Grain moisture sensing method and system based on RFID signal

Through the grain moisture perception method based on RFID signals, using the concept of equivalent dielectric properties, the problem of time-consuming and low accuracy of grain moisture detection in the existing technology is solved, and more accurate moisture perception and detection efficiency are improved.

CN118914302APending Publication Date: 2024-11-08KAIFENG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410982256.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The prior art has the problem of time-consuming, expensive and inconvenient use in food moisture detection, and machine learning-based methods are difficult to reveal the perception mechanism of wireless signals, and the recognition accuracy is limited by sample size.

Method used

Using the grain moisture perception method based on RFID signal, by building a physical perception platform, establishing the relationship between label shape variables and equivalent dielectric characteristics, determining the critical value of volume density, calculating the resonance frequency, mapping the grain moisture, and proposing equivalent dielectric The concept of nature, breaking away from the dependence on food body density and achieving more accurate moisture perception.

Benefits of technology

It realizes more accurate perception of food moisture changes, reveals the essential problems of wireless sensing, simplifies the model, gets rid of the dependence on food body density, and improves detection accuracy and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118914302A_ABST
    Figure CN118914302A_ABST
Patent Text Reader

Abstract

The invention discloses a grain moisture sensing method and system based on an RFID signal, which is applied to the technical field of grain moisture sensing, and comprises the following steps: constructing a physical sensing platform; initial modeling is carried out on the parameters influencing the tag performance, simplification is carried out, and the relationship among tag deformation quantity, equivalent dielectric characteristics and tag resonant frequency is obtained; determining a volume density critical value according to the relationship among the label deformation quantity, the increment of the equivalent dielectric property and the resonant frequency; under the condition that the volume density is greater than the volume density critical value, calculating to obtain the current resonant frequency of the tag; and solving the equivalent dielectric characteristic by utilizing the relation among the current deformation quantity and the equivalent dielectric characteristic of the label and the current resonant frequency of the label, and mapping the grain moisture. According to the method, the grain density is larger than the critical value of the body density, the equivalent dielectric property does not change along with the grain density, the dependence on the grain body density rho is eliminated, and the change of target moisture is sensed or mapped more accurately.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of grain moisture sensing, and more particularly to a grain moisture sensing method and system based on RFID signals. Background Art

[0002] Moisture is one of the important components of stored grains. The moisture content determines the vitality of grains and the direction of their life activities. Problems such as temperature, insect pests and mold during grain storage are all related to the moisture content of grains. Changes in moisture will also lead to changes in the physical properties of grains, such as hygroscopicity, conductivity and dielectric properties. Therefore, the development of an effective grain moisture detection solution is an urgent problem to be solved in the industry.

[0003] Traditional methods for detecting grain moisture, such as drying, capacitance, resistance, microwave and neutron meter, are time-consuming, expensive and inconvenient to use. In recent years, radio frequency (RF) technology has been used to detect moisture, temperature and mildew in stored grain. For example, a recent research result combined a multi-class SVM method and used tag impedance to sense the temperature and moisture of wheat. Although the impedance method explains the intrinsic reasons for signal changes to a certain extent, it still lacks theoretical support in terms of the perception mechanism.

[0004] Most existing technologies use machine learning methods to perceive targets, which usually requires a large number of data samples for training, and its recognition accuracy is limited by the sample size. In addition, the "black box" method based on machine learning usually cannot reveal the perception mechanism of wireless signals.

[0005] Therefore, how to more accurately sense or map the change of target moisture is an urgent problem to be solved by those skilled in the art. Summary of the invention

[0006] In view of this, the present invention provides a grain moisture sensing method and system based on RFID signals to solve the problems in the background technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] On the one hand, the present invention discloses a grain moisture sensing method based on RFID signals, comprising:

[0009] Build a physical perception platform;

[0010] Initially model the parameters that affect the tag performance and simplify them to obtain the relationship between the tag deformation and equivalent dielectric properties and the tag resonant frequency;

[0011] Determine the critical value of volume density according to the relationship between the tag deformation amount, the increment of equivalent dielectric characteristics and the resonant frequency;

[0012] Under the condition that the body density is greater than the body density critical value, the current resonant frequency of the tag is calculated;

[0013] The relationship between the current deformation of the tag and the equivalent dielectric properties and the current resonant frequency of the tag is used to solve the equivalent dielectric properties and map the moisture content of the grain.

[0014] Preferably, in the above-mentioned grain moisture sensing method based on RFID signals, the parameters affecting the tag performance are initially modeled, and the formula is as follows:

[0015]

[0016] Where c is the speed of light, which is a constant in air under standard atmospheric pressure, υ is the Poisson's ratio, which ranges from 0<υ<1, L is the label length, L" is the compensation value of the label length caused by the edge field effect, and ε r The relative dielectric constant of the label material property is a constant, H and W are the thickness and width of the label medium respectively, and s represents the strain force, which represents the food pressure on the label and has no unit dimension.

[0017] Preferably, in the above-mentioned grain moisture sensing method based on RFID signals, the relationship between the tag deformation amount and the equivalent dielectric property and the tag resonant frequency is obtained as follows:

[0018]

[0019] f n is the tag resonant frequency, L n , ε eq,n Respectively represent the strain s n Length and equivalent dielectric properties of the lower tab.

