Animal dynamic trajectory tracking method and system based on RFID

Abstract

The invention provides an animal dynamic trajectory tracking method and system based on RFID, and the method comprises the steps: obtaining the data of a monitoring region, detecting an animal behavior mode in real time in the monitoring region, and carrying out the analysis; dividing the whole monitoring area based on an analysis result; dynamically adjusting the grid density according to animal behavior modes in different areas; on the basis of grid adjustment, correspondingly optimizing the signal intensity in the region; integrating the multi-dimensional signals into a comprehensive signal intensity matrix for track prediction; according to the comprehensive signal intensity matrix, constructing a prediction model combining Kalman filtering and Bayesian signal adjustment to predict the future trajectory of the animal; the predicted future trajectory of the animal is corrected based on the animal behavior mode, and the accuracy and precision of the trajectory are improved by dynamically adjusting the prediction of the trajectory. According to the invention, not only are the accuracy and reliability of animal trajectory tracking improved, but also the energy efficiency of the system is remarkably improved, and the wide development of the RFID animal tracking technology in practical application is promoted.

Description

Technical Field

[0001] The present invention belongs to the field of dynamic track tracking, and in particular relates to an animal dynamic track tracking method and system based on RFID. Background Art

[0002] In modern agriculture, breeding and animal research, animal tracking systems based on RFID (radio frequency identification) technology have been widely used. RFID technology uses radio waves to identify the tags worn by animals to achieve real-time tracking of the animal's location. This technology plays an important role in animal management, disease prevention and control, behavioral research and precision agriculture due to its non-contact, long-distance and no need for direct visual contact. Traditional RFID animal tracking systems are generally composed of RFID tags, readers and data processing platforms. The reader transmits radio waves to communicate with the RFID tags worn by animals to obtain the animal's location data. These data can help farms, zoos, scientific research institutions, etc. monitor the animal's location in real time, make management decisions, and analyze the animal's behavior patterns.

[0003] However, despite the wide application of RFID technology in animal tracking, the existing RFID tracking system still has many problems to be solved. Especially in large-scale animal groups or complex environments, traditional RFID technology faces a series of challenges such as positioning accuracy, energy efficiency and data processing. First of all, insufficient positioning accuracy is one of the main defects of the current RFID system. The reading accuracy of RFID tags is usually affected by many factors, such as signal attenuation, environmental interference, and obstacle occlusion. Especially in large animal groups, when multiple animals are in the same area, the signals interfere with each other, resulting in increased positioning errors. Most traditional RFID systems rely on static or uniform grid layouts, which greatly affects the positioning accuracy in high-density areas. Secondly, energy efficiency is also one of the pain points of traditional RFID animal tracking systems. In large-scale animal tracking systems, multiple RFID readers usually need to work 24 hours a day, which not only increases the energy consumption of the equipment, but also increases the difficulty and cost of system maintenance. Due to the lack of dynamic adjustment mechanism in existing systems, it is often impossible to adjust the grid layout and reading frequency according to the density of animal activities and environmental changes, resulting in unnecessary energy waste. Third, traditional RFID systems usually lack the ability to intelligently analyze and adapt to animal behavior, which makes it difficult to maintain stable tracking under the complex behavior patterns of group animals. For example, when multiple animals gather together or stay in a certain area, it is difficult for the system to accurately distinguish the location of each animal, which in turn affects the tracking effect.

[0004] In order to solve the above problems, in recent years, some researchers have tried to combine RFID technology with other sensor technologies, such as accelerometers, temperature sensors, attitude sensors, etc., to make up for the shortcomings of RFID. However, these improvements often only address a single problem and have not achieved a real breakthrough in technology. For example, although the introduction of other sensors can improve positioning accuracy, the complexity of signal processing and data fusion increases, which affects the real-time performance and response speed of the system. In addition, some systems rely on static algorithms and settings and cannot adapt to changing environments and animal behaviors, which makes these systems still have certain limitations in dynamically changing breeding environments. Furthermore, there is no effective means in the existing technology to perform behavioral pattern recognition and intelligent trajectory prediction. Most systems can only provide the real-time location of animals and lack the ability to deeply analyze and predict animal behavior patterns. Therefore, how to combine RFID with advanced intelligent algorithms to achieve comprehensive optimization of positioning accuracy, energy efficiency and behavior recognition has become a technical problem that needs to be solved in the field of animal tracking. Summary of the invention

[0005] The purpose of this invention is to propose an animal dynamic trajectory tracking method and system based on RFID. By combining regional dynamic tracking with intelligent grid algorithm, the positioning accuracy problem of traditional RFID system in dense areas is solved; through multi-dimensional signal reverse calculation and behavior pattern recognition technology, the problems caused by signal interference and group behavior complexity are solved. This technological innovation not only improves the accuracy and reliability of animal trajectory tracking, but also significantly improves the energy efficiency of the system, and promotes the widespread development of RFID animal tracking technology in practical applications.

