Multimodal wireless radio frequency sensing methods and systems

Abstract

This invention provides a multimodal wireless radio frequency sensing method and system, relating to the fields of wireless sensing and radio frequency communication technology. The method includes: determining the current relative humidity based on the deviation between the extracted first resonant frequency and the acquired reference resonant frequency, combined with the humidity scaling relationship under the current ambient temperature; based on the extracted first return loss amplitude, eliminating the high-frequency internal resistance loss caused by the determined current ambient temperature and the leakage current resistance loss caused by the determined current relative humidity to obtain the pure gas-sensitive resistor loss, and determining the target gas concentration based on the pure gas-sensitive resistor loss. This invention achieves accurate self-compensation for temperature and humidity interference while maintaining the advantages of passive wireless deployment.

Description

Technical Field

[0001] This invention relates to the field of wireless sensing and radio frequency communication technology, and in particular to multimodal wireless radio frequency sensing methods and systems. Background Technology

[0002] Wireless passive radio frequency sensing technology, due to its characteristics of requiring no battery power and no physical wiring connection, has broad application prospects in gas monitoring in extreme industrial environments such as flammable and explosive environments and high-temperature confined spaces. Existing wireless passive gas sensors are usually based on the LC resonance principle, coating the capacitive region of the LC resonator with a gas-sensitive material. They transmit radio frequency excitation signals through the readout end and receive the echo signals reflected from the sensing end, inferring the gas concentration based on the resonant frequency shift or amplitude change of the echo signal.

[0003] However, the existing wireless passive gas sensing technologies mentioned above have the following main drawbacks in practical applications: First, most existing wireless LC gas sensors employ a single sensing unit structure. The dielectric constant and conductivity of the gas-sensitive material not only vary with the target gas concentration but are also significantly affected by ambient temperature and relative humidity. In complex industrial environments, temperature fluctuations can cause thermal drift of the LC resonant frequency, while humidity changes can alter the dielectric layer capacitance and increase leakage current losses. Because a single sensing unit cannot distinguish whether the signal change is caused by the target gas or by ambient temperature and humidity interference, the final calculated gas concentration data contains errors, resulting in significant baseline drift.

[0004] Secondly, to eliminate temperature and humidity interference, existing technologies typically employ external commercial temperature and humidity sensors for compensation. These external sensors are mostly active devices, requiring independent power supplies and wired data transmission. This not only increases system power consumption and maintenance costs but also compromises the flexibility and security of deploying wireless passive sensors in extreme environments. If an additional passive LC temperature reference unit is used, existing technologies lack an effective method for jointly decoupling the signals from the temperature reference unit and the gas-sensitive unit, making it difficult to achieve accurate separation of multiple physical quantities.

[0005] Third, signal feature extraction relies on a single dimension. Existing signal processing methods often focus only on a single dimension, such as the resonant frequency shift or return loss amplitude variation of the echo signal. However, temperature, humidity, and gas concentration affect LC resonant circuits through different mechanisms, manifesting as changes in capacitance, resistance loss, and combined impedance, respectively. Relying solely on a single dimension of signal features cannot construct a complete equivalent circuit model, making it impossible to physically eliminate high-frequency internal resistance losses caused by temperature and leakage current resistance losses caused by humidity, thus limiting detection accuracy and reliability. Summary of the Invention

[0006] This invention provides a multimodal wireless radio frequency sensing method that integrates a sensing architecture with dual LC units for both humidity and temperature sensing, and combines a two-dimensional signal feature extraction and decoupling algorithm to achieve accurate self-compensation for temperature and humidity interference while maintaining the advantages of passive wireless deployment.

[0007] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: In a first aspect, a multimodal wireless radio frequency sensing method is applied to a multimodal wireless radio frequency sensing chip system, the system including a readout end and a sensing end, the sensing end including a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit, comprising the following steps: Step 1: The readout end transmits an radio frequency excitation signal to the sensing end, receives the echo signals reflected by the air-humidity sensitive LC sensing unit and the temperature sensitive LC sensing unit, and extracts the first resonant frequency and the first echo loss amplitude of the air-humidity sensitive LC sensing unit, as well as the second resonant frequency of the temperature sensitive LC sensing unit from them. Step 2: Based on the extracted second resonant frequency and combined with a preset temperature fitting relationship, determine the current ambient temperature; Step 3: Based on the determined current ambient temperature, query the preset reference matrix to obtain the reference resonant frequency and reference return loss amplitude of the zero gas concentration and zero humidity environment corresponding to the current ambient temperature. Step 4: Based on the deviation between the extracted first resonant frequency and the obtained reference resonant frequency, and combined with the humidity scaling relationship under the current ambient temperature, determine the current ambient relative humidity; Step 5: Based on the extracted first return loss amplitude, remove the high-frequency internal resistance loss caused by the current ambient temperature and the leakage current resistance loss caused by the current ambient relative humidity to obtain the pure gas-sensitive resistor loss, and determine the target gas concentration based on the pure gas-sensitive resistor loss.

[0008] Secondly, a multimodal wireless radio frequency sensing system, which implements the method described above, includes a readout end and a sensing end; The sensing end includes a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit. The surface of the interdigitated capacitor of the humidity-sensitive LC sensing unit is coated with a polyaniline sensitive film. The readout unit includes an RF transceiver antenna, a vector network analyzer, and a processing module; The vector network analyzer transmits radio frequency excitation signals to the sensing end through the radio frequency transceiver antenna, receives the echo signals reflected by the air humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit, and extracts the first resonant frequency and the first return loss amplitude of the air humidity-sensitive LC sensing unit, as well as the second resonant frequency of the temperature-sensitive LC sensing unit.

[0009] Thirdly, a computing device includes: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to perform the method as described.

[0010] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0011] The above-described solution of the present invention has at least the following beneficial effects: This application effectively solves the gas detection distortion problem caused by the cross-sensitivity of temperature and humidity in existing technologies by constructing a dual-unit architecture including a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit, and combining it with a two-dimensional signal decoupling algorithm. Specifically, the second resonant frequency of the temperature-sensitive LC sensing unit is used to independently calculate the current ambient temperature, and based on this, the reference resonant frequency and reference return loss amplitude of a zero gas concentration and zero humidity environment are obtained by querying the reference matrix, which can physically eliminate the influence of temperature drift on the baseline. Furthermore, when determining the target gas concentration, the high-frequency internal resistance loss caused by the current ambient temperature and the leakage current resistance loss caused by the current relative humidity are eliminated based on the first return loss amplitude to obtain the pure gas-sensitive resistor loss. This processing method enables this application to maintain high-precision gas concentration detection capability in complex industrial environments with wide temperature range and high humidity fluctuations, and reduces the false alarm rate and false negative rate caused by environmental interference.

[0012] This application employs a passive wireless deployment of the sensor by using an interaction method between the readout and sensing ends based on the LC magnetic coupling resonance principle. The sensing end requires no built-in battery or physical cable connection; data reading is accomplished solely through the energy interaction between the radio frequency excitation signal and the echo signal. This structure overcomes the difficulties in deployment and high maintenance costs of traditional wired or battery-powered sensors in extreme environments such as high temperatures, enclosed spaces, and flammable / explosive conditions. The passive wireless characteristic not only eliminates the need for periodic battery replacements but also avoids the safety hazards of wired connections under harsh operating conditions.

[0013] This application constructs a signal processing closed loop that senses multiple physical quantities by simultaneously extracting the first resonant frequency, the first return loss amplitude, and the second resonant frequency. Unlike existing technologies that rely solely on single frequency shifts or amplitude changes for detection, this application utilizes the sensitivity of the resonant frequency to changes in capacitance (i.e., humidity-dominated changes) and the sensitivity of the return loss amplitude to changes in resistance (i.e., gas-dominated changes), combined with an equivalent circuit model for parameter separation. This dual-dimensional feature extraction mechanism ensures effective decoupling of temperature and humidity interference signals from the target gas signal at the theoretical model level, improving the system's signal-to-noise ratio and response sensitivity to trace gas changes. Attached Figure Description

[0014] Figure 1 The images show the return loss-resonance frequency response results under simulation (black, obtained from ANSYS HFSS 2023 simulation) and actual testing (red).

[0015] Figure 2 SEM images of PANI under optimal quantification control (ammonium persulfate: aniline molar ratio = 1.05: 1, pH = 0.5).

