate and trajectory fluctuation index under each motion level are extracted.

[0029] For each set of data, single-point error estimation is performed on the positioning error. The error variance and maximum error value under the same motion level are calculated using a sliding window method to generate an error distribution parameter set for the current motion intensity.

[0030] A response accuracy evaluation function is constructed, taking the refresh frequency, displacement change rate, error variance, and maximum error corresponding to each motion level as multivariate inputs, and calculating the response accuracy score of the Global Positioning System module under different motion intensity levels through a multi-parameter nonlinear fitting method.

[0031] In a preferred embodiment, the multi-parameter nonlinear fitting method refers to:

[0032] The refresh frequency, displacement change rate, error variance, and maximum error are used as four input parameters, which are arranged in a fixed order to form an ordered input vector, denoted as the parameter sequence set.

[0033] For each set of parameters, the four inputs are first mapped to intervals so that all their values ​​are reduced to a closed interval greater than zero and less than one. The reduction function R(a) = 1 / (1+a) is used, where a is any input value.

[0034] The reduced parameters are injected as coordinate points into the constructed quadrilateral polyline path according to the order of arrangement. Each polyline connects the coordinates between two parameters, forming a combination of three polylines.

[0035] Calculate the cosine of the inclination angle of the three broken lines in sequence and find their root mean square, denoted as the angle function value; then input the angle function value as the sole independent variable into the defined response accuracy scoring function:

[0036] E(y) = exp(−y²), where y is the angle function value and E(y) is the response accuracy score for the corresponding motion intensity level.

[0037] In a preferred embodiment, the step of comparing the recognition capabilities of the two modules based on a unified timeline includes:

[0038] The sensitivity score of the RFID module and the response accuracy score of the GPS module are time-stamped and bound according to the same time period number to construct a time series comparison table corresponding to the module scores, so as to ensure the temporal consistency and sample comparability of the scoring data;

[0039] For each time period, the two scores bound together are normalized by using the amplitude scaling function T(p)=p / (1+p) to limit the value range of the two scores to between zero and one, thus generating standardized score pairs.

[0040] Standardized rating pairs are used as input vectors and projected sequentially onto a two-dimensional plane coordinate system to construct a set of directed rating paths. The relative angle values ​​of each rating pair in the coordinate space are used as the discrimination parameter, and the cosine of the angle is used as the discrimination parameter.

[0041] Construct a scoring fusion function, taking the cosine value of the angle between each scoring pair as the unique input. Input the cosine value of the angle into the scoring fusion function F(c)=1−c², and calculate the applicability score of the two modules corresponding to the current exercise intensity under the current exercise intensity level, where c is the cosine value of the angle and F(c) is the fusion output result.

[0042] In a preferred embodiment, the step of constructing a dynamic sampling frequency control mechanism based on the applicability score results includes:

[0043] The standard intensity level matrix defines each level interval as the reference structure for sampling frequency control, clarifies the upper and lower boundaries of the signal sampling frequency corresponding to each motion intensity level, and stores this mapping relationship in the dynamic control parameter library.

[0044] Based on the module applicability score obtained within the current time period, match the corresponding exercise intensity level range, call the frequency control parameter group bound to the level in the parameter library, and obtain the corresponding signal sampling frequency modulation range.

[0045] Within the selected sampling frequency modulation range, and considering the applicability score's position within the grade range, the specific sampling frequencies of the RFID module and the GPS module are determined and marked as scheduling execution values.

[0046] The scheduling execution value is transmitted to the control channels of the RFID module and the GPS module to update the actual sampling frequency of each module, thereby completing the synchronous collection and time alignment of pet identity information and location information.

[0047] In a preferred embodiment, the RFID+GPS integrated pet identification and positioning system specifically includes:

[0048] The exercise intensity modeling module is used to build an exercise intensity reference model, collect historical movement trajectory data of different types of pets, extract acceleration change features corresponding to typical movement behaviors, and construct a standard intensity level matrix for quantifying exercise status based on the combination rules of acceleration amplitude, movement frequency and duration.

[0049] The RFID performance evaluation module is used to obtain the response characteristics of the RFID module to different motion states, collect the frequency of signal response intensity change of the RFID module during identity recognition, and combine the reading stability index and recognition rate to generate the sensitivity score of the RFID module under various motion intensities.

[0050] The Global Positioning Performance Evaluation Module is used to obtain the response characteristics of the Global Positioning System (GPS) module to different motion states. It collects the GPS module's positioning refresh frequency, displacement distance change value, and positioning error value under various motion states, and generates a response accuracy score for the GPS module under different levels of motion intensity.

[0051] The recognition capability fusion module is used to fuse the recognition performance of the RFID module and the GPS module on a unified time axis, match and calculate the sensitivity score and response accuracy score, and generate the applicability score of the two modules under the current motion intensity.

[0052] The dynamic sampling frequency control module is used to construct a dynamic sampling frequency control mechanism based on the applicability score results. It calls the frequency control parameters corresponding to the current motion intensity level in the standard intensity level matrix, determines the specific sampling frequency modulation values ​​of the radio frequency identification module and the global positioning system module, and performs real-time update control.

[0053] The information synchronization acquisition module is used to receive and execute the sampling scheduling instructions generated by the dynamic sampling frequency control module, driving the radio frequency identification module and the global positioning system module to synchronously acquire identity information and location information within the same time period, so as to achieve consistent matching of pet identification and location data.

[0054] The technical effects and advantages of this invention are as follows:

[0055] This invention constructs a dynamic sampling frequency control mechanism based on applicability scoring. This mechanism can adjust the sampling frequencies of the RFID module and the GPS module in tandem according to different levels of motion intensity, ensuring the synchronization of their signal acquisition cycles over time. This achieves precise alignment of identification data and location information during actual monitoring, avoiding time mismatches where tag identification information precedes or lags the actual location coordinates. This synchronous acquisition mechanism is particularly suitable for complex dynamic scenarios where pets are moving at medium to high speeds or engaging in vigorous behavior. While the positioning system provides real-time feedback, it ensures the integrity and effectiveness of the identification link, improving the accuracy and consistency of the fusion system's recognition of moving targets. Compared to existing solutions that rely solely on static frequency configuration or single-module acquisition, this invention significantly improves the consistency and timeliness of dual-mode identification results, enhancing the overall system's fusion stability and task execution efficiency in dynamic environments.

