Livestock positioning control device based on GNSS and Bluetooth Mesh networking

Abstract

According to the livestock positioning control device based on GNSS and Bluetooth Mesh networking, the manual dependence is greatly reduced, the informatization level is high, and long-time low-cost grazing is achieved. The Bluetooth tag converts the communication signal intensity with the surrounding wearable driving necklace into the distance with the surrounding necklace, completes the calculation of the position information of the Bluetooth tag according to the GNSS positioning information of the necklace, and transmits the position information to the nearest necklace through the Mesh network; after receiving the position information of the surrounding labels, the necklace transmits the self positioning information and the information of each label to a monitoring platform through a 4G communication technology, and the position of the livestock is monitored in real time through the monitoring platform; the necklace judges whether the livestock exceeds the buffer zone or leaves the boundary or not according to the position information of the necklace and the vertex coordinates of the virtual fence, if the livestock exceeds the buffer zone, sound stimulation is conducted on the livestock, if the livestock exceeds the boundary of the fence, electric shock stimulation is conducted on the livestock, and the aim of controlling the whole livestock group is achieved in the mode that the livestock with the leading force is controlled in the fence.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent grazing, and in particular relates to a livestock positioning control device based on GNSS and Bluetooth Mesh networking. Background Art

[0002] Currently, manual grazing is generally used in the livestock industry, which requires herders to participate in the entire grazing process. Due to the large grazing area and large number of livestock, it is difficult to look after the livestock during the grazing process, which may lead to the loss of livestock. Therefore, the traditional grazing method has problems such as high dependence on manual labor and low level of informatization, making it difficult to achieve long-term and low-cost grazing. Summary of the invention

[0003] In order to overcome the defects of the prior art, the technical problem to be solved by the present invention is to provide a livestock positioning control device based on GNSS and Bluetooth Mesh networking, which greatly reduces manual dependence, has a high level of informatization, and can achieve long-term and low-cost grazing.

[0004] The technical solution of the present invention is: a livestock positioning control device based on GNSS and Bluetooth Mesh networking, the device comprises a wearable driving collar, a Bluetooth tag, and a monitoring platform, and data is transmitted between the wearable driving collar and the Bluetooth tag via Bluetooth Mesh networking;

[0005] The Bluetooth tag converts the communication signal strength with the surrounding wearable repellent collars into the distance to the surrounding wearable repellent collars, calculates its own location information based on the GNSS positioning information of the wearable repellent collars, and transmits the location information to the nearest wearable repellent collar through the Mesh network;

[0006] After the wearable driving collar receives the location information of the surrounding Bluetooth tags, it transmits its own GNSS positioning information and the information of each Bluetooth tag to the monitoring platform through 4G communication technology. The herders can monitor the location of the livestock in real time through the monitoring platform.

[0007] The wearable drive collar determines whether the livestock exceeds the rear fence boundary buffer zone or leaves the fence boundary based on its own GNSS position information and the coordinates of the virtual fence vertices. If it exceeds the fence boundary buffer zone, the livestock will be stimulated by sound; if it exceeds the fence boundary, the livestock will be stimulated by electric shock. The goal of controlling the entire herd is achieved by controlling the leadership livestock within the fence.

[0008] The Bluetooth tag of the present invention converts the communication signal strength with the surrounding wearable driving collars into the distance with the surrounding wearable driving collars, completes the calculation of its own position information according to the GNSS positioning information of the wearable driving collar, and transmits the position information to the nearest wearable driving collar through the Mesh network; after receiving the position information of the surrounding Bluetooth tags, the wearable driving collar transmits its own GNSS positioning information and the information of each Bluetooth tag to the monitoring platform through the 4G communication technology, and the herders monitor the livestock position in real time through the monitoring platform; the wearable driving collar determines whether the livestock exceeds the rear fence boundary buffer zone or leaves the fence boundary according to its own GNSS position information and the coordinates of the virtual fence vertices, and if it exceeds the fence boundary buffer zone, the livestock is stimulated by sound, and if it exceeds the fence boundary, the livestock is stimulated by electric shock, and the goal of controlling the entire herd is achieved by controlling the leadership livestock within the fence; therefore, the dependence on manual labor is greatly reduced, the level of informatization is high, and long-term and low-cost grazing can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 This is a schematic diagram of the wearable repellent collar module.