[0016] Figure 3 XRD patterns of PANI at different oxidation degrees (molar amounts of ammonium persulfate and aniline = 1.25:1, 1.15:1, 1.05:1, 0.95:1, 0.85:1).

[0017] Figure 4 XRD patterns of PANI at different protonation levels (Ph = 0.1, 0.5, 1) when the molar ratio of ammonium persulfate to aniline is 1.05:1.

[0018] Figure 5 The image shows the C1s XPS spectrum of PANI after protonation regulation (synthetic environment pH=0.5).

[0019] Figure 6 The N1s XPS spectrum of PANI after protonation regulation (synthetic environment pH=0.5).

[0020] Figure 7 The XPS spectrum of PANI after protonation regulation (synthetic environment pH=0.5) is shown.

[0021] Figure 8 C1s XPS spectra of PANI with low protonation (synthetic environment pH>1.5).

[0022] Figure 9 N1s XPS spectra of PANI with low protonation (synthetic environment pH>1.5).

[0023] Figure 10 XPS full spectrum of PANI with low protonation degree (synthetic environment pH>1.5).

[0024] Figure 11 Bar charts showing the response of PANI to NH3 at different oxidation levels (molar amounts of ammonium persulfate and aniline = 1.25:1, 1.15:1, 1.05:1, 0.95:1, 0.85:1).

[0025] Figure 12Bar chart showing the response of PANI to NH3 at different protonation levels (synthesis environment pH = 0.1, 0.3, 0.5, 0.7, 1) when the molar ratio of ammonium persulfate to aniline is 1.05:1.

[0026] Figure 13 The response curve of PANI to 100 ppb NH3 after 10 cycles under optimal ionization control (ammonium persulfate molar: aniline molar = 1.05: 1, pH = 0.5).

[0027] Figure 14 The response curve of PANI to 100 ppm NH3 in 10 cycles under high-quality ionization regulation is shown. Detailed Implementation

[0028] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0029] like Figures 1 to 14 As shown, embodiments of the present invention propose a multimodal wireless radio frequency sensing method, applied to a multimodal wireless radio frequency sensing chip system. The system includes a readout end and a sensing end. The sensing end includes a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit, comprising the following steps: Step 1: The readout end transmits an radio frequency excitation signal to the sensing end, receives the echo signals reflected by the air-humidity sensitive LC sensing unit and the temperature sensitive LC sensing unit, and extracts the first resonant frequency and the first echo loss amplitude of the air-humidity sensitive LC sensing unit, as well as the second resonant frequency of the temperature sensitive LC sensing unit from them. Step 2: Based on the extracted second resonant frequency and combined with a preset temperature fitting relationship, determine the current ambient temperature; Step 3: Based on the determined current ambient temperature, query the preset reference matrix to obtain the reference resonant frequency and reference return loss amplitude of the zero gas concentration and zero humidity environment corresponding to the current ambient temperature. Step 4: Based on the deviation between the extracted first resonant frequency and the obtained reference resonant frequency, and combined with the humidity scaling relationship under the current ambient temperature, determine the current ambient relative humidity; Step 5: Based on the extracted first return loss amplitude, remove the high-frequency internal resistance loss caused by the current ambient temperature and the leakage current resistance loss caused by the current ambient relative humidity to obtain the pure gas-sensitive resistor loss, and determine the target gas concentration based on the pure gas-sensitive resistor loss.

[0030] In this embodiment, this application constructs a dual-unit architecture comprising a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit, and combines this with a two-dimensional signal decoupling algorithm to effectively solve the gas detection distortion problem caused by the cross-sensitivity of temperature and humidity in the prior art. Specifically, the second resonant frequency of the temperature-sensitive LC sensing unit is used to independently calculate the current ambient temperature, and based on this, a reference resonant frequency and reference return loss amplitude for a zero gas concentration and zero humidity environment are obtained by querying the reference matrix. This can physically eliminate the influence of temperature drift on the baseline. Furthermore, when determining the target gas concentration, the high-frequency internal resistance loss caused by the current ambient temperature and the leakage current resistance loss caused by the current relative humidity are eliminated based on the first return loss amplitude to obtain the pure gas-sensitive resistor loss. This processing method enables this application to maintain high-precision gas concentration detection capability in complex industrial environments with wide temperature ranges and high humidity fluctuations, reducing the false alarm rate and false negative rate caused by environmental interference.

[0031] In this embodiment, the present application employs a readout-sensor interaction method based on the LC magnetic coupling resonance principle, achieving passive wireless deployment of the sensor. The sensor requires no built-in battery or physical cable connection; data reading is completed solely through the energy interaction between the radio frequency excitation signal and the echo signal. This structure overcomes the difficulties in deployment and high maintenance costs of traditional wired or battery-powered sensors in extreme environments such as high temperatures, enclosed spaces, and flammable / explosive conditions. The passive wireless characteristic not only eliminates the need for periodic battery replacements but also avoids the safety hazards of wired connections under harsh operating conditions.

[0032] In this embodiment, this application constructs a signal processing closed loop for multi-physical quantity sensing by simultaneously extracting the first resonant frequency, the first return loss amplitude, and the second resonant frequency. Unlike existing technologies that rely solely on single frequency shifts or amplitude changes for detection, this application utilizes the sensitivity of the resonant frequency to capacitance changes (i.e., humidity-dominated changes) and the sensitivity of the return loss amplitude to resistance changes (i.e., gas-dominated changes), combined with an equivalent circuit model for parameter separation. This dual-dimensional feature extraction mechanism ensures effective decoupling of temperature and humidity interference signals from the target gas signal at the theoretical model level, improving the system's signal-to-noise ratio and response sensitivity to trace gas changes.

[0033] In another preferred embodiment of the present invention, step 1, in which the readout end transmits a radio frequency excitation signal to the sensing end, receives the echo signals reflected by the humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit, and extracts from them the first resonant frequency and the first return loss amplitude of the humidity-sensitive LC sensing unit, and the second resonant frequency of the temperature-sensitive LC sensing unit, includes: Step 11: The vector network analyzer in the readout end sends a swept electromagnetic wave to the sensing end through the radio frequency transceiver antenna, including: decomposing the total inductance of the planar spiral inductor into a superposition of self-inductance and mutual inductance, and simultaneously completing the parasitic parameter correction of the high-frequency skin effect. The specific execution steps are as follows: The substrate for the LC unit at the sensing end was selected as FR-4 epoxy board (relative permittivity εr=4.4, loss tangent tanδ=0.02), the conductor material was high-purity electrolytic copper, the metal layer thickness was t=35μm, and the resistivity of copper was ρ=1.75×10⁻⁶. -8 Ω·m, free permeability μ0 = 4π × 10 -7 H / m.

[0034] The initial geometric parameters of the planar spiral inductor are set, and the formulas for calculating the core geometric parameters are as follows: Formula for calculating coil outer diameter: ; Formula for calculating the average radius of a coil: ; in, The outer diameter of the coil of a planar spiral inductor; The value represents the inner diameter of the coil of the planar spiral inductor; in this design, it is taken as 2mm. This indicates the number of turns of the coil in the planar spiral inductor; in this design, it is set to 8 turns. The linewidth of the metal conductor representing the planar spiral inductor is 0.2 mm in this design. This represents the spacing between adjacent conductors of a planar spiral inductor; in this design, the value is 0.2 mm. The average radius of the coil in a planar spiral inductor is the arithmetic mean of the coil's inner and outer diameters. Substituting these parameters into the calculation, we obtain the coil's outer diameter. average radius Total length of conductor .

[0035] Calculate the self-inductance component Lself by substituting the above parameters: ; in, This represents the self-inductance component of a planar spiral inductor; Represents the vacuum permeability, a fixed physical constant with values ​​ranging from 1 to 10. H / m; This indicates the number of turns of the coil in the planar spiral inductor; in this design, it is set to 8 turns. The average radius of the coil representing a planar spiral inductor; The Greenhouse correction factor represents the self-inductance calculation of the planar spiral inductor, which is a commonly used empirical correction value in the industry. The linewidth of the metal conductor representing the planar spiral inductance; This indicates the coating thickness of the metal wire; in this design, the value is 35 μm. The fixed empirical coefficient representing the conductor thickness correction term; substituting the parameters, the self-inductance is calculated. .