[0056] This invention establishes a dual-module scoring system with a unified timeline by evaluating the recognition sensitivity of the RFID module and the response accuracy of the GPS module under different motion intensity levels. Based on this, a fusion scoring mechanism is constructed, enabling the system to quantify the adaptability and response performance differences of the two modules in real time under different motion states. By mapping the scoring results to the corresponding motion intensity levels and dynamically setting the sampling frequency control parameters, the system can automatically adjust the allocation of sampling frequency resources based on the changing trends of the target behavior state, achieving precise scheduling of module acquisition rhythm and avoiding sampling redundancy, resource waste, or response delay problems caused by the performance imbalance of a single module. In actual operation, this scoring fusion mechanism can adapt to various behavior states, from stationary and low-speed cruising to high-speed running, exhibiting broad adaptability and dynamic control capabilities, significantly improving the utilization efficiency of module resources and the stability of the recognition link.

[0057] This invention addresses the issues of asynchronous response and accuracy imbalance between traditional RFID and GPS modules during high-speed dynamic behavior. It proposes a control mechanism based on module scoring fusion, enabling real-time comparison of module response capabilities on a unified timeline and automatically adjusting signal acquisition frequency strategies based on the fusion score. This enhances the system's robustness against external disturbances and sudden events. When the target pet is in a state of drastic behavior such as being startled, running away, or suddenly turning, the system can improve the sensing coverage of key modules by encrypting the frequency, ensuring high synchronization between identification and positioning information even when the target is moving rapidly or traversing complex paths, guaranteeing uninterrupted tracking. This mechanism reflects the coordination consistency between modules through fusion scoring, proactively avoiding data offset risks caused by significant differences in module response capabilities. This strengthens the overall dynamic adaptability of the system and its continuous tracking capability during critical behavioral phases, effectively supporting the implementation of high-reliability, high-precision pet identification and real-time positioning tasks. Attached Figure Description

[0058] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0059] Figure 1 This is a schematic diagram of the RFID+GPS fusion pet identification and positioning method in this invention.

[0060] Figure 2 This is a schematic diagram of the RFID+GPS integrated pet identification and positioning system of the present invention. Detailed Implementation

[0061] 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.

[0062] Reference Figure 1 - Figure 2 The following examples were obtained:

[0063] Example 1: A pet identification and location method integrating RFID and GPS, comprising the following steps:

[0064] A motion intensity reference model is established. Based on historical motion trajectory data of different types of pets, acceleration features corresponding to typical motion behaviors are extracted to construct a standard intensity level matrix for quantifying motion states. This step is mainly used to establish a mapping system between behavior and intensity levels, ensuring that the system has the basic ability to accurately identify the dynamic state of pets. By collecting trajectory data of pets of multiple breeds, sizes, sexes, and ages in natural environments, combined with continuous acceleration signals obtained by the inertial measurement unit, behavioral feature samples containing key parameters such as motion frequency, speed fluctuation, and displacement period are extracted. Typical behavior types, such as standing still, walking slowly, and running quickly, are identified by combining expert annotation or cluster analysis. Subsequently, a comprehensive judgment is made based on factors such as the amplitude of acceleration change, the length and rate of change of the action cycle, etc., to divide multi-level intensity level boundaries, forming a set of multi-dimensional parameter-driven level division matrices. This matrix serves as the basic reference standard for subsequent module parameter dynamic scheduling and behavior scoring.

[0065] This step involves acquiring the response characteristics of the RFID module to motion states, collecting the frequency of signal response intensity changes during identification, and combining the stability and reading rate of signal reading to generate sensitivity scores for the RFID module at different motion intensities. The aim is to establish an evaluation mechanism for the RFID module's motion adaptability by quantitatively analyzing its performance under different motion states. Under the established standard intensity level matrix, typical behavioral scenarios at each motion level are dynamically sampled, collecting the response intensity changes of the identification module to the target identity tag, recording the frequency changes of signal fluctuations, and calculating the data loss intervals, number of successful responses, and stability during continuous reading. By setting different identification success thresholds, the number of intervals, maximum failure duration, and reading success rate per unit time are processed and incorporated into the sensitivity scoring model along with the actual sampling rate. Finally, an identification adaptability score is generated for each motion intensity level to measure the module's identification reliability under dynamic interference.

[0066] This step involves acquiring the response characteristics of a Global Positioning System (GPS) module to motion states. It collects the positioning refresh frequency, displacement distance changes, and positioning error values ​​of the GPS module under various motion states, generating response accuracy scores for different motion intensities. The aim is to perform feature modeling and performance quantification of the GPS module's performance trends under different motion states. By setting multiple motion intensity levels and sequentially executing data acquisition tasks, the position information sequence output by the module at each level is collected, the time interval between positioning points is extracted, and the positioning refresh frequency is calculated. Simultaneously, the displacement difference between adjacent coordinate points is analyzed to form a distance change time series, and the trajectory stability is estimated by combining the total path length and overall motion time. Furthermore, a sliding time window is used to process continuous position points, estimating the single-point deviation value at each moment, and statistically analyzing the variance range and maximum fluctuation amplitude of the error. Using these multiple performance indicators as input features, an accuracy response scoring model is constructed to uniformly score the positioning accuracy and stability of the module under different dynamic conditions, forming a response accuracy score result set.

[0067] Based on a unified timeline, the recognition capabilities of the two modules are compared. The sensitivity score of the RFID module and the response accuracy score of the GPS module are fused to obtain the applicability score of the two modules for the current motion intensity. This step is mainly used to realize the synchronous comparison and evaluation fusion mechanism of multi-module recognition performance, ensuring data consistency and the accuracy of scheduling decisions. The system uses a unified clock mechanism to bind the sensitivity score and response accuracy score according to timestamps, constructing a time-series comparison structure for module scores, and normalizing the scores to form comparable score pairs. Further, this is represented as a two-dimensional vector path in coordinate space. By extracting the geometric relationship features of the score pairs in space, the relative deviation of the score state is obtained. This is used as an input parameter mapped into the score fusion model, using the angle relationship as a feature factor to calculate the fused module applicability score, which serves as the basis for determining the system sampling priority and scheduling for the current period, realizing the dynamic determination of the collaborative capability of the two modules.