[0020] Preferably, in the above-mentioned grain moisture sensing method based on RFID signals, the specific steps of determining the critical value of bulk density are as follows:

[0021] For the n+1th state, we have:

[0022]

[0023] Among them, ΔL represents the deformation of the label due to the strain s or the increase in body density ρ, Δε eq represents the increment of equivalent dielectric properties;

[0024] f n+1 is the resonant frequency obtained after a certain amount of food is filled into the container and pressed; ε eq,n is the equivalent dielectric property at the current volume density;

[0025] When Δε eq= 0, we get the volume density ρ min , then we think that ρ min This is the critical value of volume density.

[0026] Preferably, in the above-mentioned grain moisture sensing method based on RFID signal, the specific steps of calculating the current resonance frequency of the tag are:

[0027] Construct a mathematical model between the tag resonant frequency and body density under critical state:

[0028]

[0029] Among them, f ρ is the resonant frequency of the tag when the body density is ρ, f0 is the tag in the initial state, in air, when the force s = 0 and ρ = 0, a, b, c are model parameters determined by linear regression from experimental data.

[0030] On the other hand, the present invention discloses a grain moisture sensing system based on RFID signals, which applies the above method and includes: a tag, an RFID antenna, a container, an analog VNA device and a computer; the tag is attached to the inner wall of the container and is in contact with the wheat; the RFID antenna is connected to the analog VNA device to send and receive electromagnetic wave signals; when the tag is excited, it will generate a signal in reverse to the space, the signal can be received by the RFID antenna, and sent to the computer through the analog VNA device.

[0031] Preferably, the above-mentioned grain moisture sensing system based on RFID signals further includes: absorbing material; the absorbing material is pasted on the inner wall of the container to absorb electromagnetic waves transmitted by the label and electromagnetic waves reflected by the cross section of the grain.

[0032] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a grain moisture sensing method and system based on RFID signals, proposes a paradigm shift in wheat moisture sensing, from a pattern-based method to a mathematical model-based method, thereby better revealing the essential problems of wireless sensing; proposes an equivalent dielectric property (i.e., ε eq ), which is different from the complex form of dielectric constant, which can be used as a real number to characterize the moisture content of grain. eq It can effectively simplify the model; study the effect of grain density (ρ) on the equivalent dielectric property ε eq The influence of ε eq A density-independent algorithm is proposed based on the relationship model between the grain and the moisture content of the grain. The algorithm makes the equivalent dielectric properties of the grain that is greater than the critical value of the bulk density unchanged with the grain density, thus getting rid of the dependence on the grain bulk density ρ and more accurately sensing or mapping the changes in the target moisture. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0034] Figure 1 A schematic diagram of the change of dielectric constant with frequency provided by the present invention;

[0035] FIG. 2( a ) is a multi-physics simulation model of an RFID provided by the present invention;

[0036] FIG2( b ) is a schematic diagram of the simulation model results provided by the present invention;

[0037] Figure 3 A schematic diagram of the relationship between the dielectric constant and the volume density provided by the present invention;

[0038] FIG4( a ) is a schematic diagram of phase and external force provided by the present invention;

[0039] FIG4( b ) is a schematic diagram of frequency and external force provided by the present invention;

[0040] FIG. 4( c ) is a diagram of S provided by the present invention. 11 Schematic diagram of external forces;

[0041] Figure 5 The standardized frequency versus volume density curve provided by the present invention;

[0042] Figure 6 A flow chart of the method provided by the present invention;

[0043] Figure 7 A schematic diagram of a three-layer propagation model provided by the present invention;

[0044] Figure 8 A schematic diagram showing outliers being removed after preprocessing provided by the present invention;

[0045] Fig. 9 The variation of Δf provided by the present invention within different ρ value ranges;

[0046] Figure 10 (a) shows the ε of wheat with different moisture content provided by the present invention. eq Relationship curve with ρ;

[0047] Figure 10(b) shows the relationship between wheat moisture and ε provided by the present invention. eq relationship curve. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] The embodiment of the present invention discloses a grain moisture sensing method based on RFID signals, such as Figure 6 As shown, including:

[0050] Build a physical perception platform;

[0051] Initially model the parameters that affect the tag performance and simplify them to obtain the relationship between the tag deformation and equivalent dielectric properties and the tag resonant frequency;

[0052] Determine the critical value of volume density according to the relationship between the tag deformation amount, the increment of equivalent dielectric characteristics and the resonant frequency;

[0053] Under the condition that the body density is greater than the body density critical value, the current resonant frequency of the tag is calculated;

[0054] The relationship between the current deformation of the tag and the equivalent dielectric properties and the current resonant frequency of the tag is used to solve the equivalent dielectric properties and map the moisture content of the grain.

[0055] What you need to know is:

[0056] In the actual process of measuring grain moisture, if the label contacts the grain, the label will be deformed due to the extrusion of the grain. The label will be deformed under the pressure of the grain, and the deformation of the label will affect its reflection signal, which is one of the factors affecting the stability of the system. In addition, the bulk density of the grain also affects the reflection signal of the label.

[0057] The present invention cleverly captures the relationship between volume density, strain force and dielectric properties, and derives a mathematical model to obtain a stable ε eq The proposed mathematical model-based approach to grain moisture perception is a paradigm shift from the pattern-based approach in previous work, and its accuracy does not depend on the number of data samples and can better reveal the perception mechanism problem.