[0036] The mutual inductance component Lmutual is calculated by adding the mutual inductance between the turns of a coil with current flowing in the same direction. The formula is as follows: ; in, This represents the total mutual inductance generated by the unidirectional currents between all turns of a planar spiral inductor. Represents the vacuum permeability, a fixed physical constant, with values ​​ranging from... H / m; This indicates the number of turns of the coil in the planar spiral inductor; in this design, it is set to 8 turns. The average radius of the coil representing a planar spiral inductor; The inter-turn spacing number represents the number of turns between two coil turns, with a value ranging from 1 to... Positive integers; The linewidth of the metal conductor representing the planar spiral inductance; The formula represents the spacing between adjacent conductors of a planar spiral inductor; where For the inter-turn spacing sequence number, The mutual inductance is calculated by summing all terms up to 7. .

[0037] Total inductance and high-frequency parasitic parameter correction, total inductance calculation formula: ; in, This represents the total inductance of the planar spiral inductor; This represents the self-inductance component of a planar spiral inductor; This represents the total inter-turn mutual inductance of the planar spiral inductor; substituting the parameters, the total inductance is calculated. .

[0038] Simultaneously, based on the skin effect model, high-frequency parasitic resistance correction is completed. First, the skin depth at the operating frequency is calculated, as shown in the following formula: ; in, It represents the skin depth of a metallic conductor under a high-frequency alternating electromagnetic field, and is the conductor depth when the current density decays to 1 / e of the surface area. This represents the resistivity of a metallic conductor. In this design, the resistivity of copper is taken as... Ω·m; Pi is a fixed constant. This indicates the circuit's operating frequency; the nominal resonant frequency for this design is 34MHz. The permeability of a metallic conductor is represented by the permeability of copper, a non-ferromagnetic metal, which is approximately equal to the permeability of free space. Substituting the nominal resonant frequency of 34MHz, the skin depth is calculated as follows: .

[0039] Substituting the formula for high-frequency parasitic resistance, we can then complete the internal resistance calculation: ; in, This represents the parasitic series resistance of an inductor under a high-frequency alternating electromagnetic field. Represents the resistivity of a metallic conductor; Indicates the total length of the metal wire; Indicates the linewidth of a metal conductor; Indicates the skin depth of a metallic conductor; This represents the natural constant, with a fixed value of approximately 2.71828; This indicates the coating thickness of the metal conductor; substituting the parameters, the high-frequency internal resistance at the nominal frequency is calculated. This provides basic parameters for subsequent loss separation.

[0040] Parasitic parallel capacitance of inductors is calculated based on a coplanar waveguide model. The calculation is broken down into three components: Parasitic capacitance at horizontal edges between traces Calculation formula: Modulus calculation formula: Capacitance calculation formula: ;in, denoted by the first-kind complete elliptic integral modulus for calculating horizontal edge capacitance, which is dimensionless; Indicates the spacing between adjacent conductor lines; Indicates the linewidth of the metal conductor; This indicates the parasitic capacitance at the horizontal edge between traces; Indicates the correction factor for the coplanar waveguide half-structure; Represents the vacuum permittivity, a fixed physical constant, with various values. F / m; This represents the relative permittivity of the substrate; in this design, the FR-4 substrate has a value of 4.4. The modulus is The first kind of complete elliptic integral; The complementary modulus is The first kind of complete elliptic integral, where .

[0041] Direct parasitic capacitance of trace sidewall Calculation formula: ; in, This indicates the direct parasitic capacitance on the sidewall of the trace; Represents the vacuum permittivity, a fixed value. F / m; This indicates the relative permittivity of the substrate; for FR-4 substrate, the value is 4.4. Indicates the thickness of the plating on the metal wire; Indicates the spacing between adjacent conductors; substrate parasitic capacitance. Calculation formula: Modulus calculation formula: ; Capacitance calculation formula: ; in, The modulus of the first kind of fully elliptic integral is used to calculate the substrate parasitic capacitance. Represents the hyperbolic tangent function; Indicates the spacing between adjacent conductor lines; Indicates the linewidth of the metal conductor; : The thickness of the substrate; Indicates substrate parasitic capacitance; This represents the empirical coefficient for substrate coupling correction; Represents the vacuum permittivity, a fixed value. F / m; This represents the relative permittivity of the substrate material; for FR-4 substrate, the value is 4.4. The modulus is The first kind of complete elliptic integral; The complementary modulus is The first kind of complete elliptic integral, where .

[0042] Substituting the parameters, we get: , , .

[0043] Formula for calculating total parasitic parallel capacitance: ; in, This represents the total parasitic parallel capacitance of a planar spiral inductor; These are, respectively, the horizontal edge parasitic capacitance, the sidewall direct parasitic capacitance, and the substrate parasitic capacitance; Indicates the total length of the metal wire; This indicates the correction value for the invalid trace length at the outer diameter end of the coil; The number of turns of the planar spiral inductor is used to calculate the total parasitic parallel capacitance. .

[0044] The entire process of interdigital capacitor parameter design based on conformal mapping method is as follows: It accurately calculates the static capacitance value of the interdigital capacitor and clarifies the core structural differences between humidity-sensitive and temperature-sensitive LC units. The specific execution steps are as follows: Interdigitated electrodes of the humidity-sensitive LC sensing unit: finger width Spacing finger length cross-logarithm Yes, the interdigitated surface is coated with a protonated polyaniline-sensitive film (synthesis environment pH=0.5, oxidant to monomer molar ratio 1.05:1), and the composite relative permittivity is... The temperature-sensitive LC sensing unit's interdigitated electrodes: Use the same linewidth, line spacing, and finger length as the humidity-sensitive unit, but with 15 pairs of interdigitated fingers. The interdigitated finger surface has no polyaniline sensitive film, only air filling, and the relative permittivity... It is guaranteed to be sensitive only to temperature and unaffected by humidity or ammonia concentration.

[0045] The core parameter calculation for the conformal mapping method begins with calculating the elliptic integral modulus, as shown in the following formula: ; in, This represents the first-kind fully elliptic integral modulus of the interdigital capacitor calculated using the conformal mapping method. Indicates the finger width of the interdigitated electrode; This represents the distance between adjacent fingers of the interdigitated electrode; substituting the parameters, the module is calculated. complementary modulus ; Through numerical calculation using the first type of complete elliptic integral, we obtain , .

[0046] formula for interdigital capacitance: ; in, This indicates the static capacitance value of the interdigitated electrodes; Represents the vacuum permittivity, a fixed physical constant, with various values. F / m; This represents the relative permittivity of the dielectric material filling the space between the interdigitated electrodes; Indicates the number of interdigitated pairs of the interdigitated electrode; Indicates the effective length of a single interdigital finger; The modulus is The first kind of complete elliptic integral; The complementary modulus is The first kind of complete elliptic integral.

[0047] Substitute parameters to calculate: Formula for calculating total system capacitance: ; in, This represents the total system capacitance of the LC resonant unit; This represents the total parasitic parallel capacitance of a planar spiral inductor; This represents the static capacitance value of the interdigitated electrodes; substituting it into the formula for the natural resonant frequency: ; in, This represents the inherent resonant frequency of the LC resonant unit; The conversion factor between angular frequency and frequency; This represents the total inductance of the planar spiral inductor; This represents the total system capacitance of the LC resonant unit; This represents the high-frequency parasitic series resistance of an inductor.

[0048] Based on the parameters calculated by the above theory, three-dimensional electromagnetic field simulation and optimization were completed, and the hardware design parameters and frequency sweep range were finally determined. The specific execution steps are as follows: In SolidWorks, 1:1 3D models of the humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit were created, including the FR-4 substrate, copper spiral inductor, interdigitated electrodes, and polyaniline sensitive thin film layer of the humidity-sensitive unit. The model was then imported into ANSYS HFSS2023 software, and material properties, radiation boundary conditions (boundary distance from the model ≥ 3 times the working wavelength to avoid boundary reflection interference), and a 50Ω lumped excitation port were set to fully match the impedance of the actual VNA port tested.

[0049] The inherent resonant frequencies and quality factor Q values ​​of the two LC units are determined to verify that the deviation between the theoretical calculation values ​​and the simulation values ​​is ≤2%, ensuring the accuracy of the theoretical design. The S11 frequency response curves of the entire frequency band from 10MHz to 300MHz are obtained, and two independent resonant absorption valleys are obtained, which perfectly match the nominal resonant frequencies calculated by theory. The electric field intensity vector diagram and current density distribution cloud diagram at the resonant frequency are output to verify that the electric field is concentrated in the interdigital capacitor region and the magnetic field is concentrated in the spiral inductor region, ensuring that the electric field coupling efficiency of the sensitive material is maximized.