[0010] Figure 2 Workflow diagram for wearable repellent collar.

[0011] Figure 3 This is a flowchart of the Bluetooth tag workflow.

[0012] Figure 4 Bluetooth tag positioning method. DETAILED DESCRIPTION

[0013] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0014] In order to make the description of the present disclosure more detailed and complete, the following is an illustrative description of the implementation and specific embodiments of the present invention; however, this is not the only form of implementing or using the specific embodiments of the present invention. The implementation covers the features of multiple specific embodiments and the device steps and sequences used to construct and operate these specific embodiments. However, other specific embodiments can also be used to achieve the same or equivalent functions and step sequences.

[0015] This livestock positioning and control device based on GNSS and Bluetooth Mesh networking includes a wearable driving collar, a Bluetooth tag, and a monitoring platform. The wearable driving collar and the Bluetooth tag transmit data via Bluetooth Mesh networking.

[0016] The Bluetooth tag converts the communication signal strength with the surrounding wearable repellent collars into the distance to the surrounding wearable repellent collars, calculates its own location information based on the GNSS positioning information of the wearable repellent collars, and transmits the location information to the nearest wearable repellent collar through the Mesh network;

[0017] After the wearable driving collar receives the location information of the surrounding Bluetooth tags, it transmits its own GNSS positioning information and the information of each Bluetooth tag to the monitoring platform through 4G communication technology. The herders can monitor the location of the livestock in real time through the monitoring platform.

[0018] The wearable drive collar determines whether the livestock exceeds the rear fence boundary buffer zone or leaves the fence boundary based on its own GNSS position information and the coordinates of the virtual fence vertices. If it exceeds the fence boundary buffer zone, the livestock will be stimulated by sound; if it exceeds the fence boundary, the livestock will be stimulated by electric shock. The goal of controlling the entire herd is achieved by controlling the leadership livestock within the fence.

[0019] The Bluetooth tag of the present invention converts the communication signal strength with the surrounding wearable driving collars into the distance with the surrounding wearable driving collars, completes the calculation of its own position information according to the GNSS positioning information of the wearable driving collar, and transmits the position information to the nearest wearable driving collar through the Mesh network; after receiving the position information of the surrounding Bluetooth tags, the wearable driving collar transmits its own GNSS positioning information and the information of each Bluetooth tag to the monitoring platform through the 4G communication technology, and the herders monitor the livestock position in real time through the monitoring platform; the wearable driving collar determines whether the livestock exceeds the rear fence boundary buffer zone or leaves the fence boundary according to its own GNSS position information and the coordinates of the virtual fence vertices, and if it exceeds the fence boundary buffer zone, the livestock is stimulated by sound, and if it exceeds the fence boundary, the livestock is stimulated by electric shock, and the goal of controlling the entire herd is achieved by controlling the leadership livestock within the fence; therefore, the dependence on manual labor is greatly reduced, the level of informatization is high, and long-term and low-cost grazing can be achieved.

[0020] Preferably, the wearable driving collar includes: a main control computing module, a power management module, a storage module, a positioning transmission module, a motion information collection module and a control stimulation module; the main control computing module controls the peripheral logic, data analysis and processing, and Bluetooth Mesh networking; the power management module provides corresponding voltages to each module; the storage module stores fence information, location data that has not been sent in time, and stores different types of sounds required for sound stimulation; the positioning transmission module acquires the livestock GNSS location information and communicates with the monitoring platform; the motion information collection module monitors the movement status of the livestock in real time; the control stimulation module emits sound and electric shock stimulation to confine the livestock within the set fence boundary.