[0068] A dynamic sampling frequency control mechanism is constructed based on the suitability score results. The signal sampling frequencies of the RFID and GPS modules are set according to the current activity level, and the sampling frequencies are adjusted in real time to achieve synchronous acquisition and matching of pet identity and location information. This step uses dynamic control to allocate sampling resources and switch acquisition strategies between the RFID and GPS modules under different activity states. The system uses the suitability score as the decision input, first matching the activity level and retrieving the corresponding sampling frequency range from the standard activity level matrix. Then, based on the relative position of the score within that range, the system sets the sampling frequency value that the two modules should currently execute and sends this value as a frequency scheduling command to the control interfaces of each module. Upon receiving the control command, the modules update their frequencies, achieving real-time response to the current behavior state. By synchronizing the time of identity and location information, the system ensures cross-module collaboration at the sampling cycle level, ultimately achieving a high degree of consistency in the fusion of pet identification and location information.

[0069] The steps for constructing a standard intensity level matrix for quantifying movement states include: collecting time-series trajectory data of pets in their natural environment based on behavioral data samples from multiple breeds, sizes, and age groups, and acquiring continuous acceleration signals using an inertial measurement unit (IMU). Breeds include, but are not limited to, mammals such as Bichon Frises, Golden Retrievers, Border Collies, French Bulldogs, British Shorthairs, and Ragdolls; sizes include small, medium, and large; and age groups include infants (3 months to 1 year), adults (1 to 7 years), and seniors (7 years and older). Natural environment activities include typical behavioral scenarios such as walks with owners, running in parks, foraging independently, and jumping within enclosures. Time-series trajectory data refers to position sequences collected at equal time intervals, with a recording interval of once per second and a duration of at least 600 seconds, to cover the complete movement cycle. The inertial measurement unit is a triaxial accelerometer, installed on the pet's neck collar. The sampling frequency is set to 100 Hz, and it can record more than 300 acceleration vectors per unit time, including linear acceleration values ​​in the X-axis forward and backward direction, the Y-axis left and right direction, and the Z-axis up and down direction. The unit of measurement is meters per square second, and the recording range is from -16 to +16 meters per square second.

[0070] The collected acceleration data is normalized to extract behavioral feature indicators containing multiple preset motion behavior dimensions, and the corresponding typical motion behaviors are identified based on the feature change patterns. Normalization involves processing the acceleration data from each channel with zero mean and converting it into a dimensionless signal sequence at the standard deviation scale to remove the offset caused by differences in individual body size, posture, and gravity direction on signal intensity. The extracted behavioral feature indicators include: the number of periodic acceleration peaks (used to determine running or jumping frequency), the maximum acceleration value (measuring explosive movements), the rate of change of acceleration (used to distinguish between uniform and abrupt changes), the average acceleration level (assessing the duration of the movement), and the amplitude modulation structure (determining whether it is rhythmic behavior). Taking a 300-second sample as an example, if there are more than 10 acceleration peaks periodically every 30 seconds along the Z-axis, and the X and Y axes show obvious synchronous amplitude oscillations, it can be preliminarily judged as high-speed running behavior; if the average acceleration value across all three axes is less than 0.5 m / s², and there are no obvious rhythmic fluctuations, it can be judged as a stationary state. Based on the combined changing trends of multiple indicators, and combined with template matching algorithms or classification models, behavior recognition is performed. The typical behaviors identified include six types of typical movement states: stationary, slow walking, fast running, short sprint, continuous jumping, and sudden turning.

[0071] Based on the behavioral classification results, corresponding exercise intensity grading rules are constructed. Level boundaries are set according to the combination patterns of acceleration variation amplitude, movement frequency, and duration, forming a standard intensity level matrix. The exercise intensity grading rules are based on six identified typical behaviors, constructing intensity grading levels from low to high, ranging from level 0 to level 5. Level 0 represents complete stillness, and level 5 represents high-speed, continuous, and uninterrupted explosive movement. The grading criteria are as follows: When the average acceleration is less than 0.5 m / s², and the movement duration is greater than 120 seconds, it is classified as a level 0 still state; when the acceleration fluctuation is between 0.5 and 1.5 m / s², the movement frequency is less than 1 time per second, and the duration is greater than 30 seconds, it is classified as a level 1 low-intensity walking state; when the peak acceleration exceeds 2.5 m / s², the movement frequency is greater than 2 times per second, and the movement duration exceeds 20 seconds, it is classified as a level 3 rapid movement state; when the peak acceleration exceeds 4.5 m / s², the frequency exceeds 3.5 times per second, and the duration is not less than 10 seconds, it is classified as a level 5 high-intensity explosive state. The above-mentioned grading rules are encapsulated in matrix form, with rows representing motion levels (level 0 to 5) and columns representing three indicators (acceleration amplitude, frequency, and duration). The matrix cells store the boundary thresholds for each level, forming a standard intensity level matrix for system control and scoring model reference.

[0072] The steps for obtaining the response characteristics of the RFID module to motion states include:

[0073] Under different exercise intensity levels, continuous reading data from the RFID module on pet identification tags was recorded, resulting in a continuous time series of signal strength changes. The RFID module is a device used to identify passive electronic tags via radio frequency. The RFID tag attached to the pet is a passive transponder containing identification information. The identification module sends radio frequency signals to the tag within a certain distance and receives the tag's response signal. When the pet's movement changes, its posture, speed, and environmental obstructions affect the continuity and strength of the module's signal readings. At exercise intensity level 0 (i.e., a stationary state), the continuous reading success rate can stably reach more than once per second, with signal strength fluctuations within 3 dB. At exercise intensity level 5 (high-intensity jumping and running), continuous data readings are frequently interrupted, the success rate drops to below 0.2 times per second, and the signal strength fluctuation range can reach 10 dB. The recorded data is a time series with time as the horizontal axis and signal strength as the vertical axis. Each sampling point represents the actual received signal strength value of the module at that moment, expressed in decibels and milliwatts. The sampling period was 10 times per second, and the recording length was set to 30 seconds, resulting in 300 signal data points for subsequent fluctuation analysis.