[0050] With the optimization goals of resonant frequency isolation ≥20MHz, maximizing quality factor (Q) and achieving no overlap of resonant peaks over a wide temperature range, parameterized scanning was performed on the number of turns, line width, and line spacing of the inductor, as well as the finger width, finger spacing, and finger length of the interdigitated electrodes. Ultimately, the optimal hardware parameters were determined, resulting in a Q value ≥45 and a resonant frequency drift range of 25MHz~45MHz under all operating conditions of -20℃ to 80℃ and 0-90%RH. For the temperature-sensitive LC unit, the optimized Q value ≥50 and a resonant frequency drift range of 296.5MHz-299.0MHz under all temperature ranges of -20℃ to 80℃, unaffected by humidity or ammonia concentration.

[0051] In HFSS, temperature variables of -20℃ to 80℃, dielectric constant variables (corresponding to 0-90%RH humidity), and ammonia concentration variables (corresponding to 0-100ppm) were set to complete the full-condition frequency sweep simulation. Finally, the frequency sweep bandwidth of the VNA was locked at 10MHz to 300MHz, which completely covers the full-condition resonant frequency drift range of the two LC units.

[0052] Hardware execution flow for frequency sweeping electromagnetic wave transmission: Based on the frequency sweep range locked by the above simulation optimization, the core operating parameters of the vector network analyzer (VNA) are configured as follows: port characteristic impedance 50Ω, linear frequency sweep mode, frequency sweep bandwidth 10MHz~300MHz, frequency point resolution 10kHz (to ensure accurate capture of resonance peaks), RF output power 0dBm (to adapt to the magnetic coupling energy requirements of passive LC units and avoid signal nonlinear distortion), intermediate frequency bandwidth 100Hz (to suppress environmental electromagnetic interference and improve signal-to-noise ratio), and average sampling times 3 (to reduce random noise).

[0053] The VNA's RF output port is connected to a planar helical RF transceiver antenna via a low-loss SMA coaxial cable. Before testing, a full dual-port SOLT (Short-Circuit-Open-Load-Straight-Through) calibration is performed to completely eliminate systematic errors caused by parasitic parameters of the cable, connectors, and antenna. The antenna polarization direction is adjusted to be in the same direction as the planar helical inductor of the LC unit at the sensing end to ensure maximum near-field magnetic coupling efficiency. The VNA continuously transmits linearly swept frequency RF electromagnetic waves to the sensing end according to the pre-configured parameters. Through the near-field magnetic coupling resonance principle, it provides excitation energy to the passive humidity-sensitive and temperature-sensitive LC sensing units, causing the two LC units to resonate at their respective inherent resonant frequencies.

[0054] Step 12: The vector network analyzer receives the reflection and scattering parameter signals reflected by the humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit, including: the RF transceiver antenna simultaneously receives the RF echo signals reflected by the two LC resonant units at the sensing end through magnetic coupling while transmitting swept-frequency electromagnetic waves; the echo signals carry the resonant characteristic information of the LC units and are synchronously transmitted to the receiving port of the VNA through a low-loss coaxial cable. The VNA performs down-conversion, analog-to-digital conversion, and digital demodulation processing on the received echo signals; based on a preset 50Ω port characteristic impedance, the port reflection coefficient is first calculated using the following formula: ; in, The voltage reflection coefficient of the VNA test port is dimensionless and characterizes the degree of matching between the load impedance and the characteristic impedance of the transmission line. This represents the total load impedance of the LC resonant unit at the sensing end, equivalent to the VNA port. This represents the characteristic impedance of the RF transmission line and the VNA port. In this design, it is fixed at 50Ω, which is the common standard impedance in the RF testing industry.

[0055] Then, the return loss S11 is calculated using the reflection coefficient, as shown in the following formula: ; in, This represents the return loss, which characterizes the proportion of the incident signal reflected by the load. The smaller the value, the stronger the resonant absorption. The conversion factor between voltage amplitude and decibel value; This represents the magnitude of the reflection coefficient; finally, the S11 parameters corresponding to each frequency point within the full frequency sweep range of 10MHz~300MHz are acquired, generating the original S11 frequency response curve; the VNA transmits the original S11 frequency response data to the processing module at the readout end, and the processing module performs three levels of preprocessing on the original data: A 5-point moving average algorithm is used to smooth the S11 amplitude curve, suppressing curve spikes caused by environmental electromagnetic noise and preserving the core characteristics of the resonance valley. The curve is normalized based on the maximum S11 value across the entire frequency band to eliminate the overall baseline shift caused by fluctuations in magnetic coupling efficiency at different test distances. The preprocessed S11 curve is divided into intervals corresponding to the humidity-sensitive unit and the temperature-sensitive unit, according to the predefined resonant frequency ranges of the two LC units. The processing module performs an initial verification of the preprocessed curve to determine whether there are resonance absorption valleys with a depth ≥3dB in either interval. If they exist, the process proceeds to the next step; otherwise, the VNA is triggered to re-acquire the signal, avoiding invalid data interference with subsequent calculations.

[0056] Step 13: Track the lowest trough of the reflection and scattering parameter signal curve in the frequency domain, mark the lowest trough frequency corresponding to the humidity-sensitive LC sensing unit as the first resonant frequency, and mark the lowest trough amplitude corresponding to the humidity-sensitive LC sensing unit as the first return loss amplitude. This includes: the processing module locking the pre-processed, segmented frequency domain interval (25MHz~45MHz) dedicated to the humidity-sensitive unit, completely isolating the temperature-sensitive unit's resonant signal and out-of-band environmental noise interference, ensuring that only the effective resonant characteristics of the humidity-sensitive unit are extracted; within the locked frequency domain interval, an ergonomic global extremum search algorithm is used to compare the S11 amplitude of all frequency points in the interval point by point, and locate the global minimum point of the S11 curve (i.e., the lowest trough of the curve); this lowest trough corresponds to the inherent resonant point of the humidity-sensitive RLC parallel resonant circuit, at which point the imaginary part of the total impedance of the circuit is 0, the energy loss is the maximum, and the return loss amplitude is the deepest, which is completely matched with the RLC equivalent circuit model of the present invention.

[0057] The horizontal axis frequency value corresponding to the global minimum point is formally marked as the first resonant frequency of the humidity-sensitive LC sensor unit; the vertical axis S11 amplitude (dB unit) corresponding to the global minimum point is formally marked as the first return loss amplitude of the humidity-sensitive LC sensor unit. Final parameter validity verification: The processing module performs a final verification on the two marked parameters. The verification rules are: resonance valley depth ≥ 3dB, meeting the quality factor requirements for effective resonance; the first resonant frequency is within the preset full-condition drift range of 25MHz-45MHz; the first return loss amplitude is within the preset effective range of -12dB to 2dB; if all three verifications pass, the parameters are output to the subsequent decoupling algorithm module; if any one fails, the VNA is triggered to re-acquire the signal.

[0058] Step 14: Mark the lowest trough frequency of the curve corresponding to the temperature-sensitive LC sensing unit as the second resonant frequency. This includes: the processing module locking the pre-processed, segmented frequency domain interval (296MHz~299MHz) dedicated to the temperature-sensitive unit, completely isolating it from the interval of the humidity-sensitive unit, and thoroughly eliminating interference from the resonant signal and out-of-band noise of the humidity-sensitive unit; the interdigitated surface of the temperature-sensitive LC unit has no polyaniline sensitive film, and its resonant frequency is only affected by the ambient temperature and is not affected by humidity or ammonia concentration; within the locked frequency domain interval, the same traversal global extremum search algorithm as in step 13 is used to compare the S11 amplitude of all frequency points in the interval point by point, and locate the global minimum point (lowest trough of the curve) of the S11 curve. This point corresponds to the inherent resonant frequency of the temperature-sensitive LC unit, and its frequency drift is only caused by changes in ambient temperature.

[0059] Core parameter marking and validity verification: The abscissa frequency value corresponding to the global minimum point is officially marked as the second resonant frequency of the temperature-sensitive LC sensing unit. Simultaneously, parameter validity verification is performed, with the following rules: resonance valley depth ≥ 3dB, meeting the effective resonance requirement; the second resonant frequency is within the preset full-temperature drift range of 296.5MHz~299.0MHz. If the verification passes, the frequency value is output to the subsequent temperature calculation module; if the verification fails, signal resampling is triggered.