[0006] In order to achieve the above object, a first aspect of the present invention provides an animal dynamic trajectory tracking method based on RFID, the method comprising the following steps:

[0007] S1. Acquire monitoring area data, detect animal behavior patterns in real time and analyze them in the monitoring area; the monitoring area includes animal position matrix, animal speed matrix, animal acceleration matrix and received signal strength;

[0008] S2. Divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce the signal interference function to optimize the signal layout and reduce mutual interference between different areas;

[0009] S3. Dynamically adjust the grid density according to the animal behavior patterns in different areas to meet the dynamically changing positioning needs; among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or slow activity areas;

[0010] S4. Based on the grid adjustment, the signal strength in the corresponding area is optimized to ensure that the signal coverage in different areas matches the positioning accuracy, and a signal interference optimization model is constructed to reduce the overlap and conflict between signals; for high-density activity areas, the signal strength is increased to ensure the accuracy of positioning; and for low-density animal activity areas, the signal strength is reduced to avoid signal waste;

[0011] S5. Based on the adjusted grid density and optimized signal strength, combined with the animal behavior pattern, the multi-dimensional signals are integrated into a comprehensive signal strength matrix for trajectory prediction through weighted fusion;

[0012] S6. Constructing a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal strength matrix to predict the future trajectory of the animal;

[0013] S7. Correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

[0014] Furthermore, the animal position matrix is ​​represented as X={x i |i=1,…,N}, where x i =(x i1 ,x i2 ,…,x iD ) represents the D-dimensional position coordinates of the ith animal at time t; the animal velocity matrix is ​​represented as V = {v i |i=1,…,N}, where v i =v i1 ,v i2 ,…,v iD ) represents the speed of the ith animal; the animal acceleration matrix is ​​represented by A = {a i |i=1,…,N}, where a i =(a i1 ,a i2 ,…,a iD ) represents the acceleration of the ith animal and the received signal strength is expressed as RSSI = {r i}, indicating the RFID signal strength received at the current location.

[0015] Furthermore, the real-time detection and analysis of animal behavior patterns in the monitoring area includes:

[0016] Based on the speed, acceleration, and position data of the animals, a classification model combining physical dynamics and data-driven methods is designed. The classification model uses a weighted mean-based classification strategy to predict the mode of each animal according to its behavioral dynamics, expressed as:

[0017]

[0018] Among them, P t is the proportion of behavioral patterns of all animals in the area at time t; N is the number of animals at the current moment; ψ(x i ,v i ,a i ) is the behavior pattern classification function, according to the spatial position x of the animal i , speed v i and acceleration a i , output the corresponding behavior pattern category.

[0019] Furthermore, the S2 specifically includes:

[0020] Based on the behavior pattern analysis results obtained by S1, the entire monitoring area is divided as follows:

[0021]

[0022] Among them, ρ j (t) is the grid density of region j, indicating the degree of subdivision of each grid unit in the region; N j is the number of animals in region j; α j is a regulating factor that controls the effect of grid density on the regional scale; σ(x i ,v i ,a i ,p(t)) is the function of signal strength and grid density optimization, according to the animal's position x i , speed v i , acceleration a i And the proportion of behavior patterns in region j p(t) dynamically generates the grid density; S j (t) is the signal strength of region j; β j and γ j are the regulation factors of signal intensity; p j (t is the proportion of behavior patterns in region j, based on the behavior pattern classification result calculated by S1;

[0023] The signal interference function D ij , expressed as:

[0024]

[0025] Among them, D ij is the interference degree between area i and area j, indicating the mutual influence between signals; x i and x j are the animal positions in area i and area j respectively; S k is the signal strength of region k.

[0026] Furthermore, the S3 is specifically expressed as:

[0027]

[0028] Among them, ρ j (t)' represents the grid density of the optimized region j at time t; α j N represents the adjustment factor of region j, which adjusts the grid density based on the behavior pattern of the region; j represents the number of animals in area j, which is used to measure the activity density of the area; η(x i ,v i ,a i ,p j (t)) represents the behavior pattern function, which describes the grid requirements of animals in different behavior patterns. The behavior pattern function η(x i ,v i ,a i ,p j (t)) The grid requirements are adjusted according to the animal's behavior patterns. When the animal is moving at high speed, the grid needs to be denser, while when it is stationary, the grid division is relatively sparse.

[0029] Furthermore, the S4 is specifically expressed as:

[0030]

[0031] Among them, S j (t)' represents the signal strength of the optimized region j at time t; β j represents a signal regulator that adjusts the signal strength based on the animal behavior pattern in the area; γ j represents the signal strength adjustment factor, taking into account the activity status of animals in the area; ρ j (t represents the grid density of region j at time t, which affects the distribution accuracy of the signal; the signal strength S after optimization j (t)' and the optimized grid density ρ j (t)' is inversely proportional;

[0032] The signal interference optimization model is as follows:

[0033]

[0034] Among them, D ij ' represents the interference degree between area i and area j; x i and x j represents the spatial position of animals in region i and region j respectively; S k represents the signal strength of region k; φ(p i (t),p j(t)) represents the behavior pattern function, which indicates the similarity of the behavior patterns between region i and region j.