[0021] Preferably, the method for determining whether livestock has exceeded the fence boundary is:

[0022] The fence boundary has n coordinate points P1, P2, ..., P n , P i , the longitude and latitude are lon i With lat i , n coordinate points are sent by the monitoring platform, and at the same time the monitoring platform sends a coordinate point center inside the fence, with longitude and latitude lon c With lat c ,

[0023] Convert longitude and latitude to planar coordinates using the Mercator projection:

[0024]

[0025] Where i = 1, 2, ..., n, the radius of the earth R = 6378137 m,

[0026] In the plane coordinate system, a straight line is determined based on two points, and n boundary lines are calculated, l1, l2, ..., l n ,

[0027] Among them, l i By P i With P i+1 Calculated, i = 1, 2, ..., n-1; l n By P n Calculated with P1,

[0028] l i :y i =a i × x i +b i (3)

[0029] a i With b i is the slope and intercept of the i-th boundary line,

[0030] Calculation and judgment i , i=1,2,...,n,

[0031]

[0032] The judge' is calculated based on the GNSS position information of the wearable repellent collar i and judgement i If they match, the livestock is inside the fence; otherwise, it is outside the fence boundary.

[0033] Preferably, the main control computing module selects a low-power MCU. In order to increase the battery life of the collar, two modes, working mode and standby mode, are designed. The main control computing module is responsible for switching the collar working mode. At the same time, the main control module reasonably completes the adjustment of the collar working interval according to the motion information collection data, completes the update of the fence boundary and the judgment of whether it exceeds the fence according to the information of the positioning transmission module, and the main control computing module also completes the Bluetooth Mesh networking and broadcasts information in the Mesh network through the Bluetooth chip;

[0034] The power management module includes lithium battery, solar panel, voltage conversion circuit, and charging circuit;

[0035] The storage module includes SRAM, Flash and SD card. SRAM is used to store data in working mode, Flash is used to store non-lost data in program and standby mode, and SD card is used to store audio during sound stimulation.

[0036] The positioning transmission module includes GNSS and 4G related circuits and antennas. The main control calculation module reads the location information of the positioning transmission module and transmits data with the monitoring platform through 4G. The motion information acquisition module uses sensors to collect real-time motion information, and the main control calculation module uses this information to adjust the collar working time.

[0037] The control stimulation module includes sound stimulation and electric shock stimulation. Sound stimulation relies on the voice chip to read the audio in the SD card, and then plays the sound through the power amplifier and speaker; electric shock stimulation uses the transformer coil to achieve the purpose of boosting, and uses an optocoupler relay to control the generation of high voltage to improve the safety of the collar.

[0038] Preferably, the Bluetooth tag includes a main control Bluetooth module and a power management module;

[0039] The main control Bluetooth module uses the Bluetooth chip to implement functions, which are used for the workflow of Bluetooth positioning tags, completing Bluetooth Mesh networking and receiving information in the Mesh network, RSSI data collection and processing and calculation of its own location information, as well as data transmission with other devices in the Mesh network;

[0040] The power management module includes lithium batteries, solar panels, voltage conversion circuits, and charging circuits. Lithium batteries with high energy density and support for charging and discharging are selected to power the Bluetooth tags, and the corresponding DC-DC

[0041] The conversion circuit provides power for each chip in the tag, and the Bluetooth tag is charged by solar energy.

[0042] Preferably, the positioning method of the Bluetooth tag based on RSSI data in the Mesh network includes the following steps:

[0043] (1) When the wearable repellent collar is in working mode, it broadcasts heartbeat packet data every second, the Bluetooth tag receives the information, collects RSSI data, and uses the unscented Kalman filter (UKF) to preprocess the data to reduce the impact of noise;

[0044] (2) Based on the relationship between RSSI data and distance, an RSSI ranging model is established:

[0045] RSSI k =B+Nlg(d k ) (5)

[0046] where d k is the distance, in meters, RSSI k is the corrected signal strength at time k, in dBm, where B and N are constants;

[0047] (3) Convert the GNSS position information into an unconstrained optimization problem, the function form is:

[0048]

[0049] Among them, y is the coordinate vector in the vertical coordinate form of the Bluetooth tag to be calculated, y m is the coordinate vector in the ordinate form of the mth wearable repellent collar obtained by the Bluetooth tag, M is the number of wearable repellent collar information that the Bluetooth tag can obtain, ω m is the weight of the mth wearable repellent collar information, d m is the distance between the Bluetooth tag and the mth wearable repellent collar (at the latest moment).