[0060] Through precise Greenhouse inductance calculations, conformal mapping capacitor design, and HFSS full-condition simulation optimization, strong isolation of the resonant frequencies of the humidity-sensitive and temperature-sensitive LC units was achieved, avoiding resonance peak overlap interference from the source. Near-field magnetic coupling excitation using swept-frequency electromagnetic waves provided energy supply and signal reading channels for the completely passive, cable-free sensing end, completely eliminating the limitations of wired power supply and battery replacement. Through SOLT system-level calibration, three-level data preprocessing, and a dual validity verification mechanism, signal distortion caused by system parasitic parameters, environmental electromagnetic interference, and coupling distance fluctuations was eliminated from the source, paving the way for subsequent temperature and humidity decoupling... The accurate calculation of ammonia concentration provides highly reliable raw data, reducing calculation errors at the source. It simultaneously acquires signal characteristics of two core dimensions: resonant frequency (corresponding to humidity-dominated capacitance changes) and return loss amplitude (corresponding to ammonia-dominated resistance changes), which perfectly match the decoupling algorithm requirements of this invention based on RLC equivalent circuits. Through full-condition simulation verification of the sweep frequency range design, interval signal extraction, and multi-round validity verification mechanism, it ensures that the system can still stably extract effective signals in complex industrial environments with a wide temperature range of -20℃ to 80℃ and a full humidity range of 0% to 90%RH, significantly reducing the false alarm rate and missed alarm rate in complex environments.

[0061] In another preferred embodiment of the present invention, step 2, determining the current ambient temperature based on the extracted second resonant frequency and a preset temperature fitting relationship, includes: Step 21: The processing module of the readout end pre-stores the temperature frequency fitting curve of the temperature-sensitive LC sensing unit, including: setting 21 standard temperature nodes from -20℃ to 80℃ with a step size of 5℃, stabilizing each temperature node at a constant temperature for 30 minutes to ensure that the temperature-sensitive LC sensing unit is completely thermally balanced with the ambient temperature; using a vector network analyzer, following the signal acquisition process of steps 11-14, acquiring the resonant frequency of the temperature-sensitive LC sensing unit at each temperature node, repeating the test 10 times for each node, removing outliers and taking the arithmetic mean as the standard resonant frequency value of that temperature node, and finally forming the original calibration dataset of temperature-resonant frequency.

[0062] The resonant frequency of the temperature-sensitive LC sensing unit varies with temperature, determined by the nonlinear temperature-dependent thermal expansion of the metallic conductor and the dielectric constant of the substrate. Therefore, a third-order polynomial fitting model is used to construct the temperature-frequency mapping relationship, which can accurately fit the nonlinear changes over a wide temperature range. The fitting formula is as follows: ; in, This represents the ambient temperature value to be calculated; represents the fixed fitting coefficients of the third-order polynomial, which are obtained by fitting the calibration dataset using the least squares method, and are constants after fitting; This indicates the second resonant frequency of the temperature-sensitive LC sensing unit; These are the square and cubic terms of the second resonant frequency, respectively, used to fit the nonlinear characteristic between temperature and resonant frequency. Based on the measured calibration dataset, the fixed coefficients obtained through fitting are: , , , goodness of fit The full-temperature-range fitting error is ≤±0.3℃, which meets the accuracy requirements of industrial testing.

[0063] Five uncalibrated temperature nodes were randomly selected within the range of -20℃ to 80℃ to complete the actual measurement verification: after acquiring the resonant frequency, the temperature was calculated by fitting the formula. The deviation between the calculated value and the actual ambient temperature was verified to be ≤ ±0.5℃. After the verification was passed, the third-order polynomial fitting curve (including fixed fitting coefficients) and the full-temperature-range calibration dataset were solidified and stored in the non-volatile memory of the read-out processing module to form a pre-stored temperature-frequency fitting curve.

[0064] Step 22: The processing module at the readout end substitutes the extracted second resonant frequency into the temperature frequency fitting curve, including: the processing module receiving the second resonant frequency of the temperature-sensitive LC sensing unit finally output in step 14. Simultaneously, the third-order polynomial formula for the temperature-frequency fitting curve pre-stored in step 21, along with the corresponding fixed fitting coefficients, is read from the non-volatile memory. The preparation of the basic data for the solution is completed. The second resonant frequency received... A second validity check is performed, with the following rules: the frequency value must be within the effective frequency range of 296.5MHz to 299.0MHz across the entire temperature range as specified in step 21; if the frequency value exceeds this range, it is determined to be invalid data, triggering step 1 to re-execute the RF signal acquisition and resonant frequency extraction; if the frequency value is within the valid range, it is determined to be valid data to be solved, and proceeds to subsequent processing.

[0065] The second resonant frequency that passed the verification Substituting the pre-stored third-order polynomial fitting formula, the preprocessing calculations for each power term are completed, yielding the results. The accurate calculation results of the first, second, and third terms are obtained, and the preparation of all input parameters for the fitting formula is completed. The results are then output to step 23 for the final temperature calculation.

[0066] Step 23: Based on the frequency-temperature mapping relationship in the temperature-frequency fitting curve, calculate the unique temperature value that matches the second resonant frequency, and determine the unique temperature value as the current ambient temperature. This includes: substituting the parameters of the preprocessed fitting formula from step 22, performing full arithmetic operations on the third-order polynomial, and calculating the temperature value that uniquely corresponds to the input second resonant frequency. : When the second resonant frequency is extracted in step 14 Substituting into the fitting formula, we obtain... ℃; when the second resonant frequency extracted in step 14 Substituting into the fitting formula, we obtain... ℃; when the second resonant frequency extracted in step 14 Substituting into the fitting formula, we obtain... ℃. The calculated temperature value A validity check is performed, with the following rules: the temperature value must be within the preset full-temperature operating range of -20℃ to 80℃; if it exceeds this range, the calculation is deemed invalid, and the radio frequency signal is re-acquired; if it is within the valid range, it is considered a valid temperature value and proceeds to subsequent smoothing processing. To suppress numerical jumps caused by environmental electromagnetic interference and small temperature fluctuations, a moving average filter is applied to the valid temperature values ​​obtained from three consecutive calculations to obtain the final stable smoothed temperature value; this value is then officially determined as the current ambient temperature and stored in the processing module's runtime memory.

[0067] This embodiment achieves passive and interference-free temperature reference calculation, retaining the core advantages of the system. Ambient temperature calculation can be completed solely using the resonant frequency signal of the passive temperature-sensitive LC sensing unit, eliminating the need for external active temperature and humidity sensors and fully preserving the core advantages of a completely passive and cable-free system. Simultaneously, the temperature-sensitive unit is unaffected by humidity or ammonia concentration, ensuring the purity of the temperature reference from the source. A wide-temperature-range, highly reliable nonlinear mapping model has been constructed, resolving the pain point of large linear fitting errors. Through a process of multi-point calibration across the entire temperature range, third-order nonlinear fitting, and multiple rounds of field testing verification, the goodness of fit has been improved. It perfectly adapts to the nonlinear variation of the resonant frequency of the temperature-sensitive LC unit with temperature, solving the problem of sharp error amplification at both ends of traditional linear fitting in a wide temperature range. It can maintain stable calculation accuracy in the entire industrial temperature range of -20℃ to 80℃. Through three-level processing of frequency validity verification, temperature value validity verification, and moving average filtering, invalid data and jump values ​​are thoroughly filtered out. The output current ambient temperature is the sole benchmark for subsequent humidity calculation, temperature and humidity interference elimination, and ammonia concentration calculation, avoiding distortion of subsequent full-link calculations caused by temperature benchmark errors from the source. The third-order polynomial fitting model only requires basic arithmetic operations to complete the temperature calculation, without the need for complex iterative calculations. The processing module can complete the entire process calculation in milliseconds, fully meeting the real-time requirements of industrial site monitoring of toxic and harmful gases. At the same time, the extremely low computational load can adapt to the batch deployment requirements of low-power embedded processing modules.