[0058] Specifically, the dielectric constant, also known as the dielectric constant or relative dielectric constant, is an important feature that characterizes the electrical properties of dielectric or insulating materials. It refers to the ratio of the capacitance when the same material is used as a dielectric in the same capacitor to the capacitance in a vacuum. Its value indicates the relative ability of the dielectric to store electrostatic energy in an electric field. The dielectric constant is usually a complex number and can be defined as:

[0059] ε=ε'+jε”; (1)

[0060] Among them, ε' refers to the real part of the dielectric constant of the material, which represents the ability to store electric fields. ε" refers to the imaginary part of the dielectric constant, which represents the ability to dissipate electrical energy in the form of heat. For grains, moisture content is a key factor affecting their dielectric properties. The dielectric properties of grains in an electromagnetic environment are significantly correlated with their moisture content.

[0061] The relationship between the dielectric constant and frequency of wheat with different moisture contents, Figure 1 The figure shows the relationship between the dielectric constant of grains with three different moisture contents and the frequency of electromagnetic waves. It can be seen from the figure that the real part of the dielectric constant (ε') changes with frequency, that is, ε' decreases with increasing frequency, and ε' tends to be stable with further increasing frequency; on the other hand, ε" does not change much with increasing frequency.

[0062] The above analysis shows that if the frequency is fixed, both the real and imaginary parts of the dielectric constant will change with the change of moisture content, and the greater the moisture content, the greater the dielectric constant. This means that the dielectric constant has a strong correlation with both moisture content and frequency. The present invention establishes the relationship between frequency and ε through a mathematical model. eq (related to the dielectric constant), and then use ε eq Characterize the moisture content of wheat.

[0063] In fact, the dielectric constant is a complex number, which is difficult to obtain accurately using RFID experimental equipment. In order to solve this problem, the present invention proposes an equivalent dielectric property ε eq concept, which sets it to a real number to approximate the equivalent complex dielectric constant. It will be shown that ε eq There is a strong correlation between the moisture content of wheat. Therefore, we do not need to worry about the exact dielectric constant (real and imaginary parts), but only need to obtain the ε that is closely related to moisture. eq The moisture content of grain can be estimated.

[0064] Furthermore, since the dielectric property is closely related to the moisture content of wheat, the present invention uses the electromagnetic module of HFSS (High-Frequency Structure Simulator) to study the relationship between the dielectric constant and the resonant frequency of the tag.

[0065] What you need to know is that HFSS is Ansoft's three-dimensional electromagnetic simulation software. This simulation software can calculate the S parameters and full-band electromagnetic fields of three-dimensional passive structures of any shape. It is a powerful electromagnetic field simulation software that can be used to simulate and analyze high-frequency electromagnetic field problems.

[0066] The relationship between the resonant frequency of the RFID tag and its surrounding medium (such as wheat) is simulated in various electromagnetic environments. The simulation model of the RFID tag is shown in Figure 2(a). The purpose of the simulation is to prove that the dielectric constant (taking the real part as an example) affects the resonant frequency of the tag, that is, the dielectric constant changes with the resonant frequency of the tag.

[0067] The settings during the simulation are as follows: According to the label manual parameters, the label chip impedance is set to Z c =(28.3-j204.2)Ω, assuming that the RFID tag is attached to the boundary between the medium (wheat) and the air, where the boundary condition is to define the characteristics of the electromagnetic field across the discontinuous boundary, assuming that there is no gap between the tag and the medium (wheat), the dielectric constant of the box is idealized as air, and the external space of the box is assumed to be a perfectly matched layer. According to the experimental requirements and combined with the actual situation, the boundary condition is set to the radiation boundary (Radiation). Excitation is an excitation source defined on the surface of a three-dimensional object or a two-dimensional object. The excitation port is a special boundary condition type that allows energy to enter or flow out of a geometric structure. In this chapter, the excitation is set to the wave port mode to obtain ideal simulation results. The solution type is set to the mode-driven solution type, the adaptive mesh is set, the fast sweep mode is used, and the frequency range is set to 860MHz~960MHz. The dielectric constant (real part) is set to 2.3 to 2.75, which is consistent with the dielectric constant corresponding to the common moisture content of wheat. When ANSYS HFSS is running, measure and record S 11 The parameter is the input reflection coefficient of the input return loss. The simulation results are shown in Figure 2(b), where the inflection point in the smooth curve represents the tag resonant frequency. The figure shows that ε' decreases as the resonant frequency increases from 860MHz to 960MHz.

[0068] Furthermore, according to the relationship between dielectric constant and density, such as Figure 3 As shown in the figure, the dielectric constant of grain varies with its bulk density. 3 ~0.86g / cm 3 In the range of , ε' increases with the increase of density. The three curves in the figure correspond to three different moisture contents of 17.5%, 14.5% and 11.1%, which indirectly illustrates the characteristic that the dielectric constant changes with the moisture content of grain. In order to solve the density dependence problem, the present invention proposes a density-independent calibration function to obtain a stable characteristic parameter, which can be used to robustly sense the moisture content of wheat.

[0069] According to electromagnetic wave theory, when there are objects smaller than the wavelength on the propagation path of electromagnetic waves, and the number of such obstructing objects per unit volume is very large, the signal will be scattered. Scattering occurs on rough surfaces, small objects or other irregular objects. Grains are "granular" objects, and their diameters generally range from 0.3CM to 1.0CM. The frequency range used in this experiment is 860MHz to 960MHz, so the wavelength is about 36.8CM to 31.2CM. It can be seen that the wavelength of electromagnetic waves is much larger than the diameter of grain particles.

