A Beidou and RFID combined positioning method and system

Abstract

A Beidou and RFID combined positioning method and system utilizes the Beidou and RFID combined positioning method to determine whether the positioning target is indoors or outdoors based on the activation state and time of the RFID tag. If it is outdoors, Beidou is used for positioning, and RFID is used for indoor positioning. The positioning signal is corrected using the differential evolution method, and the obtained positioning coordinates are passed into the trained neural network model for a second error correction. This obtains more accurate positioning data and improves the reliability of positioning.

Description

Technical Field

[0001] The present invention belongs to the field of positioning and communication technology, and in particular relates to a Beidou and RFID combined positioning method and system. Background Art

[0002] The BeiDou satellite navigation System (BDS) is a global satellite navigation positioning and communication system independently developed by my country in the 1980s. It is one of the world's four major satellite navigation systems along with the US GPS, Russia's GLONASS and the EU's Galileo.

[0003] The error sources of Beidou navigation satellites can be roughly divided into satellite-related errors, propagation errors, and receiver errors. Differential positioning technology can reduce satellite-related errors and provide accurate positioning, with a positioning accuracy of 3m. At present, the most common positioning method for the Beidou system is combined navigation pseudo-range positioning. Beidou has good indoor positioning, but in complex terrain or indoors, Beidou's positioning accuracy is very poor. As far as indoor positioning is concerned, there are short-range wireless positioning technologies such as WIFI, UWB, and RFID. Among them, RFID positioning technology has the advantages of low cost and high positioning accuracy, and has a wide range of application prospects. Using the combined positioning method of Beidou and RFID, all-round indoor and outdoor positioning can be achieved.

[0004] However, in actual environments, the RSSI ranging method is seriously affected by multipath attenuation and obstruction such as signal reflection, scattering, and diffraction, and the measurement results have large errors. Outdoors, actual engineering requirements may require centimeter-level positioning, and the positioning accuracy provided by Beidou positioning is not satisfactory.

[0005] In the existing technology, Beidou / UWB combined positioning is used to achieve seamless indoor and outdoor positioning. The Beidou positioning system is used to determine the outdoor measurement point location, and UWB technology is used to complete indoor fixed-point measurement. In principle, UWB technology is difficult to complete elevation measurement, and can only determine the straight-line distance between the measurement initial point and the point to be measured. Indoors, not only the straight-line distance but also the elevation measurement must be completed. Beidou and inertial navigation combined positioning system can complete indoor and outdoor combined positioning very well, but inertial navigation measurement requires knowing the location of the initial point, and the measurement cost is high.

[0006] Differential evolution is a random population search method with strong global optimization capabilities for solving nonlinear, high-dimensional and complex optimization problems. The differential evolution method is based on populations and mainly includes three operation steps: mutation, crossover and selection. Its advantages are simple principles, fewer core steps, less control by parameters, more mutation methods, strong adaptability to different objective functions and good practicality.

[0007] In recent years, artificial intelligence has developed rapidly and provided intelligent solutions to various problems. As a branch of neural network, RNN network has good numerical mapping ability and can solve nonlinear problems and time series well.

[0008] By using differential evolution combined with a bidirectional LSTM neural network, the position error between the tested label and the real label is continuously reduced to improve the positioning accuracy. Summary of the invention

[0009] In order to solve the deficiencies in the prior art, the present invention provides a Beidou and RFID combined positioning method and system, which utilizes the Beidou and RFID combined positioning mode. If it is outdoors, Beidou is used for positioning, and RFID is used for indoor positioning. The positioning signal is corrected using the differential evolution method, and the obtained positioning coordinates are passed into the trained neural network model for a second error correction. Thereby obtaining more accurate positioning data and improving the reliability of positioning. The present invention continues to conduct in-depth research on the existing neural network model to improve the defects and deficiencies of the existing neural network. Although the traditional differential evolution method can adaptively adjust the optimal solution method, it cannot solve the problem of nonlinearity. The traditional neural network has a strong nonlinear fitting property and can map various nonlinear relationships, but the defects are also more obvious, such as being prone to falling into local optimality, slow convergence speed, and other problems, resulting in a high false alarm rate of detection. The present invention selects the differential evolution method to optimize it and constructs an improved adaptive differential evolution neural network model. Figure 4 It is a training loss curve diagram of the neural network model after optimization by the neural network model and the differential evolution algorithm. The star line is the loss curve of the neural network model sequence, and the dotted line is the loss curve of the neural network model after optimization by the differential evolution method. It is found that the dotted line has a shorter convergence time and a smaller loss value, indicating that the positioning method of the present invention has a faster training speed and higher accuracy.