[0068] In another preferred embodiment of the present invention, step 3, based on the determined current ambient temperature, queries a preset reference matrix to obtain the reference resonant frequency and reference return loss amplitude for a zero-gas-concentration and zero-humidity environment corresponding to the current ambient temperature, including: Step 31: The processing module of the readout end stores the reference matrix. The reference matrix records the standard resonant frequency and standard return loss amplitude at different temperature nodes under conditions of no target gas and no humidity interference. This includes: setting 21 standard temperature nodes from -20℃ to 80℃ in 5℃ increments, corresponding one-to-one with the temperature calibration nodes in Step 21; stabilizing each temperature node for 30 minutes to ensure complete thermal equilibrium between the humidity-sensitive LC sensing unit and the ambient temperature, with no temperature drift hysteresis; using a vector network analyzer, following the signal acquisition process from Steps 11 to 14, acquiring the resonant frequency and return loss amplitude of the humidity-sensitive LC sensing unit under zero gas and zero humidity conditions at each temperature node; repeating the test 10 times for each node, using the Grubbs criterion to remove outliers and taking the arithmetic mean as the standard parameter for that temperature node, ultimately forming the original reference dataset of temperature nodes, standard resonant frequencies, and standard return loss amplitudes. Examples of core nodes are shown in Table 1 below.

[0069] Table 1 Core Nodes

[0070] Based on the collected benchmark dataset, a two-dimensional structured benchmark matrix is ​​constructed, and the matrix is ​​defined by the following formula: ; in, The pre-stored structured reference matrix is ​​a two-dimensional array with n rows and 3 columns. In this design, n=21, which is consistent with the number of temperature nodes. This represents the standard temperature node value corresponding to the i-th row of the matrix, in °C, where i is the matrix row index, ranging from 1 to 21; This represents the i-th row of the matrix and the corresponding temperature node. The standard resonant frequency of the humidity-sensitive LC sensor unit in a zero-gas, zero-humidity environment; This represents the i-th row of the matrix and the corresponding temperature node. The standard return loss amplitude of the humidity-sensitive LC sensor unit in a zero-gas, zero-humidity environment was measured. Five uncalibrated temperature nodes were randomly selected within the range of -20℃ to 80℃ for experimental verification. Reference parameters were collected at the corresponding temperatures, and the deviation between the reference value calculated by matrix interpolation and the measured value was ≤ ±0.2%. After successful verification, the structured reference matrix was... Together with the temperature-frequency fitting curve pre-stored in step 21, it is solidified and stored in the non-volatile memory of the read-end processing module to form a pre-stored reference matrix that can be called in real time.

[0071] Step 32, the processing module at the read end uses the determined current ambient temperature as the index key to perform a search in the reference matrix, including: the processing module reads the current ambient temperature output in step 23 from the running memory. Simultaneously, the structured reference matrix pre-stored in step 31 is read from non-volatile memory. Complete the dataset to prepare the basic data for retrieval; use the current ambient temperature as the index key. A validity check is performed, with the following rules: the temperature value must be within the full temperature range of -20℃ to 80℃ covered by the baseline matrix; if the temperature value exceeds this range, the index is deemed invalid, triggering step 2 to re-execute the temperature calculation process; if the temperature value is within the valid range, it is deemed a valid index key, and the subsequent retrieval process begins. This is based on a valid current ambient temperature. Using the index key, traverse the base matrix. Standard temperature nodes for all rows The distance between the current temperature and each standard node is calculated using the absolute difference formula, as follows: ; in, This represents the absolute difference between the current ambient temperature and the standard temperature node in the i-th row of the matrix; This represents the absolute value operation; The current ambient temperature output in step 23 is used as the search index key; This represents the standard temperature node value corresponding to the i-th row of the baseline matrix; after traversal, select... The minimum value corresponds to one or two standard temperature nodes: if A minimum value of 0 indicates that the current temperature perfectly matches the standard nodes of the matrix, and this node is directly identified as the unique nearest neighbor temperature node; if If the minimum value is not 0, lock the two adjacent standard temperature nodes (upper and lower limit nodes) with the smallest difference as the nearest neighbor temperature node group.

[0072] Step 33: Retrieve the standard resonant frequency corresponding to the temperature node closest to the current ambient temperature from the reference matrix, and use it as the reference resonant frequency. This includes: retrieving the reference matrix pre-stored in step 31. In the process, retrieve the standard resonant frequency corresponding to the nearest temperature node locked in step 32: if it is a single perfectly matched node, directly retrieve the frequency corresponding to that node. If there are two adjacent node groups, retrieve the corresponding low-temperature node for each. Corresponding to high temperature nodes If the current temperature falls between two adjacent standard nodes, to eliminate the discretization error caused by the 5℃ step size, a linear interpolation algorithm is used to calculate the accurate reference resonant frequency. The interpolation formula is as follows: ; in, This indicates the final determined reference resonant frequency; This represents the standard resonant frequency corresponding to the adjacent low-temperature standard node T1; This represents the standard resonant frequency corresponding to the adjacent high-temperature standard node T2; This indicates that the two adjacent standard temperature nodes locked in step 32 satisfy the following conditions. ; This represents the current ambient temperature output in step 23; for perfectly matched node scenes, no interpolation is needed, and the retrieved value is directly used. The calculated reference resonant frequency is used as the reference frequency to be verified. The validity of the calculated reference resonant frequency is verified according to the following rules: the frequency value must be within the effective frequency range of 25MHz to 45MHz under all operating conditions of the humidity-sensitive LC unit; if it is outside the range, the reference parameter is determined to be invalid, and step 32 is triggered to re-execute the search; if it is within the valid range, the value is officially determined as the reference resonant frequency.

[0073] Step 34: Retrieve the standard return loss amplitude corresponding to the temperature node closest to the current ambient temperature from the reference matrix, and use it as the reference return loss amplitude. This includes: retrieving the reference matrix pre-stored in step 31. In the process, retrieve the standard return loss amplitude corresponding to the nearest temperature node locked in step 32: if it is a single perfectly matched node, directly retrieve the corresponding value of that node. If there are two adjacent node groups, retrieve the corresponding low-temperature node for each. Corresponding to high temperature nodes The temperature node retrieved in step 33 is completely consistent with that in step 33, ensuring the uniformity of the interpolation reference. It is fully aligned with the interpolation logic in step 33, employing a linear interpolation algorithm with the same temperature node to calculate the accurate reference return loss amplitude. The interpolation formula is as follows: ; in, This represents the final determined reference return loss amplitude; This represents the standard return loss amplitude corresponding to the adjacent low-temperature standard node T1; This represents the standard return loss amplitude corresponding to the adjacent high-temperature standard node T2; This indicates that the two adjacent standard temperature nodes locked in step 32 satisfy the following conditions. ; This represents the current ambient temperature output in step 23; for perfectly matched node scenes, no interpolation is needed, and the retrieved value is directly used. The calculated reference return loss amplitude is used as the reference value to be verified. The validity of the reference return loss amplitude is verified. The verification rule is: the amplitude must be within the effective range of -12dB to -2dB of the humidity-sensitive LC unit under all operating conditions; if it is outside the range, the reference parameter is determined to be invalid, and step 32 is triggered to re-execute the search; if it is within the effective range, the value is officially determined as the reference return loss amplitude.

[0074] In another preferred embodiment of the present invention, step 4, determining the current relative humidity based on the deviation between the extracted first resonant frequency and the acquired reference resonant frequency, combined with the humidity scaling relationship at the current ambient temperature, includes: Step 41, the processing module at the readout end calculates the frequency deviation between the first resonant frequency and the reference resonant frequency, including: the processing module synchronously reads the first resonant frequency of the humidity-sensitive LC sensor unit under the measured environment output in step 13 from the running memory. And the zero-gas, zero-humidity reference resonant frequency output in step 33, which perfectly matches the current ambient temperature. Complete the basic data preparation before calculation. Use the difference formula to quantitatively calculate the frequency deviation, as follows: ; in, This indicates the frequency deviation between the measured resonant frequency and the reference resonant frequency of the humidity-sensitive LC sensor unit. This represents the first resonant frequency of the humidity-sensitive LC sensor unit under the measured environment, as output in step 13. This represents the reference resonant frequency of a zero-gas, zero-humidity environment at the current ambient temperature, as output in step 33; [The text abruptly ends here, likely due to an incomplete sentence or missing information.] A validity check is performed, with the following rules: the absolute value of the deviation must be within the preset full-condition valid range of 0.2MHz; if it exceeds the range, the deviation is deemed invalid, triggering a re-execution of RF signal acquisition and resonant frequency extraction; if it is within the valid range, Formal output.