[0035] Furthermore, it is characterized in that the comprehensive signal strength matrix is ​​expressed as:

[0036]

[0037] Among them, S jcom (t)) represents the comprehensive signal strength of region j at time t; f(p j (t)) represents the behavior mode adjustment function, according to the behavior mode p of region j j (t) dynamically adjusts the signal to reflect the signal response under different activity states; γ(t) represents the time correction factor, which takes into account the influence of time factors on signal strength;

[0038] Among them, through The signal strength and grid density are normalized so that the weights of different regions and signals can be adjusted consistently; the behavior mode adjustment function f(p j (t)) dynamically adjusts the signal response according to the animal's behavior; the time correction factor γ(t) simulates the time-varying characteristics of the signal in a dynamic environment according to the changes in the signal at different times.

[0039] Furthermore, the S6 specifically includes:

[0040] Calculate the predicted position x for each region pred (t+Δt):

[0041] x pred (t+Δt)=x current (t)+v(t)·Δt+F(t)·S com (t)

[0042] Among them, x pred (t+Δt) represents the animal position at the predicted time t+Δt; x current (t0 represents the animal position at the current time t; v9t) represents the animal's velocity vector at time t; F(t) represents the signal fusion matrix, which is used to map the fused signal strength to the position prediction; S com (t) represents the integrated signal strength;

[0043] Add a Bayesian correction term:

[0044] x pred (t+Δt)=x pred (t+Δt)+α·(S com 9t)-x pred (t))

[0045] Among them, α represents the Bayesian weight factor, which controls the balance between the fused prediction signal and the initial trajectory.

[0046] Furthermore, the S7 specifically includes:

[0047] Based on the predicted location x for each region pred 9t+Δt0, introduce the behavior pattern matrix P t As a correction term, the trajectory prediction is dynamically adjusted to improve the accuracy and precision of the trajectory; wherein the behavior pattern matrix P t It reflects the behavioral preferences of animals in different areas to avoid excessive prediction errors; the specific correction is expressed as:

[0048] x corrected 9t+Δt0=x pred 9t+Δt0+δx9t)

[0049] Among them, x corrected (t+Δt) represents the corrected trajectory prediction, x pred (t+Δt) represents the previous predicted trajectory; δx(t represents the trajectory correction, based on the animal's behavior pattern matrix P t Dynamic adjustment;

[0050] The trajectory correction value δx(t) is calculated according to the following formula:

[0051] δx(t)=P t ·G(t)

[0052] Among them, P t represents the behavior pattern matrix, which represents the animal behavior characteristics in different areas, G(t) represents the trajectory correction factor, which adjusts the trajectory based on environmental changes and movement status;

[0053] Among them, the behavior pattern matrix P t It is used to dynamically adjust the trajectory prediction and can automatically correct the trajectory when the animal's behavior changes; the trajectory correction factor G(t) allows the animal's trajectory prediction to be optimized under different environmental conditions.

[0054] In a second aspect of the present invention, an animal dynamic trajectory tracking system based on RFID is provided, the system comprising:

[0055] A regional data acquisition unit is used to acquire monitoring area data, detect animal behavior patterns in real time in the monitoring area and analyze them; the monitoring area includes an animal position matrix, an animal speed matrix, an animal acceleration matrix and a received signal strength;

[0056] The area division initialization unit is used to divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce the signal interference function to optimize the signal layout and reduce the mutual interference between different areas;

[0057] The grid density initialization unit is used to dynamically adjust the grid density according to the animal behavior patterns in different areas to cope with the dynamically changing positioning needs. Among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or slow activity areas.

[0058] The area division optimization unit is used to optimize the signal strength in the corresponding area based on the grid adjustment, ensure that the signal coverage and positioning accuracy in different areas match, and build a signal interference optimization model to reduce the overlap and conflict between signals; for high-density activity areas, the signal strength is improved to ensure the accuracy of positioning; and for low-density animal activity areas, the signal strength is reduced to avoid signal waste;

[0059] A grid density optimization unit, which is used to integrate multi-dimensional signals into a comprehensive signal strength matrix for trajectory prediction by weighted fusion based on the adjusted grid density and optimized signal strength in combination with animal behavior patterns;

[0060] An animal trajectory prediction unit, used to construct a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal strength matrix to predict the future trajectory of the animal;

[0061] The animal trajectory correction unit is used to correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

[0062] The beneficial technical effects of the present invention are at least as follows:

[0063] In view of the problems of insufficient positioning accuracy, energy efficiency waste and lack of behavior analysis capabilities in the prior art, the present invention combines multi-dimensional signal reverse calculation and behavior pattern recognition to fundamentally overcome these technical defects. The innovation of the present invention is that through dynamic area adjustment, signal optimization and intelligent behavior recognition, it not only improves the positioning accuracy of the RFID tracking system in complex environments, but also greatly reduces energy consumption, while enhancing the intelligence and adaptability of the system.

[0064] First of all, one of the key innovations of the present invention is the regional local dynamic tracking and adaptive grid algorithm. In view of the low positioning accuracy problem in the prior art in dense or complex environments with animal groups, the present invention dynamically divides the tracking area by real-time monitoring of the animal's behavior patterns, especially using higher grid density and frequency for areas where animals are active, thereby avoiding the accuracy loss caused by the uniform grid layout in traditional systems. In high-density areas, the system will automatically increase the number of RFID readers and increase the signal acquisition frequency to ensure positioning accuracy; while in low-density areas, the system will reduce the grid density, save energy, and improve the overall operation efficiency of the system. Through this adaptive grid adjustment mechanism, the present invention can accurately track the location of animals within a maximum range of 12 meters, while significantly reducing energy consumption.