[0010] The present invention adopts the following technical solution.

[0011] A Beidou and RFID combined positioning method comprises the following steps:

[0012] Step 1: Use the Beidou / RFID system combination to perform rough positioning through polygon positioning;

[0013] Step 2, optimizing the rough positioning by differential evolution method;

[0014] Step 3: Use the trained neural network to optimize the positioning result obtained in step 2.

[0015] Preferably, in step 1, the Beidou system obtains the rough positioning result of the outdoor target through the simultaneous equations of the ranging results of 4 satellites; the indoor positioning is completed by the RFID system, and the rough positioning result of the indoor target is obtained by measuring the arrival time and intensity parameters of the radio frequency signal combined with the polygonal positioning method.

[0016] Preferably, in step 2,

[0017] The measured distance between the satellite or reader and the tag is obtained by Beidou / RFID combined positioning as d i (i=1,2,3,4,…,N), the measured distance is used as the reference distance for estimating the position of the tag to be measured; suppose the coordinates of the tag to be measured are (x0,y0,z0), and the coordinates of the reader are (x i ,y i , z i )(i=1,2,3,4,…,N), thus the distance error between the i-th reader and the tag to be tested and the distance error between the reader and the reference tag is obtained, which is expressed as a function as follows:

[0018]

[0019] For each label to be tested, a distance error function is listed according to the above formula, and then the mutation, crossover and selection operations of the differential evolution method are performed respectively, and the evolution is iterated multiple times. Finally, the minimum distance error value is taken and used as the estimated position coordinate of the label to be tested.

[0020] Preferably, step 2 specifically includes:

[0021] Step 2.1, normalize and preprocess the data collected by coarse positioning;

[0022] Step 2.2, configure parameters and set the population size to N p , that is, the number of reference labels is N; the number of labels to be tested is M; the individual dimension is D, which represents the two-dimensional positioning area; the mutation rate is set to F; the crossover rate is C R , set the maximum number of iterations;

[0023] Step 2.3, generate the objective function. First, randomly generate the coordinates of M tags to be tested, and use the received signal strength to calculate the reference distance between the reader and the reference tag; then use the known N reference tag positions as the initial optimization coordinates, and continuously optimize the position coordinates of the tags to be tested based on the minimum function distance error, so that the error gradually approaches zero;

[0024] Step 2.4, take the N reference label coordinates as the input of the differential evolution method, then use the mutation strategy of DE / rand / 1 / bin for mutation operation, then use the binary method for crossover operation, and finally get the final estimated coordinates according to the minimum distance error function;

[0025] Step 2.5, determine the termination condition and output the estimated coordinate value.

[0026] In step 2.5, the judgment termination condition is:

[0027] 3) The maximum number of iterations is reached;

[0028] 4) Achieve a given distance error value.

[0029] Preferably, step 3 comprises:

[0030] Step 3.1, network construction;

[0031] The number of hidden layer units is H. During forward propagation, forward calculation is performed from time 1 to time t to obtain and save the output of the hidden layer at each moment. During backward propagation, reverse calculation is performed from time t to time 1 to obtain and save the output of the hidden layer at each moment. Finally, the final output is obtained by combining the output structure of the corresponding moment of forward propagation and backward propagation at each moment.

[0032] Step 3.2, model training;

[0033] Assume the real coordinates of the object are (x, y, z). After Beidou / RFID locates the object, the output coordinates are The loss function chooses the mean square error function

[0034]

[0035] In the formula, n represents the size of the data. The n groups of data are divided into training set, validation set, and test set in a ratio of 8:1:1. The training data is trained 50,000 times, and the optimization method uses the stochastic gradient descent method with a learning rate of 0.005.

