Rfid-based unwrapping phase-position model tag positioning method and system

Abstract

The application belongs to the technical field of RFID application, and discloses a phase-position model label positioning method and system based on RFID. The method comprises the following steps: S1, obtaining phase data of each RFID label at different time points; S2, obtaining the corresponding relationship between the antenna position at different time points and the unwrapped phase by combining the relationship between the unwrapped phase and time; S3, judging the unwrapped risk points and the time corresponding to the unwrapped risk points, numbering each unwrapped risk point and constructing an unwrapped risk point serial number matrix; and S4, constructing an optimization model, solving the optimization model to obtain the required output, and adjusting the position of the RFID label according to the solved value. Through the application, the problem that the phase collection of dense labels cannot meet the space sampling theorem is solved.

Description

Technical Field

[0001] This invention belongs to the field of RFID application related technology, and more specifically, relates to a method and system for RFID-based unwrapped phase-position model tag positioning. Background Technology

[0002] RFID (Radio Frequency Identification) technology, as a core technology of the Internet of Things (IoT), has been widely used in logistics, retail, warehousing, and manufacturing as a means of identification, significantly improving the efficiency and accuracy of item identification. Furthermore, using RFID technology for item and robot positioning can further enhance the effectiveness of RFID applications, and integrating location information can elevate traditional item management to a whole new level. Using mobile robots to inventory and locate RFID tags is an important means of improving warehouse automation and inventory efficiency.

[0003] Current RFID positioning systems based on mobile robots or mobile platforms typically mount RFID antennas on the mobile platform. The antennas move, measuring the tags to form a positioning system similar to synthetic aperture radar. Positioning algorithms based on this system usually use probabilistic holograms to locate the tags. This method first divides the space to be located into a grid, then calculates the probabilistic hologram value for each grid using the acquired RFID tag phase data. Finally, by comparison, the grid with the highest probabilistic hologram value is selected as the positioning result. While this method can achieve centimeter-level accuracy, its accuracy depends on a large number of grid calculations, resulting in a significant computational burden, which increases exponentially when expanding from two-dimensional to three-dimensional positioning. To overcome this excessive computational burden, an RFID positioning method based on an unwrapped phase-position model has been proposed. First, the phase measurements are unwrapped. The unwrapping algorithm compares adjacent phase measurements to eliminate periodic phase ambiguity, thus converting the phase-position model into an unwrapped phase-position model. An objective equation is established based on this model, which has better solvability than equations based on the phase-position model. It has only one extremum and no multiple local optima, and the gradient of the function is highly singular. Finally, a nonlinear optimization method is used to solve the objective equation. This method can achieve centimeter-level positioning accuracy with low computational load. However, the unwrapping algorithm of this method relies heavily on continuous phase acquisition. If two consecutive phase measurements cannot satisfy the spatial sampling theorem, the unwrapping result will be incorrect. This situation frequently occurs in practical RFID positioning scenarios, especially in scenarios with dense tags, where the feasibility and accuracy of the positioning algorithm are difficult to guarantee. Summary of the Invention

[0004] To address the above-mentioned deficiencies or improvement needs of existing technologies, this invention provides a method and system for RFID-based unwrapped phase-position model tag positioning, solving the problem that phase acquisition of dense tags cannot satisfy the spatial sampling theorem.

[0005] To achieve the above objectives, according to one aspect of the present invention, an improved RFID-based unwrapped phase-position model tag positioning method is provided, the method comprising the following steps:

[0006] S1 When the mobile robot moves with the antenna and communicates with the RFID tag, record the odometer data of the mobile robot over time, collect the communication signal between the antenna and the RFID tag and extract the phase data that changes over time in the communication signal, and match the phase data with the RFID tag one by one to obtain the phase data of each RFID tag at different times.

[0007] S2 performs phase unwrapping on the phase data of each RFID tag to obtain the relationship between the unwrapped phase and time. Using the positional relationship between the antenna and the mobile robot and the odometer data of the mobile robot changing over time, the position of the antenna at different times is obtained. Combining the relationship between the unwrapped phase and time, the correspondence between the antenna position and the unwrapped phase at different times is obtained.