[0074] The stability and sustainability of the identification signal under various motion levels are evaluated by acquiring the number of interruptions, average read success rate, and maximum loss interval duration during the signal reading process, forming a corresponding stability index matrix. The number of interruptions refers to the number of interruptions that occur during continuous reading, transitioning from successful reads to read failures. The average read success rate is the number of times an RFID tag is successfully identified per unit time divided by the total number of sampling attempts. For example, if 300 samples are taken in 30 seconds and 150 tag data are successfully read, the average read success rate is 0.5. The maximum loss interval duration represents the longest period of consecutive read failures. For example, at motion intensity level 4, if there is no signal response for 8 consecutive seconds, the maximum loss interval is 8 seconds. The statistical results of the three indicators for each motion intensity level are filled into the three columns of the matrix. The number of rows in the matrix equals the number of intensity levels, for a total of six levels, with each row corresponding to one level, forming a six-row, three-column stability index matrix. This matrix is ​​used to quantitatively describe the signal stability performance of the RFID module under different intensities, reflecting the performance fluctuation trend of the module in dynamic identification in a real working environment.

[0075] By combining the reading rate and stability index matrix, the recognition performance of the RFID module under different motion intensity levels is quantitatively scored, a set of sensitivity score datasets for module applicability evaluation is generated and input into the module applicability scoring model, and the sensitivity scores of the RFID module corresponding to different motion intensities are output.

[0076] During signal acquisition, the continuous signal strength change time series is marked as follows: The signal strength change time series refers to the ordered sequence of received signal strength values ​​continuously recorded by the RFID module at each moment within a fixed time interval. The signal unit is decibels and milliwatts (dBmW), and the sampling frequency is 10 Hz, i.e., 10 sampling points per second. Time points with signal strength higher than a set threshold are marked as valid read points. The threshold is defined as the minimum valid read value when the RFID module stably receives signals in a static state, specifically set to -65 dBmW. Time points with signal strength lower than the set threshold or no signal response are marked as invalid read points. Within a standard 30-second recording period, the total number of sampling points is 300. If the signal strength of five consecutive sampling points is lower than -70 dBmW within a certain period, this segment is considered a continuous signal loss state. The number of times the valid read interval is interrupted is counted as the interval count. The interval count refers to the number of events during signal acquisition where the valid read point sequence is interrupted by one or more invalid read points and the counting restarts. For example, if there are 20 consecutive valid reads followed by 3 consecutive invalid reads, and then the state is restored to valid status, it is considered as one intermittent event; if there are 5 more invalid reads after that, and the state is restored, then the cumulative number of intermittent events is 2.

[0077] The average read success rate is calculated by comparing the number of valid read points to the total number of read points per unit time. For example, if 300 samples are taken within 30 seconds and 150 valid read points are obtained, the average read success rate is 50%. If only 90 valid read points are obtained within the same sampling period, the success rate drops to 30%. The longest consecutive period without valid read points is defined as the maximum loss interval. The maximum loss interval reflects the most severe continuous communication interruption. For example, in a motion scenario, if a situation occurs with nine consecutive seconds of no valid read points at a sampling frequency of 10 Hz, then 90 consecutive invalid read points constitute the maximum loss interval, recorded as nine seconds.

[0078] In the process of quantifying and scoring the recognition performance of the RFID module under different motion intensity levels, a suitability scoring model is constructed using a sequence structure modeling approach: recognition performance refers to the ability of the RFID module to maintain stable, continuous, and high-success-rate readings under a specific motion intensity level. The motion intensity levels are set to levels zero to five according to the aforementioned standard level matrix, representing the state division from stillness to high-intensity burst motion.

[0079] The original read rate value, intermittent count value, average read success rate value, and maximum loss interval value for each motion level are sequentially encoded into a numerical vector of the same length. The read rate value is the number of successful reads per second, for example, 0.6 times per second; the intermittent count is the number of read interruption events within every 30 seconds, for example, four times; the average read success rate is in percentage form, such as 45 percent; and the maximum loss interval is the longest communication interruption time in seconds, such as six seconds. These four data items are sequentially arranged into a one-dimensional numerical vector of length four.

[0080] The input vectors corresponding to each set of motion levels are projected onto a two-dimensional numerical plane in a fixed order to form multiple sets of state paths. Specifically, the reading rate and average reading success rate are mapped to the horizontal axis, and the number of intermittents and the maximum loss interval are mapped to the vertical axis, forming four coordinate points in the two-dimensional plane. These points are connected in vector order to form three continuous broken line paths. Each segment reflects the changing trend between two adjacent indicators. The state path contains angle, length, and direction information.

[0081] Curve fitting is performed on each state path, and the discrete gradient value sequence of the curve is calculated and its variance is statistically analyzed. This variance serves as the response fluctuation index of the RFID module at each motion intensity level. The discrete gradient value refers to the slope change between adjacent line segments, representing the degree of fluctuation of the performance parameter between indicators. For example, if the slope of the first line segment is 0.8, the second segment is -0.2, and the third segment is 0.5, then the discrete gradient sequence is 0.8, -1.0, 0.7, and the calculated gradient variance is approximately 0.68. The larger the variance, the more drastic the parameter fluctuation of the module at that motion level.

[0082] The response fluctuation index is input into a predefined logarithmic mapping function, with the response fluctuation index as the independent variable, to calculate the corresponding sensitivity score. The logarithmic mapping function is defined as S(x) = ln(1+x), where x is the response fluctuation index and S(x) is the sensitivity score. The domain of the function is non-negative real numbers, and the range is positive real numbers. The score increases monotonically with the increase of the response fluctuation index, indicating that the higher the degree of fluctuation in recognition performance, the higher the score, and the worse the module's adaptability to the current motion state. Taking a response fluctuation index of x = 0.68 as an example, substituting into the function, we get S(x) = ln(1+0.68) ≈ ln(1.68) ≈ 0.518. After completing this scoring process for all motion intensity levels, a complete sensitivity score vector is obtained, which is used in the subsequent module recognition capability fusion scoring step.