[0036] The beneficial effect of the present invention is that, compared with the prior art,

[0037] 1. Differential evolution and neural network correction are used to achieve higher positioning accuracy;

[0038] 2. Use RFID for indoor positioning, low cost;

[0039] 3. Use Beidou positioning outdoors and RFID positioning indoors to achieve seamless spatial positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of a Beidou and RFID combined positioning method of the present invention;

[0041] Figure 2Schematic diagram of neural network structure;

[0042] Figure 3 A comparison chart of the positioning accuracy of the present invention and the positioning accuracy of the prior art;

[0043] Figure 4 It is a schematic diagram comparing the loss curves of the training set of the positioning method of the present invention and the training set of the training set using only the neural network model. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0045] A Beidou and RFID combined positioning method, such as Figure 1 As shown, the following steps are included:

[0046] Step 1: Use the Beidou / RFID system combination to perform rough positioning through polygon positioning;

[0047] The principle of Beidou satellite positioning is that the satellite rotates around the center of the earth at a fixed period at an altitude of 20,000 meters above the ground, which means that any point on the ground is monitored by four satellites at any time. During the operation of the system, the position accuracy of the satellite is known, and the three satellites in the application system can obtain the coordinate equation of the spatial position of a certain point. The three equations can be combined to solve the unknown number of the spatial position. However, due to the error between the satellite clock and the ground clock, in order to improve the positioning accuracy, there will actually be four unknown numbers in the equation calculation, and the fourth satellite needs to be introduced to complete the establishment of the equation group.

[0048] The RFID-based indoor positioning method mainly determines the location of the target by measuring the arrival time, strength and other parameters of the radio frequency signal in combination with the corresponding method. According to the TOA (Time of arrival) principle, the target position is obtained through multilateral positioning and the improved mean method. The positioning function is mainly divided into two stages: ranging and positioning. In the ranging stage, the distance between the target and the known reader is calculated for target positioning. The more accurate the distance, the better the positioning effect. In the positioning stage, polygonal positioning is used for positioning to obtain the positioning coordinates of the target.

[0049] Step 2, optimizing the rough positioning result of step 1 by differential evolution method;

[0050] The measured distance between the satellite or reader and the tag is obtained by Beidou / RFID combined positioning as d i (i=1, 2, 3, 4, ..., N), the measured distance is used as the reference distance to estimate the position of the tag to be measured. Assume that the coordinates of the tag to be measured are (x0, y0, z0) and the coordinates of the reader are (x i ,y i , z i )(i=1, 2, 3, 4, ..., N), the distance error between the i-th reader and the tag to be tested and the distance error between the reader and the reference tag are obtained, which can be expressed as a function as follows:

[0051]

[0052] Therefore, for each label to be tested, a distance error function can be listed according to the above formula, and then the mutation, crossover and selection operations of the differential evolution method are performed respectively, and the evolution is iterated multiple times to finally take the minimum distance error value and use this value as the estimated position coordinate of the label to be tested.

[0053] The differential evolution method includes the following steps:

[0054] 1) Population initialization. R individuals are randomly and uniformly generated in the solution space of joint positioning, and each individual consists of a k-dimensional vector

[0055] X i (0)=(x i，1 (0), x i，2 (0), x i，3 (0), ..., x i，k (0))(i=1,2,3,...,R)

[0056] The value of the jth dimension of the i-th individual is determined as follows:

[0057] X i，j (0) = L j_min +rand(0,1)(L j_max →L j_min )

[0058] i=1,2,3,...,R

[0059] j = 1, 2, 3, ..., k

[0060] 2) Mutation. In the gth iteration, three individuals X are randomly selected from the population. p1 (g), X p2 (g), X p3 (g), and p1≠p2≠p3≠i, the generated mutation vector is:

[0061] H i (g) = Xp1 (g)+F(X p2 (g)-x p3 (g)),

[0062] where Δ p2，p3 (g) = X p2 (g)-X p3 (g) is the difference vector, and F is the scaling factor

[0063] 3) Crossover. Crossover can produce a better next generation

[0064] 4) Selection operation: To generate the next generation of target vectors, the best individuals are selected based on the adaptability of the target vector and the test vector.

[0065]

[0066] For each individual, X i (g+1) is better than or equal to X i (g) Achieve overall optimality through mutation, crossover, and selection.