[0070] Therefore, electromagnetic waves will scatter in the grain. However, as the density of grain increases, the internal voids become smaller and smaller, the "granular" characteristics of the grain become less and less obvious, and the scattering phenomenon weakens. Therefore, at this time, this medium (grain) is closer to a uniform medium. When electromagnetic waves are vertically incident on the surface of the medium, there are two main propagation modes, namely reflection and transmission, and the electromagnetic wave signal tends to be stable. Based on the above analysis, it is believed that when the volume density exceeds 0.86g / cm 3 Or higher, there should be a critical value that makes the dielectric constant stable, that is, the dielectric constant no longer changes when the density exceeds the critical value. However, it is difficult to reach the critical value in the laboratory. In fact, the purpose of this chapter is not to obtain the critical value of density, but to determine the dielectric constant of wheat at the critical density value, which can be used to map the moisture content information of grain.

[0071] In order to further optimize the above technical solutions, in the actual application of Tag2Model, the label will be deformed due to the extrusion of grain, which will affect the reflected signal of the label. The label is attached to the inner side of the container wall, and the label is in direct contact with the grain, which is more conducive to Tag2Model to capture the label reflection signal with rich grain information. The label is usually deformed due to the extrusion force of the wheat pile. In order to study how the deformation of the label affects the signal, the experiment evaluated the changes in three factors when the label is deformed, namely phase, frequency and S parameters. In order to simulate the label extrusion phenomenon, the following label bending experiment was carried out. The label is attached to a printed circuit board (dielectric substrate FR4, this material has little effect on electromagnetic signals), and the antenna is placed directly above the label. The substrate is gradually compressed by an extruder to simulate the process of the label being squeezed and deformed by grain to observe the phase, frequency and S parameters of the tag signal. 11 The phase can be expressed as follows:

[0072]

[0073] Where d is the distance between the tag and the RFID antenna, λ air is the wavelength of the electromagnetic wave in the air. For a fixed frequency electromagnetic wave λ airis a constant. The above equation shows that the phase is a function of the distance d from the tag to the RFID antenna. As the substrate is constantly bent toward the RFID antenna to form an "arch" (maximum distance is about 1.5CM), the distance d is changing, so the phase Φ must change. In order to eliminate the effect of d on the phase, the height of the extruder is adjusted in real time during the experiment (the adjustment process is called distance compensation), the purpose is to ensure that the distance between the center of the tag and the antenna remains unchanged regardless of how the tag is deformed.

[0074] Through the above settings, the changes of various characteristics during the deformation of the label are gradually observed and recorded. In the experiment, each time the extruder is rotated one circle, the external extrusion force level on the label is expressed as s(i). A total of 10 circles are rotated, then i = 1, 2, ... 10. The results are as follows Figure 4(a)-4(c) As shown. Among them, Figure 4(a) shows the change of RFID phase before and after distance offset when the tag is deformed due to pressure. Before distance compensation, the phase changes with the pressure, while after distance compensation, the phase remains almost unchanged. Figures 4(b) and 4(c) respectively show the changes of frequency and S parameters during the deformation of the tag due to external force. It can be seen from the figure that the tag frequency changes with the change of pressure: the greater the deformation of the tag, the lower the frequency. Figure 4(c) plots the S 11 The curve with pressure level directly shows the effect of label deformation on S 11 impact.

[0075] In summary, the deformation of the label caused by the squeezing force has little effect on the label phase, but has a significant effect on the frequency and S parameters.

[0076] Furthermore, the parameters that affect label performance are initially modeled as follows:

[0077]

[0078] Where c is the speed of light, which is a constant in air under standard atmospheric pressure, υ is the Poisson's ratio, which ranges from 0<υ<1, L is the label length, L" is the compensation value of the label length caused by the edge field effect, and ε r The relative dielectric constant of the label material property is a constant, H and W are the thickness and width of the label medium respectively, and s represents the strain force, which represents the food pressure on the label and has no unit dimension.

[0079] Furthermore, the influence of tag deformation caused by pressure s on the signal resonant frequency. Since tag deformation is usually small and difficult to measure, pressure s can be used to represent tag deformation. In engineering, pressure change s is usually defined as the relative change in length when an object is deformed, as shown in formula (5), where L is the original length of the tag, that is, the state when no force is applied, and ΔL represents the change in the length of the tag after force is applied to the tag.

[0080]

[0081] Define the state of the tag in the air as the initial state, at which the pressure is zero, that is, s = 0. Let f0 represent the resonant frequency of the tag in the initial state. The resonant frequency of the tag under force s can be expressed as:

[0082] f=f0×(1-s) (6)

[0083] Define the normalized resonant frequency Δf as:

[0084]

[0085] From equation (7), we can know that Δf is inversely proportional to force s. The preliminary research conclusion is that force s is related to the resonant frequency f, and the volume density (ρ) is related to s (assuming that the volume of the grain container is constant). Therefore, ρ should be related to f. Based on the above analysis, a volume density model is defined to study the relationship between volume density and resonant frequency f. Δf can be redefined as shown in equation (8), where f ρ and ρ Respectively represent the resonant frequency and pressure of the tag under the body density ρ. Note: f ρ Can be read by VNA equipment and is a known quantity.