[0083] The steps for obtaining the response characteristics of a Global Positioning System (GPS) module to motion states include:

[0084] Under set multi-level exercise intensity, location information data continuously output by the GPS module was collected, and the time intervals in the collected data were statistically analyzed to calculate the positioning refresh frequency for each level, constructing a frequency time series. The GPS module is a spatial positioning device used to obtain the pet's geographic coordinates and has the ability to periodically emit location signals. The set exercise intensity levels are divided into six levels, representing six different exercise states: stationary, slow walking, steady running, rapid sprinting, intermittent jumping, and vigorous movement. At each level, GPS location information was collected continuously for ten minutes, with a sampling period of one second, meaning latitude and longitude coordinates and a timestamp were recorded once per second. The positioning refresh frequency refers to the number of times the location information is effectively updated per unit of time, measured in times per second. In the stationary state, the GPS module refresh frequency remains stable at once per second; in the vigorous movement state, due to signal drift and obstruction, the actual refresh frequency may drop to 0.7 times per second. By continuously recording the refresh count per second, a refresh frequency time series of length 600 was obtained.

[0085] Based on the coordinate changes of adjacent points in the frequency time series, the displacement distance change per unit time is calculated point by point. Combined with the total path length and duration, the average displacement rate and trajectory fluctuation index for each movement level are extracted. The coordinate changes of adjacent points are obtained by calculating the spherical distance between each pair of consecutive latitude and longitude coordinates, in meters. The displacement distance change per unit time is the straight-line distance the pet moves per second. In 600 sample points, 599 displacement differences are obtained for each pair of adjacent points, forming a displacement change sequence. The average displacement rate is the sum of all differences divided by the total duration, in meters per second. In slow walking mode, the average rate is approximately 0.6 meters per second; in rapid sprinting mode, the average rate can reach 2.5 meters per second. The trajectory fluctuation index is a statistical value measuring the degree of path change, calculated as the standard deviation of the turning angle change of the path formed by three adjacent points. It reflects the stability of the movement; for example, in violent jumping, this index can reach as high as 45 degrees.

[0086] For each set of data, single-point error estimation is performed on the positioning error. A sliding window approach is used to calculate the error variance and maximum error value under the same motion level, generating an error distribution parameter set for the current motion intensity. Positioning error refers to the spatial deviation between the location information output by the Global Positioning System (GPS) module at any sampling time point and the reference trajectory path, measured in meters. The reference trajectory path is a local motion trend line obtained by smoothing a continuous sequence of historical location points, simulating the actual movement trend of a target entity in space without relying on additional external measurement equipment. Single-point error estimation involves calculating the distance between the current location information point and the corresponding coordinate position at the time point on the reference path fitted within the sliding time window; this distance is the current single-point error.

[0087] The reference path was constructed using a sliding window approach, with a window width of ten seconds. Sampling was performed once per second within the window, containing a total of ten location points. Least squares smoothing was used to fit the trajectory of these ten points, generating a locally estimated path segment. At the center of the window, the Euclidean distance from the current recorded point to its nearest neighbor on the fitted path was calculated based on the fitting results, and this distance was used as the positioning error value for that point.

[0088] At each motion intensity level, the above operation is repeated for all sampling time points to obtain a complete error time series. Furthermore, the error variance and maximum error value are statistically analyzed within each sliding window to form the error distribution parameter set for that level, used to evaluate the position response accuracy of the GPS module under the current motion state. A response accuracy evaluation function is constructed, using the refresh frequency, displacement change rate, error variance, and maximum error corresponding to each motion level as multivariate inputs. A multi-parameter nonlinear fitting method is used to calculate the response accuracy score of the GPS module at different motion intensity levels.

[0089] The multi-parameter nonlinear fitting method refers to using four input parameters—refresh frequency, displacement rate of change, error variance, and maximum error—arranged in a fixed order as an ordered input vector, denoted as the parameter sequence set. The refresh frequency represents the number of times the GPS module provides valid location information per unit time, measured in times per second (e.g., 0.9). The displacement rate of change represents the average speed of the pet's movement over a continuous period, measured in meters per second (e.g., 1.2). The error variance refers to the statistical dispersion of the position deviation within a sliding time window (e.g., 0.08). The maximum error is the largest deviation observed within the same window, measured in meters (e.g., 1.5). For each parameter sequence, the four inputs are first interval-mapped to reduce their values ​​to a closed interval greater than zero and less than one, using the reduction function R(a) = 1 / (1+a), where a is any input value. Taking the input parameter sequence [0.9, 1.2, 0.08, 1.5] as an example, substituting them into the reduction function yields R(0.9)≈0.526, R(1.2)≈0.454, R(0.08)≈0.926, and R(1.5)≈0.4, forming the reduction parameter sequence [0.526, 0.454, 0.926, 0.4].

[0090] The reduced parameters are used as coordinate points in their ordered sequence and injected into a constructed quadrilateral polyline path. Each polyline connects the coordinates between two parameters, forming a three-segment polyline combination. The polyline path is defined in a two-dimensional number axis space, where the horizontal axis represents the parameter sequence index order and the vertical axis represents the reduced parameter values. The three polyline segments connect the following point pairs: (1,0.526)-(2,0.454), (2,0.454)-(3,0.926), and (3,0.926)-(4,0.4).

[0091] Calculate the cosine of the inclination angle of each of the three broken line segments sequentially and find their root mean square, denoted as the angle function value. The cosine of the inclination angle is obtained by using the cosine of the angle between the line connecting the two points and the horizontal axis. For example, the cosine of the angle of the first broken line segment is cosθ1≈0.993; the cosine values ​​of the angles of the three broken line segments are c1, c2, and c3 respectively. Assuming c1=0.993, c2=0.736, and c3=0.776, then y≈0.842.