[0067] Optimizing Beidou / RFID positioning through differential evolution method includes the following specific implementation steps:

[0068] Step 2.1, data collection. Select multiple positioning targets, evenly distribute them indoors and outdoors, and adjust their positions at regular intervals. For outdoor positioning, use the Beidou positioning system to locate the positioning target multiple times in different time periods and weather conditions to obtain positioning data; for indoor positioning, use the RFID positioning system to locate the positioning target multiple times in different lighting conditions and time periods to obtain positioning data;

[0069] Step 2.2, data preprocessing. The data is the positioning data optimized by the differential evolution method. Based on the principle of reducing the adjustment range of weights, the range of input data and output data in the network is limited between 0 and 1. The positioning coordinate data is large, so the input data needs to be normalized. There are generally two normalization methods, the maximum and minimum normalization method and the mean variance normalization method;

[0070] Step 2.3, set parameters. Set the population size to N p , that is, the number of reference labels is N; the number of labels to be tested is M; the individual dimension is D, representing the two-dimensional positioning area; the mutation rate is set to F; the crossover rate is C R , set the maximum number of iterations;

[0071] Step 2.4, generate the objective function. First, randomly generate the coordinates of M tags to be tested, and use the received signal strength to calculate the reference distance between the reader and the reference tag; then use the known N reference tag positions as the initial optimization coordinates, and continuously optimize the position coordinates of the tags to be tested based on the minimum function distance error, so that the error gradually approaches zero;

[0072] Step 2.5, differential evolution method optimization. Take the N reference label coordinates as the input of the differential evolution method, then use the mutation strategy of DE / rand / 1 / bin for mutation operation, then use the binary method for crossover operation, and finally get the final estimated coordinates according to the minimum distance error function;

[0073] Step 2.6, determine the termination condition. Two conditions are used to determine the termination: if the maximum number of iterations is reached, the coordinates output at this time are used as the best estimated coordinates; if the given distance error value is reached, the estimated coordinates at this time are output and the number of iterations is recorded.

[0074] Step 3: Use the neural network to further optimize the positioning result of step 2 through model training;

[0075] Step 3.1, network construction.

[0076] The network structure is as follows Figure 2 As shown, the number of hidden layer units is H. During forward propagation, forward calculation is performed from time 1 to time t (t=3), and the output of the hidden layer at each moment is obtained and saved. During backward propagation, reverse calculation is performed from time t to time 1, and the output of the hidden layer at each moment is obtained and saved. Finally, the final output is obtained by combining the output structure of the corresponding moment of forward propagation and backward propagation at each moment;

[0077] Step 3.2, model training.

[0078] Assume the real coordinates of the object are (x, y, z). After Beidou / RFID locates the object, the output coordinates are The loss function chooses the mean square error function

[0079]

[0080] In the formula, n represents the size of the data, which is 200000. The 200000 data sets are divided into training set, validation set, and test set in a ratio of 8:1:1. The 160000 data sets are trained 50000 times, and the optimization method uses the stochastic gradient descent method with a learning rate of 0.005.

[0081] A Beidou and RFID combined positioning system comprises a coarse positioning module, a differential evolution optimization module and a neural network optimization module.

[0082] The coarse positioning module uses the combination of Beidou / RFID system to perform coarse positioning through polygon positioning. The Beidou system obtains the coarse positioning result of outdoor targets through the simultaneous equations of the ranging results of 4 satellites; indoor positioning is completed through the RFID system, and the coarse positioning result of indoor targets is obtained by measuring the arrival time, strength and other parameters of the radio frequency signal combined with the polygon positioning method;

[0083] The differential evolution optimization module normalizes and preprocesses the data collected by coarse positioning, configures parameters and generates an objective function to perform mutation, crossover and selection operations of the differential evolution method, iterates evolution multiple times, and finally takes the minimum value of the distance error, and uses this value as the estimated position coordinate of the tag to be tested;

[0084] The neural network optimization module further optimizes the positioning results through network construction and model training.