[0075] Step 42, based on the equivalent circuit model, identify the frequency deviation, including: decomposition of the equivalent circuit model and resonant frequency influence factors: complete the quantitative decomposition of the influence factors, as shown in the following formula: ; in, This represents the inherent resonant frequency of the LC resonant unit; The total inductance value of the planar spiral inductor is only affected by the ambient temperature, and the temperature effect has been completely eliminated by the reference resonant frequency in step 33. The total capacitance of the LC resonant unit system consists of inductive parasitic capacitance and interdigital capacitance. The interdigital capacitance changes linearly with the relative permittivity of the dielectric and is the core dominant term for the change of resonant frequency. The total equivalent series resistance of the LC resonant unit is composed of the high-frequency internal resistance of the inductor, the resistance of the gas-sensitive film, and the humidity leakage current resistance. It only affects the amplitude of the resonant peak and its pulling effect on the resonant frequency can be ignored. The resistance loss term only has an observable effect on the resonant frequency when the resistance value changes by an order of magnitude. The resistance change caused by the target gas cannot reach this order of magnitude, and therefore has no significant effect on the frequency. Based on the above equivalent circuit model, three core exclusionary arguments are completed to pinpoint the sole source of the deviation: Temperature interference has been completely eliminated, and the reference resonant frequency has been reached. It is the standard value of zero gas and zero humidity at the current temperature, therefore The effects of temperature-induced thermal drift inductance and capacitance have been completely eliminated, with no residual temperature interference. Target gas interference is negligible; the adsorption of target ammonia only causes a change in the conductivity of the polyaniline sensitive film, reflected in the total equivalent resistance. According to the equivalent circuit model, the effect of this resistance change on the resonant frequency is ≤0.02MHz, which is much smaller than the MHz-level frequency deviation caused by humidity and can be completely ignored. Regardless of the target gas concentration; The sole source of capacitance change is humidity: water molecules in the environment are adsorbed by the polyaniline sensitive membrane, which significantly increases the relative permittivity of the dielectric between the interdigital electrodes. This increase in permittivity directly leads to changes in interdigital capacitance. Increase, thereby increasing the total capacitance of the system. Increasing the resonant frequency eventually leads to The frequency decreases, resulting in a negative frequency deviation, which differs from the value calculated in step 41. Physical property matching; after the above model identification and verification, the output of step 41 is confirmed. The effective frequency deviation is caused solely by ambient humidity, without residual interference from temperature or the target gas.

[0076] Step 43: Obtain the pre-stored humidity scaling relationship at the current ambient temperature. This humidity scaling relationship establishes a mapping between frequency deviation and relative humidity value, including: controlling the environment in the sealed test chamber to zero target ammonia concentration to eliminate target gas interference; the temperature control is completely consistent with the standard temperature nodes in Step 31, covering 21 standard temperature nodes from -20℃ to 80℃, with a relative humidity gradient of 10% to 90%RH set at each temperature node, a step size of 10%RH, and temperature and humidity control accuracies of ±0.2℃ and ±2%RH, respectively; after each temperature and humidity combination point has stabilized for 30 minutes, strictly follow the process in Steps 11-14 to collect the resonant frequency of the humidity-sensitive LC sensor unit, and calculate the frequency deviation at the corresponding humidity by combining it with the reference resonant frequency at the same temperature. Each point was tested 10 times. Outliers were removed using the Grubbs criterion, and the arithmetic mean was taken to form a three-dimensional calibration dataset of temperature nodes, frequency deviation, and relative humidity. For each standard temperature node, a third-order polynomial fitting model was used to construct the mapping relationship between frequency deviation and relative humidity at that temperature. The fitting formula is as follows: ; in, This represents the relative humidity value of the environment to be calculated; The fixed fitting coefficients of the third-order polynomial at the corresponding temperature node are obtained by fitting the calibration dataset using the least squares method, and are constants after fitting. Indicates the frequency deviation; These are the square and cubic terms of the frequency deviation, used to fit the nonlinear relationship between humidity and frequency deviation. Each standard temperature node corresponds to a set of independent fitting coefficients, and the goodness of fit... The fitting error across the entire humidity range is ≤ ±3%RH; the fitting models for all temperature nodes together form a pre-stored humidity scaling relation library, which is permanently stored in the non-volatile memory of the processing module. The processing module reads the current ambient temperature output in step 23. The system verifies that the temperature falls within the effective temperature range of -20℃ to 80℃. Once the verification is successful, the search process begins. For indexing, among the 21 pre-stored standard temperature nodes, match the nearest neighbor temperature node / adjacent node group that is completely identical to the one in step 32: If If it perfectly matches the standard node, directly retrieve the set of fitting coefficients corresponding to that node; if Between two adjacent standard nodes, a two-node linear interpolation algorithm is used to calculate the corresponding set of fitting coefficients at the current temperature, ensuring that the scaling model perfectly matches the current temperature. The matched third-order polynomial mapping model of humidity-frequency deviation at the current temperature is then formally determined as the humidity scaling relationship required for this solution.

[0077] Step 44: Input the frequency deviation into the humidity scaling relationship to calculate the corresponding current ambient relative humidity, including: inputting the effective value output in step 42... Substitute the third-order polynomial formula for the humidity scaling relationship at the current temperature retrieved in step 43, complete the full arithmetic operation, and solve for the result that is consistent with the input. The only corresponding relative humidity value .

[0078] The solution obtained The value undergoes double validity verification: The humidity value must be within the physically valid range of 0~100%RH. If it exceeds this range, the calculation is deemed invalid, triggering step 41 to recalculate the frequency deviation. The frequency deviation corresponding to the humidity value must be consistent with the trend of the calibration dataset (the higher the humidity, the greater the negative frequency deviation). If the trend is inconsistent, the calculation is deemed invalid, triggering a re-execution of signal acquisition. To suppress small fluctuations in ambient humidity and jumps in calculated values ​​caused by electromagnetic noise, a moving average filter is applied to the valid humidity values ​​obtained from three consecutive calculations to obtain the final stable and smoothed humidity value. This value is then formally determined as the current ambient relative humidity.

[0079] In another preferred embodiment of the present invention, step 5, based on the extracted first return loss amplitude, removes the high-frequency internal resistance loss caused by the determined current ambient temperature and the leakage current resistance loss caused by the determined current ambient relative humidity to obtain the pure gas-sensitive resistor loss, and determines the target gas concentration based on the pure gas-sensitive resistor loss, including: Step 51: The processing module at the readout end obtains the total equivalent parallel impedance of the sensing end based on the first return loss amplitude, including: the processing module synchronously reads the first return loss amplitude of the humidity-sensitive LC sensing unit under the measured environment output in step 13 from the running memory. (dB unit) Characteristic impedance of VNA port and RF transmission line (Standard impedance for the RF testing industry), system parasitic parameters at the nominal resonant frequency of the humidity-sensitive LC unit, completing the basic data preparation before inverse impedance decoding. The return loss amplitude in dB units is converted to the reflection coefficient magnitude in the linear domain, providing a basis for subsequent impedance inverse decoding. The conversion formula is as follows: ; in, The voltage reflection coefficient magnitude of the VNA test port is dimensionless and represents the proportion of the incident radio frequency signal reflected by the load at the sensing end. This represents the amplitude of the first return loss of the air-humidity sensitive LC sensor unit output in step 13; The standard conversion formula for voltage amplitude to decibel value is given, where 20 is a fixed coefficient for decibel conversion in the voltage dimension. Based on radio frequency transmission line theory, the total equivalent parallel impedance of the sensing end is solved by inversely using the magnitude of the reflection coefficient. The core formula and its modified derivation are as follows: Original reflection coefficient definition formula: ; Inverse transformation of the total load impedance formula: ; in, This represents the total equivalent parallel impedance of the LC resonant unit at the sensing end to the VNA test port; This represents the characteristic impedance of the VNA port and the RF transmission line, which is fixed at 50Ω. This represents the voltage reflection coefficient at the VNA test port. It is a real number in the resonant state, and its magnitude is calculated in the steps described above. Since this step only deals with the impedance at the resonant frequency (the amplitude of the first return loss corresponds to the resonant valley frequency), the imaginary part of the LC resonant circuit is 0 at this time, and the total impedance is purely resistive. Therefore, the inverse solution yields... That is, the total equivalent parallel resistance. The conductivity value corresponding to the total loss The validity of the total equivalent parallel impedance obtained by inverse kinematics is verified. The verification rule is: the impedance value must be within the preset effective range of 10Ω~1000Ω under all operating conditions; if it exceeds the range, the impedance inverse kinematics is deemed invalid, and step 1 is triggered to re-execute RF signal acquisition and return loss amplitude extraction; if it is within the effective range, (including) , The output is then formally sent to steps 52 and 53 for subsequent loss separation processing.