[0092] The angle function value is then input as the sole independent variable into the defined response accuracy scoring function: E(y) = exp(−y²), where y is the angle function value and E(y) is the response accuracy score for the corresponding motion intensity level. Substituting y = 0.842 in the example above, we get E(y) = exp(−0.842²) ≈ exp(−0.709) ≈ 0.491. The response accuracy score is a quantitative result of the GPS module's ability to recognize changes in target trajectory at a given motion level. A higher score indicates better module response quality, while a lower score indicates unstable response or larger errors at that level. Since a smaller angle function value indicates stronger synergy among the four input parameters, after exponential function mapping, a larger score indicates better response capability of the GPS module at the current motion intensity. This method is repeated for all motion intensity levels to complete the response accuracy scoring for all levels, generating a set of response capability score vectors that correspond one-to-one with the standard intensity level matrix. These vectors are used for comparison and integration with the sensitivity score of the RFID module in the subsequent fusion scoring mechanism.

[0093] The steps for comparing the recognition capabilities of the two modules based on a unified timeline include: binding the sensitivity score of the RFID module and the response accuracy score of the GPS module with time tags according to the same time period number, constructing a time series comparison table corresponding to the module scores, and ensuring the temporal consistency and sample comparability of the score data. The time tags are discrete time point identifiers with consecutive numbers, such as T1, T2, T3, etc. Each time tag is bound to a set of score data pairs. For example, T1 corresponds to a score of 0.63 for the RFID module and a score of 0.47 for the GPS module, forming a score pair [0.63, 0.47].

[0094] For each time period, the two scores bound together are normalized using the amplitude scaling function T(p) = p / (1+p), which limits the range of the two scores to between zero and one, thus generating standardized score pairs. Taking the score pair [0.63, 0.47] as an example, substituting into the amplitude scaling function, we get T(0.63) = 0.63 / 1.63 ≈ 0.386 and T(0.47) = 0.47 / 1.47 ≈ 0.320, resulting in the standardized score pair [0.386, 0.320].

[0095] Standardized scoring pairs are used as input vectors and projected sequentially onto a two-dimensional plane coordinate system to construct a set of directed scoring paths. The discrimination parameter is based on the relative angle between each scoring pair in the coordinate space, with the cosine of the angle as the discrimination parameter. In the two-dimensional plane, the horizontal axis represents the RFID module scoring, and the vertical axis represents the GPS module scoring. The corresponding point on the coordinate axis for each scoring pair is (0.386, 0.320). If vector [1, 0] represents the pure RFID module scoring direction, then the cosine of the angle between the scoring pair vector and the vector is: c = (0.386 × 1 + 0.320 × 0) / √(0.386² + 0.320²) ≈ 0.386 / √(0.149 + 0.102) ≈ 0.386 / 0.502 ≈ 0.769.

[0096] A scoring fusion function is constructed, using the cosine of the angle between each scoring pair as the unique input. This cosine is then input into the scoring fusion function F(c) = 1 − c², calculating the applicability score of the two modules for the current exercise intensity level. Here, c is the cosine of the angle, and F(c) is the fusion output. In the example above, c ≈ 0.769, which, when substituted into F(c) = 1 − (0.769)² ≈ 1 − 0.591 ≈ 0.409. The scoring fusion function quantifies the overall fit by analyzing the differences in the responses of the two modules. A fusion score closer to 1 indicates a greater inconsistency in the responses between the two modules, while a score closer to 0 indicates a higher degree of synergy.

[0097] The steps for constructing a dynamic sampling frequency control mechanism based on applicability scoring results include: using the level intervals defined in the standard intensity level matrix as the benchmark reference structure for sampling frequency control, clarifying the upper and lower boundaries of the signal sampling frequency corresponding to each movement intensity level, and storing this mapping relationship in the dynamic control parameter library. The standard intensity level matrix is ​​an intensity grading index system constructed for different pet behavior states. For example, levels L1 to L5 represent stationary, low-speed movement, medium-speed movement, high-speed movement, and vigorous running states, respectively. Each level is bound to a frequency control interval. For example, the RFID module is bound to a sampling frequency of 0.5–1 Hz in the L1 interval, and the GPS module is bound to a sampling frequency of 0.2–0.5 Hz. All mapping relationships are stored in the dynamic control parameter library in a hash table structure.

[0098] Based on the module suitability score obtained within the current time period, the corresponding motion intensity level range is matched, and the frequency control parameter group bound to that level in the parameter library is called to obtain the corresponding signal sampling frequency modulation range. The suitability score is obtained by nonlinearly calculating the cosine value c of the angle between the scores of the RFID module and the GPS module using the fusion function F(c)=1−c², with a value range of 0 to 1. The closer the fusion score is to 1, the greater the difference in response characteristics between the two modules within that time period, indicating that they are not suitable for collaborative work, and the system should prioritize the more stable module; the closer the fusion score is to 0, the more consistent the responses of the two modules to the current state, and the system can adopt an average sampling configuration. For example, when F(c)=0.72, it is mapped to motion intensity level L4 according to the preset interval threshold, and the sampling frequency modulation range corresponding to level L4 is extracted from the parameter library. For example, the frequency range of the RFID module is 2.0~3.5Hz, and that of the GPS module is 1.5~2.5Hz.

[0099] Within the selected sampling frequency modulation range, and considering the applicability score's position within the grade interval, the specific sampling frequencies of the RFID module and the GPS module are determined and marked as scheduling execution values. To ensure fine-grained adjustment of the score mapping, a linear interpolation method within the grade interval is introduced. For example, grade L4 corresponds to an applicability score interval of [0.65, 0.85]. If the current score is 0.72, its position on the linear mapping curve is (0.72−0.65) / (0.85−0.65)=0.35. Therefore, the sampling frequency is advanced by 35% from the lowest to the highest value. That is, the RFID module sampling frequency is 2.0+0.35×(3.5−2.0)=2.525Hz, and the GPS module sampling frequency is 1.5+0.35×(2.5−1.5)=1.85Hz.