[0086]

[0087] Taking a wheat sample with a moisture content of 7.7% as an example, the relationship between the standardized resonant frequency of the tag and the volume density at different volume densities was tested. Figure 5 As shown. It is observed that there is an obvious quadratic linear relationship between ρ and Δf(ρ), where R 2 =0.99(R 2 Indicates the goodness of fit, R 2 =1.0 indicates the best fit), and the fitting curve is shown in model (9):

[0088] Δf(ρ)=0.026×ρ 2 -0.041×ρ+0.015 (9)

[0089] In the laboratory, the value of ρ cannot be higher than 0.84 g / cm 3 . However, as ρ increases further, Δf(ρ) should converge to a relatively stable value. In other words, the above model may have a critical value x, so that when ρ>x, Δf(ρ) tends to be stable and unchanged;

[0090] There are two main factors that affect the resonant frequency: the deformation of the tag due to force and the properties of the material to which the tag is attached, represented by L and ε respectively. eq Indicated by L n , εeq,n Respectively represent the strain s n The length and equivalent dielectric properties of the lower label. Since L' can usually be ignored, model (4) can be simplified to:

[0091]

[0092] For the n+1th state, we have:

[0093]

[0094] Where ΔL represents the deformation of the label due to the strain s or the increase in volume density ρ (unitless), Δε eq represents the increment of the equivalent dielectric property. Models (10) and (11) are two observed values ​​of the resonant frequency in two adjacent states (which can be read by VNA). It can be concluded that ρ or s affects f through the following two factors, namely ΔL and Δε eq However, when the density ρ increases to a certain size, the resonant frequency f tends to be stable.

[0095] further, Figure 1 represents the relationship between volume density and dielectric constant. It can be seen from the figure that the dielectric constant increases with the increase of volume density. In order to eliminate the dependence of dielectric constant on volume density, a volume density-independent algorithm is proposed. This algorithm determines ε eq The critical value that does not change with ρ. Let f n represents the resonant frequency at the current volume density, f n+1 It is the resonant frequency obtained after a certain amount of food is filled into the container and pressed. ε eq,n is the equivalent dielectric property at the current volume density, Δε eq The increment of the equivalent dielectric properties after the volume density increases. eq = 0, we get the volume density ρ min , then we can consider ρ min is the critical value, which makes ε eq The minimum value that does not change with the volume density. Let Δf = f n+1 -f n represents the observed value of two adjacent resonant frequencies, where f n+1 and f n can be read from the analog VNA / reader device, then the following procedure is used to prove that there is a critical value ρ min , so that ε eq Does not vary with ρ:

[0096] According to formula (10) and formula (11), we have:

[0097]

[0098] Applying Taylor series expansion, we get:

[0099]

[0100] Due to Δε eq <<ε eq,n ,Right now According to the definition of Taylor's formula, for the high-order terms on the right side of formula (13), It can be approximately equal to 0, retaining only the first two terms, namely: Substituting it into equation (12), the difference Δf between two adjacent resonant frequencies can be approximated as:

[0101]

[0102] According to the definition of formula (5), we have: From formula (14), we can find Δε eq as follows:

[0103]

[0104] When Δε eq = 0, which means that the equivalent dielectric property no longer increases, then ε eq,n Approaching stability, as follows:

[0105]

[0106] Among them, 0 <f n+1 <f n ,Right now This means that as the density or pressure increases, the resonant frequency shifts to the left (decreases). In fact, after filling the container with more grain several times (i.e. slowly increasing ρ), it can be observed that the resonant frequency f tends to stabilize. For two consecutive observations f n+1 、f n , when f n+1 ≈f n When , we can get s from formula (16): n = 0. At this point, the tag resonant frequency is no longer "sensitive" to the density ρ. Therefore, Δf(ρ) = 0. Substituting into model (9), solve the equation: Δf(ρ) = 0.026×ρ 2 -0.041×ρ+0.015=0, we get ρ * =1.05. For model (9), when ρ≥1.05, ε eq It does not change with the change of body density. Without loss of generality, a mathematical model (17) between the tag resonant frequency and body density under the critical state is constructed:

[0107]

[0108] Among them, f ρ is the resonant frequency of the tag when the body density is ρ, and f0 is the tag in the initial state (in air, when force s = 0 and ρ = 0). a, b, c are model parameters, which can be determined by linear regression from experimental data.

[0109] Specific experimental process: First, assume that the volume of the grain container is fixed and constant, and gradually add a certain amount of grain (for example, 100 grams each time). Then the bulk density of the grain increases from 0.65 g / cm in the natural state to 3 Gradually increase to 0.84g / cm 3 . Measure and record the resonance frequency f at different ρ ρ Then, the experimental data is used to regress the mathematical model (17) to solve the parameters a, b, and c. Finally, the density-independent ε eq It can be solved by formula (10).

[0110] The flowchart of the proposed Tag2Model algorithm is shown in Figure 6 First, the critical value ρ is solved using the volume density-independent algorithm * , corresponding to Δε eq = 0. Then, regression analysis is used to solve model (17) to obtain f ρ , and then find ε eq , and is used to map grain moisture. Observe the analytical model (17), if ρ→+∞ and b<0, then “a×ρ b " is a very small number and can be approximately equal to 0. Therefore, model (17) can be simplified as follows:

[0111]

[0112] Model (18) shows that when the body density is large enough, the resonant frequency f ρ It is only related to f0 and parameter c. f0 is the resonant frequency of the tag in air and is a constant. Parameter c can be obtained by fitting the regression equation with the experimental data.

[0113] On the other hand, the present invention discloses a grain moisture sensing system based on RFID signals, which applies the above method and includes: a tag, an RFID antenna, a container, an analog VNA device and a computer; the tag is attached to the inner wall of the container and is in contact with the wheat; the RFID antenna is connected to the analog VNA device to send and receive electromagnetic wave signals; when the tag is excited, it will generate a signal in reverse to the space, the signal can be received by the RFID antenna, and sent to the computer through the analog VNA device.