[0065] Secondly, the present invention solves the problem that the existing technology cannot intelligently identify and adapt to animal behavior in complex environments by introducing multi-dimensional signal reverse calculation and behavior pattern recognition technology. By combining multi-dimensional data such as RFID signal strength, acceleration, and posture, the present invention uses a reverse calculation algorithm to accurately locate the real-time position of the animal, especially in the case of signal attenuation or occlusion, and can effectively correct the position error through data fusion. In addition, by identifying the behavior patterns of animals through machine learning algorithms, the system can intelligently identify abnormal behaviors of animals (such as leaving the group alone, sudden changes in speed, etc.), and automatically adjust the tracking strategy, further improving the tracking accuracy of the system in the case of complex or dynamically changing group behaviors. Real-time recognition and prediction of behavior patterns enable the system to predict the trajectory of animals in advance, reduce positioning errors, and improve the intelligence level of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0067] Figure 1 This is a flow chart of the animal dynamic trajectory tracking method based on RFID of the present invention. DETAILED DESCRIPTION

[0068] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0069] like Figure 1 As shown, an RFID-based animal dynamic trajectory tracking method provided by an embodiment of the present invention includes:

[0070] S1. Acquire monitoring area data, detect animal behavior patterns in real time in the monitoring area and analyze them; the monitoring area includes an animal position matrix, an animal speed matrix, an animal acceleration matrix and a received signal strength.

[0071] Specifically, the monitoring area data is obtained, including: animal position matrix X = {x i |i=1,…,N}, where x i =x i1 ,x i2 ,…,x iD ) represents the D-dimensional position coordinates of the ith animal at time t; the animal velocity matrix V = {v i |i=1,…,N}, where v i =(v i1 ,v i2 ,…,v iD ) represents the speed of the ith animal; the animal acceleration matrix A = {a i |i=1,…,N}, where a i =a i1 ,a i2 ,…,a iD ) represents the acceleration of the ith animal; received signal strength RSSI = {r i}, indicating the RFID signal strength received at the current location.

[0072] Furthermore, the behavior patterns of animals in space usually include the following: stillness, slow movement, high-speed movement, group gathering, etc. In order to identify these behavior patterns, the present invention designs a classification model that combines physical dynamics and data-driven methods based on the speed, acceleration, and position data of the animals. The present invention uses a classification strategy based on weighted mean to predict the mode of each animal according to its behavioral dynamics:

[0073]

[0074] Among them, P t is the proportion of behavioral patterns of all animals in the area at time t. N is the number of animals at the current moment. ψ(x i ,v i ,a i ) is the behavior pattern classification function. According to the spatial position x of the animal i , speed v i and acceleration a i , output the corresponding behavior pattern category (such as stationary, low speed, high speed, group gathering, etc.).

[0075] Understandably, animal position x iis the specific position of the animal in two-dimensional or three-dimensional space. Animal speed v i and acceleration a i Used to indicate an animal's state of motion, especially whether it is stationary or moving rapidly.

[0076] By classifying the behavior of each animal at the current moment, we can get the proportion of different behavior patterns in the area, thus providing a basis for subsequent regional divisions.

[0077] S2. Divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce a signal interference function to optimize the signal layout and reduce mutual interference between different areas;

[0078] Specifically, based on the behavior pattern analysis results obtained in the first step, the entire monitoring area is divided. The division of the area not only depends on the geographical location, but also needs to take into account the changes in animal behavior patterns. Specifically, animals have different signal requirements under different behavior patterns. Therefore, adaptive adjustment of signal strength and grid density is crucial. The present invention designs the following optimization formula to achieve this:

[0079]

[0080] S j (t) = β j ·(γ j ·p j (t))

[0081] Among them, ρ j (t) is the grid density of region j, which indicates the degree of subdivision of each grid unit in the region. j is the number of animals in region j. j is a regulating factor that controls the effect of grid density on the region size. i ,v i ,a i ,p(t)) is the function of signal strength and grid density optimization, according to the animal's position x i , speed v i , acceleration a i The grid density is dynamically generated based on the proportion of behavior patterns in region j, p(t). j (t) is the signal strength of region j. j and γ j are the adjustment factors of signal intensity. j (t is the proportion of behavior patterns in region j, based on the behavior pattern classification results calculated in the first step.

[0082] It can be understood that the grid density ρj (t) represents the fineness of the grid division in area j. The denser the grid, the higher the positioning accuracy, but the greater the computing resources and energy consumption. Signal strength S j (t) represents the signal strength in area j. The stronger the signal, the higher the positioning accuracy.

[0083] This step intelligently adjusts the grid density and signal strength of each area according to the proportion of different animal behavior patterns to ensure the best balance between signal coverage and positioning accuracy.

[0084] Furthermore, in order to avoid signal interference between different areas, this step also introduces a signal interference optimization item to ensure that the signal strengths between adjacent areas do not interfere with each other. ij To optimize signal layout and reduce mutual interference between different areas:

[0085]

[0086] Among them, D ij is the interference degree between area i and area j, indicating the mutual influence between signals. i and x j are the animal positions in region i and region j respectively. k is the signal strength of region k.