[0050] Preferably, in step (1), the state quantity set when applying the unscented Kalman filter UKF is:

[0051]

[0052] RSSI k is the RSSI data at time k, ΔRSSI k is the difference between time k and time k-1.

[0053] Let the observed quantity be Z k , W k With V k are process noise variables and measurement noise variables respectively, and the linear model is used as the assumed basic filtering model for ease of calculation, and we get:

[0054] State transition model:

[0055] X k+1 =AX k +W k (8)

[0056] Observation model:

[0057] Z k =HX k +V k (9)

[0058] Where A is the state transfer matrix, δ t is the time step, H is the observation matrix, H =

[10] ;

[0059] The core of UKF is UT transformation, which essentially uses a certain number of Sigma points to approximate a Gaussian distribution with the same mean and covariance as the original distribution, and substitutes these points into the model system according to the weights. The symmetric sampling method is selected, and the Sigma point generation method is as follows:

[0060] (1.1) Initialize 2n+1 Sigma points

[0061]

[0062] Where λ is the scaling parameter, is the mean, P is the covariance matrix, represents the i-th column of the square root of the matrix, and (n+λ)P is a semi-positive definite matrix, and n is the dimension of the state quantity;

[0063] (1.2) Sigma point weight calculation

[0064]

[0065] in is the weight of the mean and covariance of the ith sampling point, a represents the distribution of the sampling points, and γ is used to merge the medium and high-order moments, and is a non-negative number;

[0066] UKF forecasts and updates:

[0067] (1.3) State prediction

[0068] X (i) (k+1|k)=AX (i) (k|k) (12)

[0069]

[0070] Where i = 0, 1, 2, ..., 2n, is the prior estimate of the state quantity, Q is the process noise covariance matrix;

[0071] (1.4) Observation prediction

[0072] Z (i) (k+1|k)=HX (i) (k+1|k) (15)

[0073]

[0074] Where i = 0, 1, 2, ..., 2n, is the prior estimate of the observed value, is the cross-covariance matrix between states and observations, is the observed covariance matrix, L is the measurement noise covariance matrix;

[0075] (1.5) Update

[0076] Calculate the Kalman gain:

[0077]

[0078] Status Update:

[0079]

[0080] Covariance update:

[0081]

[0082] Preferably, in step (2), in order to reduce the impact of weather factors in the outdoor environment, an RSSI ranging compensation model is established:

[0083] RSSI′ k =B+N lg(d k )+C+D lg(d k ) (twenty two)

[0084] Among them, RSSI′ k The distance is d k The signal strength measurement value at

[0085] RSSI k and RSSI′ k The error model is:

[0086] error k =RSSI′ k -RSSI k =C+D lg(d k ) (twenty three)

[0087] where error k and RSSI′ k There is a linear relationship, recorded as:

[0088] error k=αRSSI′ k +β (24)

[0089] The RSSI′ k with error k are all data obtained through observation and calculation at time k,

[0090] According to the method of finding the extreme value of multivariate function, we can get:

[0091]

[0092] Preferably, in step (3), due to the existence of factors such as GNSS chip positioning time, Mesh information propagation time, and network time non-strict synchronization, there is a time difference between the wearable drive collar acquiring the location and broadcasting it and the Bluetooth tag. Considering that in a dynamic environment, the time from the last GNSS positioning to the last GNSS positioning is t m The farther away, the greater the positioning error. The weight function is calculated as follows:

[0093]