[0114] Specifically, in the Tag2Model system setting, the tag is attached to the inner side of the container wall and is in direct contact with the wheat sample. The container is made of polyethylene material and is almost completely transparent to electromagnetic waves. Therefore, the obstruction of the container to the RFID signal can be ignored. When the container is filled with wheat, the tag may be slightly deformed due to strain. The concept of equivalent dielectric properties is proposed, and a prototype model of frequency, tag length and equivalent dielectric properties is designed, and then the relationship between frequency and density is derived. Introduction The design of Tag2Model includes: three-layer propagation model, density-independent algorithm design, modeling of the relationship between resonant frequency and equivalent dielectric properties, and data preprocessing. The purpose is to obtain a stable ε eq The value of the sensor is freed from the dependence on the grain density ρ and can more accurately sense or map the change of target moisture. Figure 7 shown.

[0115] When an RFID tag is attached to or close to a rough target (granular material, such as wheat), the interaction between the electromagnetic field and the tag involves scattering, refraction, reflection, etc. These phenomena can be explained by the basic principles of electromagnetic fields.

[0116] Tiny irregular structures on a rough surface can cause electromagnetic waves to scatter, and these scattering points can be bumps or depressions on the rough surface. When the RFID tag is pressed against a rough target, the electromagnetic waves interact with the irregular structures on the surface, resulting in scattering. This may cause the received signal to have components of multiple paths, affecting the amplitude and phase of the signal; Refraction: Refraction refers to the phenomenon that light changes direction when passing through the surface of a medium due to the different refractive index of the medium. If there are medium changes on the surface of a rough target, the RFID signal may be refracted on the target surface, changing the path and direction of signal propagation; Reflection: Reflection is the interaction between electromagnetic waves and the surface, and part of the energy returns to space. When an RFID signal encounters a rough surface, part of the energy will be reflected back into space. The intensity and angle of reflection may be affected by the surface roughness and characteristics. However, there are three attempts to eliminate or weaken the scattered and refracted signals.

[0117] First, the container is changed from organic glass to a softer polyethylene material, which has a negligible effect on the propagation of electromagnetic waves. In other words, the polyethylene material can be regarded as air, so that the complexity of the tag signal is reduced when it is transmitted through the container because there is one less layer of medium; second, the label is attached to the inside of the container so that the label is in close contact with the food, in order to reduce the scattering effect of the electromagnetic waves reflected by the label; third, the absorbing material is attached to the back of the container, which can effectively reduce the impact of signal refraction.

[0118] Specifically, the Tag2Model system consists of a passive tag (Alien Higgs 3 chip, model 9640), an RFID directional antenna (Laird2 model S9028 PCR, directional angle 65 degrees), a simulated vector network analyzer VNA model TektronixTTR506A, power 10dB), a tablet computer and a piece of absorbing material (Fe-Si-Al component). This embodiment uses a polyethylene container with a size of 31.5cm×18cm×18cm and is filled with wheat samples. The tag is attached to the inner wall of the box and is in contact with the wheat. The RFID antenna is connected to the simulated VNA device to send and receive electromagnetic wave signals. When the tag is excited, it will generate signals in the space in reverse, which can be received by the RFID antenna and then sent to the computer through the simulated VNA device for further analysis and processing. The resonant frequency was measured using a non-contact processing method. The reflection parameter S is measured by the simulated VNA, and the frequency corresponding to the maximum negative peak of S is obtained, which is the resonance of the tag.

[0119] In order to further optimize the above technical solution, it also includes: absorbing material; the absorbing material is pasted on the inner wall of the container to absorb the electromagnetic waves transmitted by the label and the electromagnetic waves reflected by the cross section of the grain.

[0120] The multipath problem in radio frequency identification technology (RFID) is caused by the signal encountering multiple paths during propagation, resulting in the receiver receiving multiple signals from different paths. This may cause problems such as signal attenuation and phase delay, affecting the performance of the RFID system. In multipath propagation, signals from different paths may meet at the receiver, resulting in signal superposition and interference. In order to reduce the impact of the multipath problem on the system: use directional antennas to reduce the propagation of signals on different paths; control the transmission power of the RFID system to keep it within an appropriate range; attach absorbing materials to the inner wall of the container to prevent signal transmission and scattering, which can effectively reduce the invalid reflection signals caused by the grain cross section. However, there are still abnormal data samples that need to be processed in the time domain and frequency domain.

[0121] It should be noted that the absorbing material is pasted on the inner wall of the container. Its main purpose is to absorb the electromagnetic waves transmitted by the label and reflected by the cross-section of the grain to obtain a "cleaner" signal, which is convenient for improving the performance of the system. Figure 8 The RFID data samples before and after outlier preprocessing (i.e., the magnitude of the S parameters) are shown. This method effectively removes outliers from the signal.

[0122] Model validation and analysis:

[0123] The ultimate goal is to expect ε eqAccurately characterize the moisture information of wheat. In order to eliminate the influence of temperature on Tag2Model, the experiment was carried out at room temperature of 20℃. Using experimental data, f ρ and ρ, where the parameters a, b, and c can be obtained by regression. It can be concluded that when ρ is satisfied * When ε≥1.05 eq It has nothing to do with bulk density. Taking wheat with a moisture content of 7.7% as an example, using experimental data to ρ and ρ to model and inversely derive ε eq .