[0087] Understandably, the signal interference D ij It is used to quantify the degree of signal interference between area i and area j. The greater the interference, the stronger the interference between signals and the worse the positioning accuracy. ij ,The present invention can optimize signal distribution between different areas, reduce signal overlap, and improve the positioning accuracy of the entire system.

[0088] S3. Dynamically adjust the grid density according to the animal behavior patterns in different areas to cope with dynamically changing positioning needs; among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or low-speed activity areas.

[0089] Specifically, according to the animal behavior patterns in different areas, the density of the grid needs to be flexibly adjusted to cope with the dynamically changing positioning needs. High-density animal activity areas (such as group gatherings or high-speed movement areas) require finer grid divisions, while static or low-speed activity areas can use coarser grid divisions.

[0090] Among them, the grid density ρ j The adjustment of (t)' depends on the following formula:

[0091]

[0092] Among them, ρ j (t)' represents the grid density of region j at time t. j N represents the adjustment factor of region j, which adjusts the grid density based on the behavior pattern of the region. j Represents the number of animals in area j, which is used to measure the activity density of the area. η(x i ,v i ,a i ,p j (t)) represents the behavior mode function, which describes the grid requirements of animals in different behavior modes. It depends on the animal's position x i , speed v i , acceleration a i And the proportion of behavior patterns in region j, p j (t) is dynamically calculated. η(x i ,v i ,a i ,p j (t)) The grid requirements are adjusted according to the animal's behavior pattern. When moving at high speed, the grid needs to be denser, while when stationary, the grid division is sparser. This adjustment takes into account the positioning accuracy requirements of different behavior patterns, thereby achieving dynamic adaptation.

[0093] S4. Based on the grid adjustment, the signal strength in the corresponding area is optimized to ensure that the signal coverage in different areas matches the positioning accuracy, and a signal interference optimization model is constructed to reduce the overlap and conflict between signals. For high-density activity areas, the signal strength is improved to ensure the accuracy of positioning; for low-density animal activity areas, the signal strength is reduced to avoid signal waste.

[0094] Specifically, based on the grid density adjustment, the next step is to optimize the signal strength in the area to ensure that the signal coverage in different areas matches the positioning accuracy. For high-density activity areas, the signal strength should be increased to ensure positioning accuracy; while for low-density areas, the signal strength can be reduced to avoid signal waste. The optimization formula for signal strength is as follows:

[0095]

[0096] Among them, S j (t) represents the signal strength of region j at time t. j Represents a signal modulator that adjusts the signal strength based on the animal behavior pattern in the area. j Represents the signal strength adjustment factor, taking into account the activity status of animals in the area (such as group density and movement speed). j (t) represents the grid density of region j at time t, which affects the distribution accuracy of the signal.

[0097] Signal strength S j (t) and grid density ρ j (t). The denser the grid, the lower the required signal strength, because a fine grid can provide more accurate positioning without excessive signal interference. j and γ j Dynamic adjustment can refine the signal distribution in the activity area according to the behavior patterns of animal groups. For example, when the animal activities are dense, the signal strength is increased to ensure stable positioning; in the area with less activity, the signal strength is reduced to avoid waste.

[0098] Furthermore, as the grids and signals of multiple areas are optimized, interference may occur between areas, affecting the stability and positioning accuracy of the overall system. This step optimizes the interference between areas, reduces overlap and conflict between signals, and improves the stability and positioning accuracy of the system. The signal interference optimization model is as follows:

[0099]

[0100] Among them, D ij represents the interference between area i and area j. i and x j S represents the spatial position of animals in region i and region j respectively. k represents the signal strength of region k. φ(p i (t),p j (t)) represents the behavior pattern function, which indicates the similarity of the behavior patterns of region i and region j. ij It reflects the degree of interference between area i and area j. The higher the interference, the more serious the signal overlap, which may affect the positioning accuracy. By adjusting the signal strength and area division, the interference can be effectively reduced. i (t),p j (t)) to reflect the similarity of the behavior patterns between regions, which can further optimize the signal distribution. Regions with similar behavior patterns may produce large signal overlaps, so it is necessary to optimize the signal strength distribution between these regions.

[0101] S5. Based on the adjusted grid density and optimized signal strength, combined with animal behavior patterns, multi-dimensional signals are integrated into a comprehensive signal strength matrix for trajectory prediction through weighted fusion.

[0102] Specifically, in the previous steps, the present invention has obtained the optimized signal strength S j (t)' and grid density ρ j (t)', and the behavior pattern matrix P tIn this step, the present invention first integrates the multi-dimensional signals into a comprehensive signal strength matrix S that can be used for trajectory prediction by weighted fusion. jcom (t). This step is the core of multi-dimensional signal fusion.

[0103] The calculation formula for setting the comprehensive signal strength is as follows:

[0104]

[0105] Among them, S jcom (t) represents the comprehensive signal strength of region j at time t. j (t) represents the signal strength of region j at time t. j (t) represents the grid density of region j at time t. f(p j (t)) represents the behavior mode adjustment function, according to the behavior mode p of region j j (t) dynamically adjusts the signal to reflect the signal response under different activity states. γ(t) represents the time correction factor, which takes into account the influence of time factors on signal strength, such as signal attenuation or change.