[0085] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0086] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0087] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0088] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A Beidou and RFID combined positioning method, characterized in that: The following steps are involved: Step 1: Use the Beidou / RFID system combination to perform rough positioning through polygon positioning; Step 2, optimizing the rough positioning by differential evolution method; Step 3, using the trained neural network to optimize the positioning result obtained in step 2; Step 3 includes: Step 3.1, network construction; The number of hidden layer units is H. During forward propagation, forward calculation is performed from time 1 to time t to obtain and save the output of the hidden layer at each moment. During backward propagation, reverse calculation is performed from time t to time 1 to obtain and save the output of the hidden layer at each moment. Finally, the final output is obtained by combining the output results of the corresponding moments of forward propagation and backward propagation at each moment. Step 3.2, model training; Assume that the real coordinates of the object are (x, y, z). After the data of Beidou / RFID positioning the object passes through the neural network, the output coordinate value is ( ), the loss function selects the mean square error function as follows: In the formula, n represents the size of the data, and the n groups of data are divided into training set, validation set, and test set in a ratio of 8:1:

1.

2. A Beidou and RFID combined positioning method according to claim 1, characterized in that: In step 1, the Beidou system obtains the rough positioning result of the outdoor target through the simultaneous equations of the ranging results of the four satellites; indoor positioning is completed through the RFID system, and the rough positioning result of the indoor target is obtained by measuring the arrival time and intensity parameters of the radio frequency signal combined with the polygonal positioning method.

3. The Beidou and RFID combined positioning method according to claim 1, characterized in that: In step 2, The measured distance between the satellite or reader and the reference tag obtained by Beidou / RFID combined positioning is , i=1,2,3,4,…,N, the measured distance is used as the reference distance for estimating the position of the tag to be measured; let the coordinates of the tag to be measured be ( , , ), the coordinates of the reader are ( , , ), i=1,2,3,4,…,N, thus the distance error between the i-th reader and the tag to be tested and the distance error between the reader and the reference tag are obtained, which can be expressed as a function as follows: For each label to be tested, a distance error function is listed according to the above formula, and then the mutation, crossover and selection operations of the differential evolution method are performed respectively, and the evolution is iterated multiple times. Finally, the minimum distance error value is taken and used as the estimated position coordinate of the label to be tested.

4. The Beidou and RFID combined positioning method according to claim 3, characterized in that: Step 2 specifically includes: Step 2.1, normalize and preprocess the data collected by coarse positioning; Step 2.2, configure parameters and set the population size to , that is, the number of reference labels is N; the number of labels to be tested is M; the individual dimension is D, which represents the two-dimensional positioning area; the mutation rate is set to F; the crossover rate is , set the maximum number of iterations; Step 2.3, generate the objective function. First, randomly generate the coordinates of M tags to be tested, and use the received signal strength to calculate the reference distance between the reader and the reference tag. Then, use the known N reference tag positions as the initial optimization coordinates, and continuously optimize the position coordinates of the tags to be tested based on the minimum function distance error, so that the error gradually approaches zero. Step 2.4, take the N reference label coordinates as the input of the differential evolution method, then use the mutation strategy of DE / rand / 1 / bin for mutation operation, then use the binary method for crossover operation, and finally get the final estimated coordinates according to the minimum distance error function; Step 2.5, determine the termination condition and output the estimated coordinate value.

5. The Beidou and RFID combined positioning method according to claim 4, characterized in that: In step 2.5, the judgment termination condition is to meet any of the following conditions: 1) The maximum number of iterations is reached; 2) Achieve a given distance error value.

6. A Beidou and RFID combined positioning system, used to implement a Beidou and RFID combined positioning method according to any one of claims 1 to 5, comprising a coarse positioning module, a differential evolution optimization module, and a neural network optimization module, characterized in that: The coarse positioning module is used to perform coarse positioning using the combination of the Beidou / RFID system through polygonal positioning. The Beidou system obtains the coarse positioning result of outdoor targets through the simultaneous equations of the ranging results of four satellites; indoor positioning is completed through the RFID system, and the coarse positioning result of indoor targets is obtained by measuring the arrival time and intensity parameters of the radio frequency signal combined with the polygonal positioning method; The differential evolution optimization module is used to normalize and preprocess the data collected by coarse positioning, configure parameters and generate objective functions to perform mutation, crossover and selection operations of the differential evolution method, iterate evolution multiple times, and finally take the minimum value of the distance error, and use this value as the estimated position coordinate of the tag to be tested; The neural network optimization module is used to further optimize the positioning results through network construction and model training.

7. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images