[0100] The scheduling execution value is transmitted to the control channels of the RFID module and the GPS module to update the actual sampling frequency of each module, completing the synchronous acquisition and time alignment of pet identity information and location information. The sampling frequency adjustment command is sent to each module by the central scheduling unit in a protocol message format. The module receives the command and executes the update in the next refresh cycle. The system pairs the identity data with the location information according to a unified timestamp to ensure a one-to-one mapping between tag identification and location data, improving the information synchronization accuracy in dynamic tracking scenarios.

[0101] Example 2: A pet identification and positioning system integrating RFID and GPS, specifically including:

[0102] The exercise intensity modeling module is used to build an exercise intensity reference model, collect historical movement trajectory data of different types of pets, extract acceleration change features corresponding to typical movement behaviors, and construct a standard intensity level matrix for quantifying exercise status based on the combination rules of acceleration amplitude, movement frequency and duration.

[0103] The RFID performance evaluation module is used to obtain the response characteristics of the RFID module to different motion states, collect the frequency of signal response intensity change of the RFID module during identity recognition, and combine the reading stability index and recognition rate to generate the sensitivity score of the RFID module under various motion intensities.

[0104] The Global Positioning Performance Evaluation Module is used to obtain the response characteristics of the Global Positioning System (GPS) module to different motion states. It collects the GPS module's positioning refresh frequency, displacement distance change value, and positioning error value under various motion states, and generates a response accuracy score for the GPS module under different levels of motion intensity.

[0105] The recognition capability fusion module is used to fuse the recognition performance of the RFID module and the GPS module on a unified time axis, match and calculate the sensitivity score and response accuracy score, and generate the applicability score of the two modules under the current motion intensity.

[0106] The dynamic sampling frequency control module is used to construct a dynamic sampling frequency control mechanism based on the applicability score results. It calls the frequency control parameters corresponding to the current motion intensity level in the standard intensity level matrix, determines the specific sampling frequency modulation values ​​of the radio frequency identification module and the global positioning system module, and performs real-time update control.

[0107] The information synchronization acquisition module is used to receive and execute the sampling scheduling instructions generated by the dynamic sampling frequency control module, driving the radio frequency identification module and the global positioning system module to synchronously acquire identity information and location information within the same time period, so as to achieve consistent matching of pet identification and location data.

[0108] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0109] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0110] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0111] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A pet identification and positioning method integrating RFID and GPS, characterized in that, Includes the following steps: Establish an exercise intensity reference model, extract acceleration features corresponding to typical exercise behaviors based on historical movement trajectory data of different types of pets, and construct a standard intensity level matrix for quantifying exercise status; The response characteristics of the RFID module to motion state are obtained, the frequency of signal response intensity change of the RFID module during the recognition process is collected, and the sensitivity score of the RFID module corresponding to different motion intensities is generated by combining the stability and reading rate of signal reading. The response characteristics of the GPS module to motion states are obtained. The positioning refresh frequency, displacement distance change value and positioning error value of the GPS module under various motion states are collected to generate the response accuracy score of the GPS module corresponding to different motion intensities. Based on a unified timeline, the recognition capabilities of the two modules are compared. The sensitivity score of the RFID module and the response accuracy score of the GPS module are fused to obtain the applicability score of the two modules for the current motion intensity. A dynamic sampling frequency control mechanism is constructed based on the applicability score results. The signal sampling frequencies of the RFID module and the GPS module are set according to the current exercise intensity level, and the signal sampling frequencies of the RFID module and the GPS module are adjusted in real time to complete the synchronous collection and matching of pet identity information and location information.

2. The RFID+GPS fusion pet identification and positioning method according to claim 1, characterized in that, The steps for constructing a standard intensity level matrix for quantifying motion states include: Based on behavioral data samples of pets of different breeds, sizes and age groups, time-series trajectory data of their activities in the natural environment were collected, and continuous acceleration signals were collected using an inertial measurement unit. The collected acceleration data is normalized, and behavioral feature indicators containing multiple preset motion behavior dimensions are extracted. The corresponding typical motion behaviors are then identified based on the feature change patterns. Based on the behavioral classification results, corresponding exercise intensity grading rules are constructed, and the grade boundaries are set according to the combination pattern of acceleration change amplitude, movement frequency and duration to form a standard intensity grade matrix.

3. The RFID+GPS fusion pet identification and positioning method according to claim 1, characterized in that, The steps for obtaining the response characteristics of the RFID module to motion states include: Under different levels of exercise intensity, the continuous reading data of the RFID module on the pet's identification tag was recorded to obtain a continuous time series of signal strength changes; The number of intervals, average success rate, and maximum loss interval duration during signal reading are obtained to evaluate the stability and persistence of the signal under various levels of motion, forming a corresponding stability index matrix. By combining the reading rate and stability index matrix, the recognition performance of the RFID module under different motion intensity levels is quantitatively scored, a set of sensitivity score datasets for module applicability evaluation is generated and input into the module applicability scoring model, and the sensitivity scores of the RFID module corresponding to different motion intensities are output.

4. The RFID+GPS fusion pet identification and positioning method according to claim 3, characterized in that, During signal acquisition, the continuous signal intensity change time series is marked as follows: The time points when the signal strength is higher than the set threshold are marked as valid read points, and the time points when the signal strength is lower than the set threshold or there is no signal are marked as invalid read points. The number of times the valid read interval is interrupted is counted as the number of intervals. The ratio of the number of valid read points to the total number of read points per unit time is calculated to obtain the average read success rate. The longest duration of a continuous absence of valid read points is taken as the maximum loss interval duration.

5. The RFID+GPS fusion pet identification and positioning method according to claim 3, characterized in that, In the process of quantifying and scoring the recognition performance of the RFID module under different motion intensity levels, a suitability scoring model is constructed using sequence structure modeling. The original read rate value, intermittent count value, average read success rate value and maximum loss interval duration value under each motion level are encoded into numerical vectors of the same length in sequence. The input vectors corresponding to each set of motion levels are projected onto the two-dimensional numerical plane in a fixed order to form multiple sets of state paths. Curve fitting is performed on each state path, the discrete gradient value sequence of the curve is calculated and its variance is statistically analyzed, which serves as the response fluctuation index of the RFID module under each motion intensity level. The response volatility index is input into the set logarithmic mapping function, and the corresponding sensitivity score is calculated with the response volatility index as the independent variable. The logarithmic mapping function is defined as S(x)=ln(1+x), where x is the response volatility index and S(x) is the sensitivity score.