[0106] Understandably, through The present invention normalizes the signal strength and grid density so that the weights of different regions and signals can be adjusted consistently. j (t)) dynamically adjusts the signal response based on the animal's behavior. For example, when the animal is moving quickly, the signal response may need to be enhanced to improve positioning accuracy. The time correction factor γ(t) takes into account the changes in the signal at different times and can more accurately simulate the time-varying characteristics of the signal, especially in dynamic environments.

[0107] S6. Construct a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal intensity matrix to predict the future trajectory of the animal.

[0108] Specifically, after weighted fusion of multi-dimensional signals, the present invention obtains a comprehensive signal strength matrix S jcom (t), next, the present invention needs to predict the future trajectory of the animal based on this fusion signal. In order to predict the trajectory more accurately, the present invention combines the physical motion model with the signal strength and grid density information, and proposes a prediction model combining Kalman filtering and Bayesian signal adjustment.

[0109] First, calculate the predicted position x of each region pred (t+Δt):

[0110] x pred (t+Δt)=x current(t)+v(t)·Δt+F(t)·S com (t)

[0111] Among them, x pred (t+Δt) represents the animal's position at the predicted time t+Δt. current (t) represents the animal's position at the current time t. v(t) represents the animal's velocity vector at time t. F(t) represents the signal fusion matrix, which is used to map the fused signal strength to the position prediction. S com (t) represents the integrated signal strength.

[0112] The prediction model combines the physical model (displacement and velocity) with the influence of signal strength to achieve dynamic and accurate trajectory prediction. The signal fusion matrix F(t) combines the comprehensive signal strength with the location information, allowing the system to optimize the prediction results based on the signal strength in different areas.

[0113] In order to further improve the accuracy of trajectory prediction, a Bayesian correction term is added:

[0114] x pred (t+Δt)=x pred (t+Δt)+α·(S com (t0-x pred (t))

[0115] Among them, α represents the Bayesian weight factor, which controls the balance between the fused prediction signal and the initial trajectory. After the Bayesian correction term is introduced, the system adjusts the trajectory prediction at different time points and optimizes it by combining historical data with the current signal strength. By adjusting α, the present invention can control the sensitivity of the prediction and further improve the prediction accuracy.

[0116] S7. Correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

[0117] Specifically, the core goal of this step is to introduce the behavior pattern matrix P t As a correction item, the trajectory prediction is dynamically adjusted to improve the accuracy and precision of the trajectory. The behavior pattern matrix reflects the behavioral preferences of animals in different areas. This feature needs to be considered in trajectory prediction to avoid excessive prediction errors.

[0118] Predict x based on the previous trajectory pred (t+Δt0, a new correction method is proposed:

[0119] x corrected (t+Δt)=x pred (t+Δt0+δx(t)

[0120] Among them, x corrected (t+Δt) represents the corrected trajectory prediction. pred (t+Δt) represents the previous predicted trajectory. δx(t represents the trajectory correction, based on the animal's behavior pattern matrix P t Dynamic adjustment.

[0121] The correction value δx(t) is calculated according to the following formula:

[0122] δx(t)=P t ·G(t)

[0123] Among them, P t represents the behavior pattern matrix, which represents the animal behavior characteristics in different areas. G(t) represents the trajectory correction factor, which adjusts the trajectory based on environmental changes and motion status. Behavior pattern matrix P t It is used to dynamically adjust trajectory predictions and automatically correct the trajectory when the animal's behavior changes. The trajectory correction factor G(t) allows further optimization of the animal's trajectory prediction under different environmental conditions. The introduction of the correction term δx(t) ensures the model's sensitivity to dynamic behavior patterns and reduces prediction errors caused by behavioral changes.

[0124] The embodiment of the present invention further provides an animal dynamic trajectory tracking system based on RFID, the system comprising:

[0125] A regional data acquisition unit is used to acquire monitoring area data, detect animal behavior patterns in real time in the monitoring area and analyze them; the monitoring area includes an animal position matrix, an animal speed matrix, an animal acceleration matrix and a received signal strength;

[0126] The area division initialization unit is used to divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce the signal interference function to optimize the signal layout and reduce the mutual interference between different areas;

[0127] The grid density initialization unit is used to dynamically adjust the grid density according to the animal behavior patterns in different areas to cope with the dynamically changing positioning needs. Among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or slow activity areas.

[0128] The area division optimization unit is used to optimize the signal strength in the corresponding area based on the grid adjustment, ensure that the signal coverage and positioning accuracy in different areas match, and build a signal interference optimization model to reduce the overlap and conflict between signals; for high-density activity areas, the signal strength is improved to ensure the accuracy of positioning; and for low-density animal activity areas, the signal strength is reduced to avoid signal waste;

[0129] A grid density optimization unit, which is used to integrate multi-dimensional signals into a comprehensive signal strength matrix for trajectory prediction by weighted fusion based on the adjusted grid density and optimized signal strength in combination with animal behavior patterns;

[0130] An animal trajectory prediction unit, used to construct a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal strength matrix to predict the future trajectory of the animal;

[0131] The animal trajectory correction unit is used to correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

[0132] In summary, the present invention solves the positioning accuracy problem of traditional RFID systems in dense areas by combining regional dynamic tracking with intelligent grid algorithms; and solves the problems caused by signal interference and group behavior complexity by multi-dimensional signal reverse calculation and behavior pattern recognition technology. This technological innovation not only improves the accuracy and reliability of animal trajectory tracking, but also significantly improves the energy efficiency of the system, promoting the widespread development of RFID animal tracking technology in practical applications.