[0008] S3 uses the correspondence obtained in step S2 to solve the phase gradient, phase gradient difference and the residual matrix corresponding to them respectively. Using the residual threshold, the untangling risk point and the time corresponding to the untangling risk point are determined according to the preset conditions. Each untangling risk point is numbered and an untangling risk point number matrix is ​​constructed accordingly.

[0009] S4 constructs an optimization model with the sequence matrix of the untangling risk points, the untangling phase, and the antenna position as inputs, and the phase adjustment value, the RFID tag position, and the phase offset as outputs. Solving the optimization model yields the required output, and the position of the RFID tag is adjusted according to the solved values.

[0010] More preferably, in step S1, the optimization model is performed according to the following relationship:

[0011]

[0012] Where J is the objective equation, γ m It is the unwrapped phase corresponding to the m-th antenna, C m K is the phase unwrapping risk matrix for the m-th antenna. m Here, λ is the phase adjustment matrix for the m-th antenna, λ is the wavelength corresponding to the operating frequency of the RFID system, and x is the phase adjustment matrix for the m-th antenna. A,m y A,m , zA,m These are the x-coordinate, y-coordinate, z-coordinate, and β values ​​of the m-th antenna, respectively. m Let represent the phase offset value of the m-th antenna, x, y, z are the x, y, and z coordinates of the position of the tag being solved, respectively, i is the i-th sampling time, N is the total number of phase sampling times for this antenna, j is the j-th adjustment matrix of this antenna, and W is the total number of adjustment matrices for this antenna.

[0013] More preferably, in step S2, the phase unwrapping is performed according to the following relationship:

[0014]

[0015] Among them, i=2,3,...,N, γ (1) It is the untangling phase at the first moment, θ (1) It is the original phase at the first moment, γ (i) It is the untangling phase at time i, θ (i) It is the original phase at time i, γ (i-1) It is the unwrapping phase at time i-1, where i is the time of signal sampling and N is the total number of signal sampling times.

[0016] More preferably, in step S2, the positions of the antenna at different times are obtained according to the following formula:

[0017]

[0018] Where, x robot y robot , These are the x-coordinates, y-coordinates, and attitude angles of the robot's pose calculated using the trajectory inference method. A,m y A,m , z A,m These are the x, y, z, and b coordinates of the m-th antenna in the world coordinate system established based on the robot's initial pose. x,m b y,m b z,m These are the x, y, and z coordinates of the position of the m-th calibrated antenna in the robot coordinate system.

[0019] More preferably, in step S3, the phase gradient is calculated according to the following formula:

[0020]

[0021] Where, Δd (i) γ is the displacement of the antenna from time i-1 to time i. (i) It is the untangling phase at time i, γ(i-1) It is the untangling phase at time i-1, G (i) It is the phase gradient at the i-th time step.

[0022] More preferably, in step S3, the phase gradient difference is calculated according to the following formula:

[0023] δ (i) =G (i) -G (i-1)

[0024] Among them, G (i) G is the phase gradient at time i. (i-1) It is the phase gradient at time i-1, δ (i) It is the phase gradient difference at the i-th time step.

[0025] More preferably, in step S3, the phase gradient residual matrix and the phase gradient difference residual matrix are calculated according to the following formulas:

[0026] ζ1=|GG|

[0027] ζ2=|δ-δ|

[0028] Where G and δ are the phase gradient and phase gradient difference corresponding to the phase data, respectively; G and δ are the linear fitting estimates of the phase gradient and phase gradient difference, respectively; and ζ1 and ζ2 are the phase gradient residual matrix and phase gradient difference residual matrix, respectively.

[0029] More preferably, in step S3, the preset conditions are performed as follows:

[0030] when season Or, when season

[0031] Where i is the i-th sampling time. and These are the indices of the unwrapping phases at time i, calculated based on the phase gradient and phase gradient difference, indicating potential unwrapping risks. Each index contains the index of the risk point. Tereshold1 and Tereshold2 represent the set thresholds for the residual matrices of the phase gradient and phase gradient difference, respectively. and These are the phase gradient residual matrix and the phase gradient difference residual at time i, respectively.