6. The RFID+GPS fusion pet identification and positioning method according to claim 5, characterized in that, The steps for obtaining the response characteristics of a Global Positioning System (GPS) module to motion states include: Under the set multi-level motion intensity, the location information data continuously output by the global positioning system module is collected, and the time intervals in the collected data are statistically analyzed to calculate the positioning refresh frequency under each level and construct a frequency time series. Based on the coordinate changes of adjacent points in the frequency time series, the displacement distance change value per unit time is calculated point by point, and combined with the total path length and duration, the average displacement change rate and trajectory fluctuation index under each motion level are extracted. For each set of data, single-point error estimation is performed on the positioning error. The error variance and maximum error value under the same motion level are calculated using a sliding window method to generate an error distribution parameter set for the current motion intensity. A response accuracy evaluation function is constructed, taking the refresh frequency, displacement change rate, error variance, and maximum error corresponding to each motion level as multivariate inputs, and calculating the response accuracy score of the Global Positioning System module under different motion intensity levels through a multi-parameter nonlinear fitting method.

7. The RFID+GPS fusion pet identification and positioning method according to claim 6, characterized in that, Multi-parameter nonlinear fitting methods refer to: The refresh frequency, displacement change rate, error variance, and maximum error are used as four input parameters, which are arranged in a fixed order to form an ordered input vector, denoted as the parameter sequence set. For each set of parameter sequences, the four inputs are first mapped to intervals so that all their values ​​are reduced to a closed interval greater than zero and less than one. The reduction function R(a) = 1 / (1+a) is used, where a is any input value. The reduced parameters are injected as coordinate points into the constructed quadrilateral polyline path according to the order of arrangement. Each polyline connects the coordinates between two parameters, forming a combination of three polylines. Calculate the cosine of the inclination angle of the three broken lines in sequence and find their root mean square, denoted as the angle function value; then input the angle function value as the sole independent variable into the defined response accuracy scoring function: E(y) = exp(−y²), where y is the angle function value and E(y) is the response accuracy score for the corresponding motion intensity level.

8. The RFID+GPS fusion pet identification and positioning method according to claim 7, characterized in that, The steps for comparing the recognition capabilities of the two modules based on a unified timeline include: The sensitivity score of the RFID module and the response accuracy score of the GPS module are time-stamped and bound according to the same time period number to construct a time series comparison table corresponding to the module scores, so as to ensure the temporal consistency and sample comparability of the scoring data; For each time period, the two scores bound together are normalized by using the amplitude scaling function T(p)=p / (1+p) to limit the value range of the two scores to between zero and one, thus generating standardized score pairs. Standardized rating pairs are used as input vectors and projected sequentially onto a two-dimensional plane coordinate system to construct a set of directed rating paths. The relative angle values ​​of each rating pair in the coordinate space are used as the discrimination parameter, and the cosine of the angle is used as the discrimination parameter. Construct a scoring fusion function, taking the cosine value of the angle between each scoring pair as the unique input. Input the cosine value of the angle into the scoring fusion function F(c)=1−c², and calculate the applicability score of the two modules corresponding to the current exercise intensity under the current exercise intensity level, where c is the cosine value of the angle and F(c) is the fusion output result.

9. The RFID+GPS fusion pet identification and positioning method according to claim 8, characterized in that, The steps for constructing a dynamic sampling frequency control mechanism based on applicability score results include: The standard intensity level matrix defines each level interval as the reference structure for sampling frequency control, clarifies the upper and lower boundaries of the signal sampling frequency corresponding to each motion intensity level, and stores this mapping relationship in the dynamic control parameter library. Based on the module applicability score obtained within the current time period, match the corresponding exercise intensity level range, call the frequency control parameter group bound to the level in the parameter library, and obtain the corresponding signal sampling frequency modulation range. Within the selected sampling frequency modulation range, and considering the applicability score's position within the grade range, the specific sampling frequencies of the RFID module and the GPS module are determined and marked as scheduling execution values. The scheduling execution value is transmitted to the control channels of the RFID module and the GPS module to update the actual sampling frequency of each module, thereby completing the synchronous collection and time alignment of pet identity information and location information.

10. A pet identification and positioning system integrating RFID and GPS, based on the pet identification and positioning method integrating RFID and GPS as described in any one of claims 1-9, characterized in that, Specifically, it includes: The exercise intensity modeling module is used to build an exercise intensity reference model, collect historical movement trajectory data of different types of pets, extract acceleration change features corresponding to typical movement behaviors, and construct a standard intensity level matrix for quantifying exercise status based on the combination rules of acceleration amplitude, movement frequency and duration. The RFID performance evaluation module is used to obtain the response characteristics of the RFID module to different motion states, collect the frequency of signal response intensity change of the RFID module during identity recognition, and combine the reading stability index and recognition rate to generate the sensitivity score of the RFID module under various motion intensities. The Global Positioning Performance Evaluation Module is used to obtain the response characteristics of the Global Positioning System (GPS) module to different motion states. It collects the GPS module's positioning refresh frequency, displacement distance change value, and positioning error value under various motion states, and generates a response accuracy score for the GPS module under different levels of motion intensity. The recognition capability fusion module is used to fuse the recognition performance of the RFID module and the GPS module on a unified time axis, match and calculate the sensitivity score and response accuracy score, and generate the applicability score of the two modules under the current motion intensity. The dynamic sampling frequency control module is used to construct a dynamic sampling frequency control mechanism based on the applicability score results. It calls the frequency control parameters corresponding to the current motion intensity level in the standard intensity level matrix, determines the specific sampling frequency modulation values ​​of the radio frequency identification module and the global positioning system module, and performs real-time update control. The information synchronization acquisition module is used to receive and execute the sampling scheduling instructions generated by the dynamic sampling frequency control module, driving the radio frequency identification module and the global positioning system module to synchronously acquire identity information and location information within the same time period, so as to achieve consistent matching of pet identification and location data.

Images