[0133] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0134] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a division of logical functions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be electrical, mechanical or other forms.

[0135] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program codes.

[0136] Although embodiments of the present invention have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. The animal dynamic trajectory tracking method based on RFID is characterized by: The method comprises the following steps: S1. Acquire monitoring area data, detect animal behavior patterns in real time and analyze them in the monitoring area; the monitoring area includes animal position matrix, animal speed matrix, animal acceleration matrix and received signal strength; S2. Divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce the signal interference function to optimize the signal layout and reduce mutual interference between different areas; S3. Dynamically adjust the grid density according to the animal behavior patterns in different areas to meet the dynamically changing positioning needs; among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or slow activity areas; S4. Based on the grid adjustment, the signal strength in the corresponding area is optimized to ensure that the signal coverage in different areas matches the positioning accuracy, and a signal interference optimization model is constructed to reduce the overlap and conflict between signals; for high-density activity areas, the signal strength is increased to ensure the accuracy of positioning; and for low-density animal activity areas, the signal strength is reduced to avoid signal waste; S5. Based on the adjusted grid density and optimized signal strength, combined with the animal behavior pattern, the multi-dimensional signals are integrated into a comprehensive signal strength matrix for trajectory prediction through weighted fusion; S6. Constructing a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal strength matrix to predict the future trajectory of the animal; S7. Correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

2. The animal dynamic trajectory tracking method based on RFID according to claim 1, characterized in that: The animal position matrix is ​​expressed as X={xi|i=1,…,N}, where xi=(xi1,xi2,…,xiD) represents the D-dimensional position coordinates of the ith animal at time t; the animal velocity matrix is ​​expressed as V={vi|i=1,…,N}, where vi=(vi1,vi2,…,viD) represents the velocity of the ith animal; the animal acceleration matrix is ​​expressed as A={ai|i=1,…,N}, where ai=(ai1,ai2,…,aiD) represents the acceleration of the ith animal and the received signal strength is expressed as RSSI={ri}, which represents the RFID signal strength received at the current position.

3. The animal dynamic trajectory tracking method based on RFID according to claim 2, characterized in that: The real-time detection and analysis of animal behavior patterns in the monitoring area includes: Based on the speed, acceleration, and position data of the animals, a classification model combining physical dynamics and data-driven methods is designed. The classification model uses a weighted mean-based classification strategy to predict the mode of each animal according to its behavioral dynamics, expressed as: Among them, Pt is the proportion of behavioral patterns of all animals in the area at time t; N is the number of animals at the current moment; ψ(xi,vi,ai) is the behavior pattern classification function, which outputs the corresponding behavior pattern category according to the animal's spatial position xi, velocity vi and acceleration ai.

4. The animal dynamic trajectory tracking method based on RFID according to claim 3 is characterized in that: The S2 specifically includes: Based on the behavior pattern analysis results obtained by S1, the entire monitoring area is divided as follows: Sj(t)=βj·(γj·pj(t)) Among them, ρj(t) is the grid density of region j, which indicates the degree of subdivision of each grid unit in the region; Nj is the number of animals in region j; αj is an adjustment factor that controls the effect of grid density on regional scale; σ(xi,vi,ai,p(t)) is a function of signal strength and grid density optimization, which dynamically generates grid density according to the animal's position xi, speed vi, acceleration ai and the proportion of behavioral patterns in region j p(t); Sj(t) is the signal strength of region j; βj and γj are adjustment factors for signal strength respectively; pj(t) is the proportion of behavioral patterns in region j, based on the behavioral pattern classification result calculated by S1; The signal interference function Dij is expressed as: Among them, Dij is the interference degree between area i and area j, indicating the mutual influence between signals; xi and xj are the animal positions in area i and area j respectively; Sk is the signal strength in area k.

5. The animal dynamic trajectory tracking method based on RFID according to claim 4, characterized in that: The S3 is specifically expressed as: Where ρj(t)' represents the mesh density of the optimized region j at time t; αj represents the adjustment factor of region j, which adjusts the mesh density based on the behavior pattern of the region; N j represents the number of animals in area j, which is used to measure the activity density of the area; η(x i ,v i ,a i ,p j (t)) represents the behavior pattern function, which describes the grid requirements of animals in different behavior patterns. The behavior pattern function η(x i ,v i ,a i ,p j (t)) The grid requirements are adjusted according to the animal's behavior patterns. When the animal is moving at high speed, the grid needs to be denser, while when the animal is stationary, the grid division is relatively sparse.