[0094] Among them, t m The time since the last positioning of the mth wearable repellent collar, d m The distance to the mth wearable repellent collar,

[0095] The coordinates obtained by the wearable repellent collar are in the longitude and latitude format under the wgs84 coordinate system. The longitude and latitude format is retained for subsequent calculations. The distance calculation formula between the coordinate data in the longitude and latitude format is:

[0096] s(y,y m ) = s′(lon, lat, lon m ,lat m ) (27)

[0097] Among them, lon and lat represent the corresponding longitude and latitude values, lon m ,lat m Represents the latitude and longitude values ​​of the mth wearable repellent collar, which is calculated as follows:

[0098]

[0099] The original optimization function becomes:

[0100]

[0101] The optimization function is solved according to the Barzilai-Borwein gradient descent method:

[0102]

[0103] Among them, y j is the coordinate result of the jth iteration, α j is the step length, is the j-th gradient,

[0104]

[0105] Use the difference method to calculate the partial derivatives respectively and get the Jacobian matrix as the gradient.

[0106] Preferably, in step (3), according to the Barzilai-Borwein gradient descent method, the step size is solved by the following method:

[0107]

[0108] When j = 0, α0 is obtained by the backtracking straight line method, and a0 = a0 / 2 is continuously set, starting from α0 = 1 and looping until the following equation is no longer true

[0109]

[0110] Among them, Δy j =[1,1] T , take μ=0.2.

[0111] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the protection scope of the technical solution of the present invention.

Claims

1. Livestock positioning control device based on GNSS and Bluetooth Mesh networking, characterized by: The device includes a wearable repellent collar, a Bluetooth tag, and a monitoring platform. The wearable repellent collar and the Bluetooth tag transmit data via Bluetooth Mesh networking. The Bluetooth tag converts the communication signal strength with the surrounding wearable repellent collars into the distance to the surrounding wearable repellent collars, calculates its own location information based on the GNSS positioning information of the wearable repellent collars, and transmits the location information to the nearest wearable repellent collar through the Mesh network; After the wearable driving collar receives the location information of the surrounding Bluetooth tags, it transmits its own GNSS positioning information and the information of each Bluetooth tag to the monitoring platform through 4G communication technology. The herders can monitor the location of the livestock in real time through the monitoring platform. The wearable drive collar determines whether the livestock exceeds the rear fence boundary buffer zone or leaves the fence boundary based on its own GNSS position information and the coordinates of the virtual fence vertices. If it exceeds the fence boundary buffer zone, the livestock will be stimulated by sound; if it exceeds the fence boundary, the livestock will be stimulated by electric shock. The goal of controlling the entire herd is achieved by controlling the leadership livestock within the fence.

2. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 1 is characterized in that: The wearable driving collar includes: a main control computing module, a power management module, a storage module, a positioning transmission module, a motion information collection module and a control stimulation module; the main control computing module controls peripheral logic, data analysis and processing, and Bluetooth Mesh networking; the power management module provides corresponding voltages to each module; the storage module stores fence information, location data that has not been sent in time, and stores different types of sounds required for sound stimulation; the positioning transmission module acquires livestock GNSS location information and communicates with the monitoring platform; the motion information collection module monitors the motion status of livestock in real time; the control stimulation module emits sound and electric shock stimulation to constrain the livestock within the set fence boundary.

3. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 2 is characterized in that: The method for judging whether livestock has exceeded the fence boundary is: The fence boundary has n coordinate points P1, P2, ..., P n , P i , the longitude and latitude are lon i With lat i , n coordinate points are sent by the monitoring platform, and at the same time the monitoring platform sends a coordinate point center inside the fence, with longitude and latitude lon c With lat c , Convert longitude and latitude to planar coordinates using the Mercator projection: Among them, i = 1, 2, ..., n, the radius of the earth R = 6378137m, In the plane coordinate system, a straight line is determined based on two points, and n boundary lines are calculated, l1, l2, ..., l n , Among them, l i By P i With P i+1 Calculated, i = 1, 2, ..., n-1; l n By P n Calculated with P1, l i :y i =a i ×x i +b i (3) a i With b i is the slope and intercept of the i-th boundary line, Calculation and judgment i , i=1,2,…,n, Judge calculated based on the GNSS position information of the wearable repellent collar i ' and judgement basis i If they match, the livestock is inside the fence; otherwise, it is outside the fence boundary.

4. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 3 is characterized in that: The main control computing module selects a low-power MCU. In order to increase the battery life of the collar, two modes, working mode and standby mode, are designed. The main control computing module is responsible for switching the collar working mode. At the same time, the main control module reasonably completes the adjustment of the collar working interval according to the motion information collection data, completes the update of the fence boundary and the judgment of whether it exceeds the fence according to the information of the positioning transmission module, and the main control computing module also completes the Bluetooth Mesh networking and broadcasts information in the Mesh network through the Bluetooth chip; The power management module includes lithium battery, solar panel, voltage conversion circuit, and charging circuit; The storage module includes SRAM, Flash and SD card. SRAM is used to store data in working mode, Flash is used to store non-lost data in program and standby mode, and SD card is used to store audio during sound stimulation. The positioning transmission module includes GNSS and 4G related circuits and antennas. The main control calculation module reads the location information of the positioning transmission module and transmits data with the monitoring platform through 4G. The motion information acquisition module uses sensors to collect real-time motion information, and the main control calculation module uses this information to adjust the collar working time. The control stimulation module includes sound stimulation and electric shock stimulation. Sound stimulation relies on the voice chip to read the audio in the SD card, and then plays the sound through the power amplifier and speaker; electric shock stimulation uses the transformer coil to achieve the purpose of boosting, and uses an optocoupler relay to control the generation of high voltage to improve the safety of the collar.

5. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 4 is characterized in that: The Bluetooth tag includes a main control Bluetooth module and a power management module; The main control Bluetooth module uses the Bluetooth chip to implement functions, which are used for the workflow of Bluetooth positioning tags, completing Bluetooth Mesh networking and receiving information in the Mesh network, RSSI data collection and processing and calculation of its own location information, as well as data transmission with other devices in the Mesh network; The power management module includes lithium batteries, solar panels, voltage conversion circuits, and charging circuits. Lithium batteries with high energy density and support for charging and discharging are selected to power the Bluetooth tags. The corresponding DC-DC conversion circuits are designed to power the chips in the tags. The Bluetooth tags are charged by solar energy.

6. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 5, characterized in that: Bluetooth tag positioning method based on RSSI data in Mesh network The method comprises the following steps: (1) When the wearable repellent collar is in working mode, it broadcasts heartbeat packet data every second, the Bluetooth tag receives the information, collects RSSI data, and uses the unscented Kalman filter (UKF) to preprocess the data to reduce the impact of noise; (2) Based on the relationship between RSSI data and distance, an RSSI ranging model is established: RSSI k =B+Nlg(d k ) (5) where d k is the distance, in meters, RSSI k is the corrected signal strength at time k, in dBm, B and N are constants; (3) Convert the GNSS position information into an unconstrained optimization problem, the function form is: Among them, y is the coordinate vector in the vertical coordinate form of the Bluetooth tag to be calculated, y m is the coordinate vector in the ordinate form of the mth wearable repellent collar obtained by the Bluetooth tag, M is the number of wearable repellent collar information that the Bluetooth tag can obtain, ω m is the weight of the mth wearable repellent collar information, d m is the distance between the Bluetooth tag and the mth wearable repellent collar.