[0032] More preferably, the sequence matrix of the untangling risk points is configured as follows:

[0033] r = r1∪r2

[0034] Where r is the index matrix of possible untangling risks in the untangling phase, which contains the index of the untangling risk points, and r1 and r2 are the index matrices of possible untangling risks in the untangling phase calculated based on the phase gradient and the phase gradient difference, respectively.

[0035] According to another aspect of the present invention, a system for the tag positioning method described above is provided, the system comprising a signal generation module, a signal acquisition module, a signal processing module, an untangling risk point judgment module, and an optimization model module, wherein:

[0036] The signal generation module includes a mobile robot, an antenna, and an RFID tag. The antenna is mounted on the mobile robot and moves with it. The RFID tag is positioned opposite the mobile robot and communicates with the antenna.

[0037] The signal acquisition module is used to acquire the odometer data of the mobile robot that changes over time, as well as the communication signal between the antenna and the RFID tag, and to map the RFID tag to the phase one by one.

[0038] The signal processing module is used to unwrap the phase, obtain the correspondence between the antenna position and the unwrapped phase at different times, and then calculate the phase residual corresponding to the phase gradient and the phase gradient difference.

[0039] The untangling risk point judgment module is used to judge untangling risk points according to preset conditions;

[0040] The optimization model module is used to adjust the position and phase of the untangling risk point using the constructed optimization model.

[0041] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0042] 1. In the context of dense tag positioning, due to the limited communication channels, the phase measurement of the tags cannot always satisfy the spatial sampling theorem, which leads to positioning errors and positioning solution failures. The method provided by this invention can handle the unwrapping failure caused by the failure to satisfy the spatial sampling theorem. First, it identifies the points where unwrapping failure may occur, and then simultaneously calculates the adjustment value of the phase of the unwrapping failure and the tag position in the proposed model, effectively dealing with this scenario and solving the positioning problem in dense tag scenarios.

[0043] 2. This invention uses phase data unwrapping to solve the residual matrix of phase gradient and phase gradient difference to identify unwrapping risk points. The phase gradient and phase gradient difference can effectively determine where unwrapping failure occurs. When phase unwrapping failure occurs, these two values ​​will show obvious outliers. Therefore, a threshold method can be used for effective judgment. Compared with other methods to solve the unwrapping failure problem, this method does not require pre-training. Therefore, it does not require the collection of a large amount of data in advance, making it a lightweight algorithm.

[0044] 3. Due to the existence of phase unwrapping failure, the present invention must consider the phase adjustment value at the unwrapping failure point when constructing the unwrapped phase-position model to ensure that the entire objective function is minimized at the correct label position. Attached Figure Description

[0045] Figure 1 This is a flowchart of an RFID-based unwrapped phase-position model tag positioning method constructed according to a preferred embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of the structure of an RFID-based unwrapped phase-position model tag positioning system constructed according to a preferred embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0048] like Figure 1 As shown, an RFID positioning method based on an improved unwrapped phase-position model is presented. This method mainly includes the following steps:

[0049] like Figure 2 As shown in the diagram, the mobile robot, antenna, and RFID tag are included. The antenna is mounted on the mobile robot and moves with it. The RFID tag is positioned opposite the mobile robot and communicates with the antenna.

[0050] For the mobile robot, S1 has a bracket on top with multiple RFID antennas, facing 90 degrees in the robot's forward movement direction and toward the tag to be communicated with. The three-dimensional positions of the antennas in the robot's coordinate system are pre-calibrated. x,m c y,m c z,m ] TThe position of the m-th antenna in the robot coordinate system is obtained through calibration. The RFID antenna and the tag communicate, and the phase data during the communication is recorded. At the same time, the robot's odometer data is also recorded.

[0051] The number of RFID antennas is greater than or equal to 2.

[0052] S2 distributes the acquired phase data according to the EPC tag, and the data from each tag is processed separately. The data θ from each tag... (i) The unwrapping phase γ is obtained by processing the data using the phase unwrapping algorithm. (i) .