6. The animal dynamic trajectory tracking method based on RFID according to claim 4, characterized in that: The S4 is specifically expressed as: Among them, S j (t)' represents the signal strength of the optimized region j at time t; β j represents a signal regulator that adjusts the signal strength based on the animal behavior pattern in the area; γ j represents the signal strength adjustment factor, taking into account the activity status of animals in the area; ρ j (t) represents the grid density of region j at time t, which affects the distribution accuracy of the signal; the optimized signal strength S j (t)' and the optimized grid density ρ j (t)' is inversely proportional; The signal interference optimization model is as follows: Among them, D ij ' represents the interference degree between area i and area j; x i and x j represents the spatial position of animals in region i and region j respectively; S k represents the signal strength of region k; φ(p i (t),p j (t)) represents the behavior pattern function, which indicates the similarity of the behavior patterns between region i and region j.

7. The animal dynamic trajectory tracking method based on RFID according to any one of claims 5-6, characterized in that: The comprehensive signal strength matrix is ​​expressed as: in, represents the comprehensive signal strength of region j at time t; f(p j (t)) represents the behavior mode adjustment function, according to the behavior mode p of region j j (t) dynamically adjusts the signal to reflect the signal response under different activity states; γ(t) represents the time correction factor, which takes into account the influence of time factors on signal strength; Among them, through The signal strength and grid density are normalized so that the weights of different regions and signals can be adjusted consistently; the behavior mode adjustment function f(p j (t)) dynamically adjusts the signal response according to the animal's behavior; the time correction factor γ(t) simulates the time-varying characteristics of the signal in a dynamic environment according to the changes in the signal at different times.

8. The animal dynamic trajectory tracking method based on RFID according to claim 7, characterized in that: The S6 specifically includes: Calculate the predicted position x for each region pred (t+Δt): x pred (t+Δt)=x current (t)+v(t)·Δt+F(t)·S com (t) Among them, x pred (t+Δt) represents the animal position at the predicted time t+Δt; x current (t) represents the animal's position at the current time t; v(t) represents the animal's velocity vector at time t; F(t) represents the signal fusion matrix, which is used to map the fused signal strength to the position prediction; S com (t) represents the integrated signal strength; Add a Bayesian correction term: x pred (t+Δt)=x pred (t+Δt)+α·(S com (t)-x pred (t)) Among them, α represents the Bayesian weight factor, which controls the balance between the fused prediction signal and the initial trajectory.

9. The animal dynamic trajectory tracking method based on RFID according to claim 8, characterized in that: The S7 specifically includes: Based on the predicted location x for each region pred (t+Δt), introduce the behavior pattern matrix P t As a correction term, the accuracy and precision of the trajectory are improved by dynamically adjusting the trajectory prediction; wherein the behavior pattern matrix P t It reflects the behavioral preferences of animals in different areas to avoid excessive prediction errors; the specific correction is expressed as: x corrected (t+Δt)=x pred (t+Δt)+δx(t) Among them, x corrected (t+Δt) represents the corrected trajectory prediction, x pred (t+Δt) represents the previous predicted trajectory; δx(t) represents the trajectory correction, based on the animal's behavior pattern matrix P t Dynamic adjustment; The trajectory correction value δx(t) is calculated according to the following formula: δx(t)=P t ·G(t) Among them, P t represents the behavior pattern matrix, which represents the animal behavior characteristics in different areas, G(t) represents the trajectory correction factor, which adjusts the trajectory based on environmental changes and movement status; Among them, the behavior pattern matrix P t It is used to dynamically adjust the trajectory prediction and can automatically correct the trajectory when the animal's behavior changes; the trajectory correction factor G(t) allows the animal's trajectory prediction to be optimized under different environmental conditions.

10. The animal dynamic trajectory tracking system based on RFID is characterized by: The system comprises: A regional data acquisition unit is used to acquire monitoring area data, detect animal behavior patterns in real time in the monitoring area and analyze them; the monitoring area includes an animal position matrix, an animal speed matrix, an animal acceleration matrix and a received signal strength; The area division initialization unit is used to divide the entire monitoring area based on the analysis results, adaptively adjust the signal strength and grid density of the area according to the division results, and introduce the signal interference function to optimize the signal layout and reduce the mutual interference between different areas; The grid density initialization unit is used to dynamically adjust the grid density according to the animal behavior patterns in different areas to cope with the dynamically changing positioning needs. Among them, fine grid division is required for high-density animal activity areas, while coarse grid division can be used for static or slow activity areas. The area division optimization unit is used to optimize the signal strength in the corresponding area based on the grid adjustment, ensure that the signal coverage and positioning accuracy in different areas match, and build a signal interference optimization model to reduce the overlap and conflict between signals; for high-density activity areas, the signal strength is improved to ensure the accuracy of positioning; and for low-density animal activity areas, the signal strength is reduced to avoid signal waste; A grid density optimization unit, which is used to integrate multi-dimensional signals into a comprehensive signal strength matrix for trajectory prediction by weighted fusion based on the adjusted grid density and optimized signal strength in combination with animal behavior patterns; An animal trajectory prediction unit, used to construct a prediction model combining Kalman filtering and Bayesian signal adjustment according to the comprehensive signal strength matrix to predict the future trajectory of the animal; The animal trajectory correction unit is used to correct the predicted future trajectory of the animal based on the animal's behavior pattern, and improve the accuracy and precision of the trajectory by dynamically adjusting the trajectory prediction.

Images