7. The livestock positioning control based on GNSS and Bluetooth Mesh networking according to claim 6 The device is characterized by: In step (1), when the unscented Kalman filter UKF is applied The state quantity is: RSSI k is the RSSI data at time l, ΔRSSI k is the difference between time l and time k-1; Let the observed quantity be Z k , W k With V k are process noise variables and measurement noise variables respectively, and the linear model is used as the assumed basic filtering model for ease of calculation, and we get: State transition model: X k+1 =AX k +W k (8) Observation model: Z k =HX k +V k (9) Where A is the state transfer matrix, δ t is the time step, H is the observation matrix, H = [1 0]; The core of UKF is UT transformation, which essentially uses a certain number of Sigma points to approximate a Gaussian distribution with the same mean and covariance as the original distribution, and substitutes these points into the model system according to the weights. The symmetric sampling method is selected, and the Sigma point generation method is as follows: (1.1) Initialize 2n+1 Sigma points Where λ is the scaling parameter, is the mean, P is the covariance matrix, represents the i-th column of the square root of the matrix, and (n+λ)P is a semi-positive definite matrix, and n is the dimension of the state quantity; (1.2) Sigma point weight calculation in is the weight of the mean and covariance of the ith sampling point, a represents the distribution of the sampling points, and γ is used to merge the medium and high-order moments, and is a non-negative number; UKF forecasts and updates: (1.3) State prediction X (i) (k+1|k)=AX (i) (k|k) (12) Where i = 0, 1, 2, ..., 2n, is the prior estimate of the state quantity, Q is the process noise covariance matrix; (1.4) Observation prediction Z (i) (k+1|k)=HX (i) (k+1|k) (15) Where i = 0, 1, 2, ..., 2n, is the prior estimate of the observed quantity, is the cross-covariance matrix between states and observations, is the observed covariance matrix, L is the measurement noise covariance matrix; (1.5) Update Calculate the Kalman gain: Status Update: Covariance update:

8. The livestock positioning control based on GNSS and Bluetooth Mesh networking according to claim 7 The device is characterized by: In step (2), in order to reduce the impact of weather factors in the outdoor environment, an RSSI ranging compensation model is established: RSSI′ k =B+Nlg(d k )+C+Dlg(d k ) (22) Among them, RSSI k ′ is the distance d k The signal strength measurement value at RSSI k With RSSI k ' k The error model is: error k =RSSI′ k -RSSI k =C+Dlg(d k ) (23) where error k With RSSI k ′ has a linear relationship, recorded as: error k =αRSSI′ k +β (24) The RSSI′ k with error k are all data obtained through observation and calculation at time k, According to the method of finding the extreme value of multivariate function, we can get:

9. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 8, characterized in that: In step (3), due to the existence of factors such as GNSS chip positioning time, Mesh information propagation time, and network time non-strict synchronization, there is a time difference between the wearable drive collar obtaining the location and broadcasting it and the Bluetooth tag. Considering that in a dynamic environment, the time from the last GNSS positioning to the last GNSS positioning is t m The farther away, the greater the positioning error. The weight function is calculated as follows: Among them, t m The time since the last positioning of the mth wearable repellent collar, d m The distance to the mth wearable repellent collar; The coordinates obtained by the wearable repellent collar are in the longitude and latitude format under the wgs84 coordinate system. The longitude and latitude format is retained for subsequent calculations. The distance calculation formula between the coordinate data in the longitude and latitude format is: s(y,y m )=s′(lon,lat,lon m ,lat m ) (27) Among them, lon and lat represent the corresponding longitude and latitude values, lon m ,lat m Represents the latitude and longitude values ​​of the mth wearable repellent collar, which is calculated as follows: The original optimization function becomes: The optimization function is solved according to the Barzilai-Borwein gradient descent method: Among them, y j is the coordinate result of the jth iteration, α j is the step length, is the j-th gradient, Use the difference method to calculate the partial derivatives respectively and get the Jacobian matrix as the gradient.

10. The livestock positioning control device based on GNSS and Bluetooth Mesh networking according to claim 9, characterized in that: In step (3), according to the Barzilai-Borwein gradient descent method, the step size is solved by the following method: When j = 0, α0 is obtained by the backtracking straight line method, and α0 = α0 / 2 is continuously set, starting from α0 = 1 until the following equation is no longer true Among them, Δy j =[1,1] T , take μ=0.2.

Images