[0053] The specific mathematical expression for the untangling algorithm is:

[0054]

[0055] S3 calculates the robot's position using a trajectory estimation method based on sensor data from the mobile robot's wheeled odometer. Subsequently, the position of the RFID antenna is calculated using the positional relationship of the calibrated antenna relative to the robot's coordinate system and rotation / translation transformations. The mathematical expression for the calculation is:

[0056]

[0057] in, The robot's position was calculated using the trajectory estimation method, [x A,m y A,m z A,m ] T This represents the position of the m-th antenna in the world coordinate system established based on the robot's initial pose. Each moment corresponds to an antenna position and the corresponding tag unwrapping phase.

[0058] In dense tag localization scenarios, due to the close spatial location of multiple tags, they crowd each other's communication channels, causing situations where the spatial sampling theorem is not satisfied between two acquisitions when the moving antenna collects data from such dense tags. Existing unwrapping algorithms cannot correctly unwrap the original tags in such cases.

[0059] Therefore, the unwrapped phase signal needs to be checked first. The phase gradient is obtained by subtracting the unwrapped phase from the previous moment and dividing by the distance the antenna traveled between those two moments. The specific mathematical expression is:

[0060]

[0061] Where, Δd (i)Let G represent the displacement of the antenna from time i-1 to time i. Since i = 2, 3, ..., N, let G... (1) =0. Then, differentiating the phase gradient yields the phase gradient difference, the specific mathematical expression of which is:

[0062] δ (i) =G (i) -G (i-1)

[0063] These two features are used to classify unwound phase acquisition points.

[0064] Calculate the residual matrix of the linear regression of the phase gradient and the phase gradient difference.

[0065] ζ1=|GG|

[0066] ζ2=|δ-δ|

[0067] Where G and δ represent the linear fit estimates of the phase gradient and phase gradient difference, respectively.

[0068] By comparing with a set threshold: when season when season r = r1∪r2, where r represents the sequence matrix of possible untangling risks in the untangling phase, which contains the index of the untangling risk point.

[0069] The specific mathematical expression for transforming the untangling risk index matrix is ​​as follows:

[0070] c (j,i) =u(ir) (j) )

[0071] in, c (j,i) It is an element in matrix C, which is an adjustment representation matrix. It is a matrix with j rows and i columns. A certain row of C represents the representation of the untangling phase that needs to be adjusted due to the untangling risk point.

[0072] S5 substitutes the unwrapping risk point index parameter C, the unwrapping phase, and the antenna position into the improved unwrapping phase-position model, which is as follows:

[0073]

[0074] Where [x,y,z] is the position of the label, β m It refers to phase shift. This model leads to a nonlinear optimization problem, mathematically expressed as:

[0075]

[0076] In this nonlinear optimization problem, the phase adjustment value K, the label position [x,y,z], and the phase offset β are... m These are the parameters to be optimized. The LM algorithm is used to optimize the objective function, and the above-mentioned parameters to be optimized are finally obtained.

[0077] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An improved RFID-based unwrapped phase-position model tag positioning method, characterized in that, This method includes the following steps: S1 When the mobile robot moves with the antenna and communicates with the RFID tag, record the odometer data of the mobile robot over time, collect the communication signal between the antenna and the RFID tag and extract the phase data that changes over time in the communication signal, and match the phase data with the RFID tag one by one to obtain the phase data of each RFID tag at different times. S2 performs phase unwrapping on the phase data of each RFID tag to obtain the relationship between the unwrapped phase and time. Using the positional relationship between the antenna and the mobile robot and the odometer data of the mobile robot changing over time, the position of the antenna at different times is obtained. Combining the relationship between the unwrapped phase and time, the correspondence between the antenna position and the unwrapped phase at different times is obtained. S3 Using the correspondence obtained in step S2, solve for the phase gradient, phase gradient difference and the residual matrix corresponding to each of them. Using the residual threshold, determine the untangling risk point and the time corresponding to the untangling risk point according to the preset conditions. Number each untangling risk point and construct the untangling risk point sequence matrix accordingly. S4 Construct an optimization model with the sequence matrix of the unwrapping risk points, the unwrapping phase, and the antenna position as inputs, and the phase adjustment value, the RFID tag position, and the phase offset as outputs. Solve the optimization model to obtain the required output, and adjust the position of the RFID tag according to the output. In step S1, the optimization model is performed according to the following relationship: in, It is the objective equation. It is the unwrapping phase corresponding to the m-th antenna. This is the phase unwrapping risk matrix for the m-th antenna. This is the phase adjustment matrix for the m-th antenna. It is the wavelength corresponding to the operating frequency of the RFID system. , , They are the first m The x-coordinate, y-coordinate, and z-coordinate values ​​of the antenna. This represents the phase offset value of the m-th antenna. x , y , z These are the x, y, and z coordinates of the position of the label being solved. i It is the first i There are 10 sampling times, where N is the total number of phase sampling times for this antenna. j This is the first of the antennas. j There are several adjustment matrices, where W is the total number of adjustment matrices for the antenna.

2. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 1, characterized in that, In step S2, the phase unwrapping is performed according to the following relationship: in, , It is the untangling phase at the first moment. It is the original phase at the first moment. It is the untangling phase at time i. It is the original phase at time i. It is the untangling phase at time i-1. N represents the time of signal sampling, and N is the total number of sampling times for the signal.

3. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 1, characterized in that, In step S2, the positions of the antenna at different times are obtained according to the following formula: in, , , These are the x-coordinates, y-coordinates, and attitude angles of the robot's pose calculated using the trajectory inference method. , , These are the x, y, and z coordinates of the m-th antenna in the world coordinate system established based on the robot's initial pose. , , These are the x, y, and z coordinates of the position of the m-th calibrated antenna in the robot coordinate system.

4. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 1, characterized in that, In step S3, the phase gradient is calculated according to the following formula: in, It is from i -1 to i The displacement of the antenna between moments. It is the untangling phase at time i. It is the untangling phase at time i-1. It is the phase gradient at the i-th time step.

5. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 4, characterized in that, In step S3, the phase gradient difference is calculated according to the following formula: in, It is the phase gradient at time i. It is the phase gradient at time i-1. It is the phase gradient difference at the i-th time.

6. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 5, characterized in that, In step S3, the phase gradient residual matrix and the phase gradient difference residual matrix are calculated according to the following formulas: in, and These are the phase gradient and phase gradient difference corresponding to the phase data, respectively. and These are the linear fit estimates of the phase gradient and the phase gradient difference, respectively. and These are the phase gradient residual matrix and the phase gradient difference residual matrix, respectively.

7. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 1, characterized in that, In step S3, the preset conditions are performed as follows: when season Or, when season ; in, i It is the first i Each sampling time, and These are the indices of the unwrapping phases at time i, calculated based on the phase gradient and phase gradient difference, where potential unwrapping risks may occur. Each indices include the indices of the unwrapping risk points. T ereshold 1 and Thereshold 2 represents the set threshold values ​​for the residual matrices of the phase gradient and the phase gradient difference, respectively. and These are the phase gradient residual matrix and the phase gradient difference residual at time i, respectively.

8. The improved RFID-based unwrapped phase-position model tag positioning method as described in claim 7, characterized in that, The sequence matrix of the untangling risk points is as follows: in, It is a sequence matrix of possible untangling risks in the untangling phase, containing indices of untangling risk points. and These are the sequence matrices of possible untangling risks in the untangling phase, calculated based on the phase gradient and the phase gradient difference.

9. A system for the tag positioning method according to any one of claims 1-8, characterized in that, The system includes a signal generation module, a signal acquisition module, a signal processing module, a risk point judgment module for untangling, and an optimization model module, wherein: The signal generation module includes a mobile robot, an antenna, and an RFID tag. The antenna is mounted on the mobile robot and moves with it. The RFID tag is positioned opposite the mobile robot and communicates with the antenna. The signal acquisition module is used to acquire the odometer data of the mobile robot that changes over time, as well as the communication signal between the antenna and the RFID tag, and to map the RFID tag to the phase one by one. The signal processing module is used to unwrap the phase, obtain the correspondence between the antenna position and the unwrapped phase at different times, and then calculate the phase residual corresponding to the phase gradient and the phase gradient difference. The untangling risk point judgment module is used to judge untangling risk points according to preset conditions; The optimization model module is used to adjust the position and phase of the untangling risk point using the constructed optimization model.

Images