[0124] First, when there is no wheat in the container, the resonant frequency f0 of the tag is read and recorded by the VNA device (f0 = 926MHz when the distance in the air is constant); secondly, the wheat sample is filled into the container until the container is full. The bulk density ρ is obtained by weighing and calculating, and the resonant frequency f of the tag is read and recorded by the VNA device. ρ Here, 100 grams of wheat is evenly compacted and placed in a container. This process is repeated 9 times to obtain the wheat density sequence {ρ n}, where n = 1, 2, ..., 9, and the bulk density range is 0.65 g / cm 3 ~0.84g / cm 3 , the corresponding resonant frequency sequence can be recorded as {f ρn}, where n = 1, 2, ..., 9; Finally, the fitting tool in the MATLAB toolbox is used to obtain the parameters a, b, and c, where a = 0.6068 × 10 -3 , b = -16.63, c = -1.633. Therefore, the resonant frequency f ρ The mathematical model of volume density ρ is expressed as follows:

[0125] f ρ =f0+0.6068×10 -3 ×ρ -16.63 -1.633 (19)

[0126] For a wheat sample with 7.7% moisture, the relationship between Δf and ρ is as follows: Fig. 9 As shown in the blue solid line, the range of ρ is 0.65 g / cm 3 ~0.84g / cm 3 , R 2 =0.968. Among them, R 2 =1 means perfect fit, R 2 To verify the validity of model (17), the ρ value range of 0.84 g / cm is plotted by the red dashed line in Figure 11. 3 ~1.05g / cm 3The figure shows that when ρ is greater than 0.84 g / cm 3 (theoretical value should be 1.05), Δf remains stable (i.e. f ρ This result verifies that when ρ is large enough, f ρ The conclusion that f tends to a constant is then used to ρ Combining formula (10) we can get the stable ε eq .

[0127] Table 1 Fitting parameters of the model at different water contents

[0128]

[0129] It is difficult to achieve the condition ρ>1.05 or higher density values ​​in a laboratory environment, such as Fig. 9 However, this does not affect the rationality of this method. In fact, TagModel does not care about the value of ρ, as long as it can obtain ε that does not change with ρ. eq Therefore, the effectiveness of the simplified model 18 is verified next when ρ→+∞, as shown in Table 1, which shows the parameters a, b, and c and R of 10 different wheat moisture samples based on model 17. 2 The regression results of R 2 All of them are above 0.9, which indicates that the regression results have high accuracy. ρ=1.05 and f ρ→∞ They represent the resonant frequencies of the tag when ρ = 1.05 and ρ = +∞, respectively. It can be seen from the table that the resonant frequencies are not much different when ρ = 1.05 and ρ = +∞, which verifies the validity of the simplified model (18) and also verifies that f ρ Or Δf is only related to f0 and parameter c in the limit state. Therefore, ε eq is a variable independent of the bulk density ρ.

[0130] The above verifies the effectiveness of the proposed model, which aims to use the equivalent dielectric ε eq Mapping wheat moisture content, ε eq It is obtained from the mathematical model (17), where L n Determined by L+ΔL, ΔL is the length compensation of L, which can usually be replaced by L", calculated by formula (4), L is the length of the label 12.31cm. Then ε eq It can be solved by formula (10). Figure 12(a) shows the equivalent dielectric characteristic curves of 10 kinds of wheat with different moisture content. Among them, L1~L10 represent the moisture sequence of {7.7%, 8.6%, 9.3%, 10.6%, 13.6%, 14.2%, 14.9%, 15.4%, 16.5%, 17.2%}. When ρ is greater than 1.05, ε eqtends to a stable value, which is close to the theoretical value, indicating that it is independent of the bulk density ρ.

[0131] It is not difficult to find from Figure 10(a) that when ρ < 0.9, the curves corresponding to the moisture sequences {L2, L3}, {L5, L6}, and {L8, L9, L10} have intersections, which means that Tag2Model may obtain incorrect ε eq , and will affect the accuracy of the sample moisture judgment. When ρ>1.05, there is no intersection between the curves, and all curves tend to be stable in this range (in fact, when ρ>0.9, these curves have already tended to be stable). Therefore, it can be concluded that when the wheat body density ρ is large enough, the equivalent dielectric property ε eq Has a certain stability, ε eq It can be used as a stable indicator for mapping wheat moisture. Based on Figure 10(a), taking ρ = 1.05 to draw Figure 10(b), it shows that ε eq There is an obvious linear relationship between wheat moisture and R 2 =0.85, wheat moisture changes with ε eq The dielectric constant is closely related to the moisture content of wheat, which proves the validity of the model proposed in this invention. It is believed that the proposed equivalent dielectric property ε is independent of the bulk density. eq It can be used as a distinct feature to characterize and estimate wheat moisture.

[0132] Table 2 Moisture prediction results of model 17

[0133]

[0134] Table 2 shows the moisture prediction results based on model (17), with an average relative error of about 7.7%. It is observed that the model becomes more accurate when the moisture increases. For example, when the moisture increases from 7.7% to 17.2%, the relative error estimated using model 14 decreases from 12.6% to 0.5%.

[0135] This example also evaluates the performance of the simplified model (18). As can be seen from Table 1, the assumption of the simplified model (i.e., b<0) is valid. Therefore, when the volume density approaches infinity, the equation “a×ρ b " approaches to infinity, model 18 only considers f0 and parameter c. The calculation results of wheat moisture are shown in Table 3, and the results are calculated by the simplified model (18). In addition, when the wheat moisture is higher, the prediction becomes more accurate. Comparing Table 2 with Table 3, it is found that the simplified model achieves satisfactory performance, with an average relative error of 10.7%, which is slightly higher than the average error of model (17). This result verifies the effectiveness of model (18), that is, if ρ approaches infinity, the resonant frequency f ρ is stable and its value depends only on parameter c. Therefore, it is possible to consider using fρ Inversion ε eq A moisture prediction model independent of bulk density is obtained.

[0136] Table 3 Simplified model 18 moisture prediction results

[0137]

[0138] It can be seen from Table 1 that parameters a, b, and c can map the moisture content of the target to be measured. In order to further verify the equivalent dielectric property ε eq The effectiveness of the feature vector is composed of a, b, and c. Mapping target M, the mapping relationship can be expressed as Where M represents the target moisture. The linear regression model was established using the MATLAB tool Fitlm We get X = [2.68, 0.154, -1.27], C = 8.70, and variance SSE = 16.57, with an average error of 9.2%. The results are shown in Table 4.

[0139] Table 4 Feature vectors Modeling estimated moisture results

[0140]

[0141] From Table 2, Table 3, and Table 4, it can be seen that model 17 has the smallest error in calculating the target moisture content, and the eigenvector The error of moisture estimation by modeling is second, and the error of target moisture estimation by model 18 is the largest. This further verifies the superiority of model 17, which uses the equivalent dielectric property ε eq Mapping moisture availability.

[0142] In summary, the present invention studies a wireless sensing method for grain moisture based on RFID technology, aiming to achieve a transformation from a pattern-based method to a mathematical model-based method in sensing means. The relationship between the dielectric constant and the resonant frequency of the tag was studied, and the correlation between the dielectric constant and the resonant frequency was simulated by HFSS finite element analysis modeling. The characteristics of the RFID signal caused by tag deformation and the influence of volume density on the dielectric constant were also studied. Based on this, the concept of equivalent dielectric properties was proposed as an effective feature for characterizing and estimating the moisture content of wheat. In order to obtain stable performance, a volume density-independent algorithm was proposed, and the performance of the algorithm was verified by the limit method. The feasibility of the proposed Tag2Model scheme was verified through a large number of experiments on a variety of wheat samples, and the use of the model method helps to reveal the mechanism problems of wireless sensing technology.

[0143] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0144] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A grain moisture sensing method based on RFID signals, characterized in that: include: Build a physical perception platform; Initially model the parameters that affect the tag performance and simplify them to obtain the relationship between the tag deformation and equivalent dielectric properties and the tag resonant frequency; Determine the critical value of volume density according to the relationship between the tag deformation amount, the increment of equivalent dielectric characteristics and the resonant frequency; Under the condition that the body density is greater than the body density critical value, the current resonant frequency of the tag is calculated; The relationship between the current deformation of the tag and the equivalent dielectric properties and the current resonant frequency of the tag is used to solve the equivalent dielectric properties and map the moisture content of the grain.

2. The grain moisture sensing method based on RFID signal according to claim 1 is characterized in that: The parameters that affect label performance are initially modeled as follows: Where c is the speed of light, which is a constant in air under standard atmospheric pressure, υ is the Poisson's ratio, which ranges from 0<υ<1, L is the label length, L" is the compensation value of the label length caused by the edge field effect, and ε r The relative dielectric constant of the label material property is a constant, H and W are the thickness and width of the label medium respectively, and s represents the strain force, which represents the food pressure on the label and has no unit dimension.

3. The grain moisture sensing method based on RFID signal according to claim 1 is characterized in that: The relationship between the tag deformation, equivalent dielectric properties and tag resonant frequency is obtained as follows: f n is the tag resonant frequency, L n , ε eq,n Respectively represent the strain s n Length and equivalent dielectric properties of the lower tab.

4. The grain moisture sensing method based on RFID signal according to claim 3 is characterized in that: The specific steps to determine the critical value of body density are as follows: For the n+1th state, we have: Among them, ΔL represents the deformation of the label due to the strain s or the increase in body density ρ, Δε eq represents the increment of equivalent dielectric properties; f n+1 is the resonant frequency obtained after a certain amount of food is filled into the container and pressed; ε eq,n is the equivalent dielectric property at the current volume density; When Δε eq = 0, we get the volume density ρ min , then min This is the critical value of volume density.

5. The grain moisture sensing method based on RFID signal according to claim 3 is characterized in that: Specific steps to calculate the current resonant frequency of the tag: Construct a mathematical model between the tag resonant frequency and body density under critical state: Among them, f ρ is the resonant frequency of the tag when the body density is ρ, f0 is the tag in the initial state, in air, when the force s = 0 and ρ = 0, a, b, c are model parameters, which are determined by linear regression from experimental data.

6. The grain moisture sensing system based on RFID signal is characterized by: The grain moisture sensing method based on RFID signals according to any one of claims 1 to 5 includes: a tag, an RFID antenna, a container, an analog VNA device and a computer; the tag is attached to the inner wall of the container and is in contact with the wheat; the RFID antenna is connected to the analog VNA device to send and receive electromagnetic wave signals; when the tag is excited, a signal is generated in the reverse direction into the space, the signal can be received by the RFID antenna, and sent to the computer through the analog VNA device.

7. The grain moisture sensing system based on RFID signal according to claim 6 is characterized in that: Also includes: Absorbing material: The absorbing material is pasted on the inner wall of the container to absorb the electromagnetic waves transmitted by the label and reflected by the cross section of the grain.