[0080] Step 52: From the total equivalent parallel impedance, separate the resistive component caused by the skin effect of the metal conductor due to the current ambient temperature, as the high-frequency internal resistance loss. This includes: the resistivity of the metal conductor changes linearly with the ambient temperature, which is the core source of the change in high-frequency internal resistance with temperature. The resistivity temperature change correction formula is as follows: ; in, Indicates the current ambient temperature The resistivity of the copper conductor; This represents the standard resistivity of a copper conductor at room temperature (20°C), and is a fixed value. Ω·m; The temperature coefficient of resistance of copper conductors is a fixed value. / ℃; This indicates the current ambient temperature output in step 23; High-frequency skin depth and parasitic resistance calculation at the current temperature: Calculate the high-frequency parasitic series resistance at the current temperature and nominal resonant frequency, then convert it to a parallel equivalent resistance. Skin depth calculation formula: ; Formula for calculating high-frequency series internal resistance: ; Equivalent conversion formula for series and parallel connections (in resonant state): ; in, It represents the skin depth of a metallic conductor at the current temperature, measured in meters (m), and characterizes the penetration depth of high-frequency current on the conductor surface. Pi is a fixed constant. The nominal resonant frequency of the humidity-sensitive LC sensor unit is expressed in Hz. Represents the vacuum permeability, a fixed physical constant, with values ​​ranging from... H / m; This represents the high-frequency parasitic series resistance of the inductor at the current temperature, expressed in Ω. The total length of the conductor representing the planar spiral inductance, expressed in meters (m). The linewidth of the conductor representing a planar spiral inductor, expressed in meters (m). This indicates the thickness of the copper conductor's plating, expressed in meters (m). This represents the natural constant, with a fixed value of approximately 2.71828; This represents the equivalent parallel resistance corresponding to the high-frequency internal resistance loss at the current temperature, in Ω, and is the only output result of this step. This represents the total inductance of the planar spiral inductor, expressed in ohms (H). The calculated... A validity check is performed, with the following rules: the resistance value must be within the preset full-temperature effective range of 500Ω~5000Ω; if it exceeds this range, the loss separation is deemed invalid, and step 51 is triggered to re-execute the impedance inverse solution; if it is within the effective range, and their corresponding conductivity values The loss is officially identified as high-frequency internal resistance loss caused by the current ambient temperature and output to step 54 for loss removal.

[0081] Step 53: Separate the resistance component caused by the water film leakage current due to the current relative humidity from the total equivalent parallel impedance, as the leakage current resistance loss. This includes: controlling the environment in the sealed test chamber to zero target ammonia concentration (0ppm) to completely eliminate target gas interference; covering 21 standard temperature nodes from -20℃ to 80℃, setting a relative humidity gradient of 10% to 90%RH at each temperature node, with a step size of 10%RH, and temperature and humidity control accuracies of ±0.2℃ and ±2%RH, respectively; after each temperature and humidity combination point is stabilized for 30 minutes, strictly follow the process of steps 11-14 to collect the return loss amplitude of the humidity-sensitive LC sensor unit, inversely solve to obtain the total equivalent parallel impedance, and after removing the high-frequency internal resistance loss at the same temperature, obtain the leakage current equivalent resistance caused only by humidity; repeat the test 10 times at each point, use the Grubbs criterion to remove outliers and take the arithmetic mean, and finally form a three-dimensional calibration dataset of temperature node-relative humidity-leakage current equivalent resistance.

[0082] For each standard temperature node, a third-order polynomial fitting model is used to construct the mapping relationship between relative humidity and leakage current equivalent resistance at that temperature. The fitting formula is as follows: ; in, This represents the equivalent parallel resistance caused solely by water film leakage current at the corresponding temperature and humidity. represents the fixed fitting coefficients of the third-order polynomial at the corresponding temperature node, which are obtained by fitting the calibration dataset using the least squares method and are constants after fitting. This represents the ambient relative humidity value, expressed in %RH, and is the only input variable in the formula. These are the square and cubic terms of relative humidity, used to fit the nonlinear characteristics between humidity and leakage current resistance. Each standard temperature node corresponds to a set of independent fitting coefficients, and the goodness of fit is... The fitting error across the entire humidity range is ≤±4%; the fitting models for all temperature nodes together form a pre-stored humidity-leakage current-resistance scaling relation library, which is permanently stored in the non-volatile memory of the processing module.

[0083] by For indexing, among the 21 pre-stored standard temperature nodes, match the nearest neighbor / adjacent node group that is completely consistent with step 32, and retrieve the fitting coefficient group of the corresponding temperature node; if Located between two adjacent nodes, a two-node linear interpolation algorithm is used to calculate the corresponding set of fitting coefficients at the current temperature, ensuring that the scaling model perfectly matches the current temperature. The output of step 44... Substituting the fitting formula at the current temperature, the unique corresponding equivalent parallel resistance of the leakage current can be calculated. Leakage current, resistance, and loss validity verification and output: The calculated... A validity check is performed, with the following rules: the resistance value must be within the preset valid range of 200Ω to 2000Ω under all operating conditions; if it exceeds this range, the loss separation is deemed invalid, and step 51 is triggered to re-execute the impedance inverse solution; if it is within the valid range, and their corresponding conductivity values The leakage current resistance loss is officially identified as caused by the current relative humidity of the environment, and is output to step 54 for loss elimination.

[0084] Step 54: Subtract the high-frequency internal resistance loss and the leakage current resistance loss from the total equivalent parallel impedance to obtain the resistance component caused only by the adsorption of the target gas, which is used as the pure gas-sensitive resistor loss. This includes: eliminating temperature and humidity losses based on the admittance superposition principle of RLC parallel resonant circuit (the total conductance in the parallel circuit is equal to the sum of the conductances of each branch). The core formula is as follows: ; ; in, The unit is S, which represents the pure gas-sensitive equivalent parallel conductance caused solely by the adsorption of the target gas. This represents the total equivalent parallel conductance output in step 51, in units of S, corresponding to the total loss. The output of step 52 represents the conductance corresponding to the high-frequency internal resistance caused by the temperature, in units of S; The output of step 53 represents the conductance corresponding to the leakage current caused by humidity, in seconds. This indicates the loss of the pure gas-sensitive resistor caused solely by the adsorption of the target gas. The protonation degree of the polyaniline sensitive film is precisely controlled, with a synthesis environment pH of 0.5 and a molar ratio of oxidant to monomer of 1.05:1. Polyaniline in this protonated state is most sensitive to the dedoping effect of ammonia. When the polyaniline sensitive film adsorbs the target ammonia, a significant dedoping effect occurs, resulting in a substantial decrease in the carrier concentration on the polyaniline molecular chain. This directly leads to a reduction in the conductivity between the interdigitated electrodes of the gas humidity-sensitive LC sensing unit, and an increase in the equivalent parallel resistance. The increase ultimately manifests as a positive change in the amplitude of the first return loss (a shallower resonance valley), perfectly corresponding to the measured characteristics of the RF signal; the calculated... Perform a validity check, with the following rules: It must be a positive value. The signal must be within the preset effective range of 100Ω to 10000Ω; if the rule is not met, the loss removal is deemed invalid, triggering a re-execution of the entire RF signal acquisition process; if the rule is met, The loss of the pure gas-sensitive resistor is officially determined and output to step 55 for target gas concentration calculation.

[0085] Step 55: Compare the pure gas resistor loss with the pre-stored gas concentration calibration table to determine the target gas concentration. This includes: setting a standard ammonia concentration gradient of 0ppm to 1000ppm in a sealed standard gas test chamber covering a full temperature range of -20℃ to 80℃ and a full humidity range of 10% to 90%RH; after each temperature, humidity and concentration combination point has stabilized for 30 minutes, strictly follow the entire process method of this invention to collect the pure gas resistor loss. Each point was tested 10 times, and the arithmetic mean was taken after removing outliers, resulting in a four-dimensional calibration dataset consisting of temperature node, humidity node, pure gas sensor, and standard ammonia concentration. For each temperature and humidity combination node, a third-order polynomial fitting model was used to construct the mapping relationship between the pure gas sensor and the ammonia concentration. The fitting formula is as follows: ; in, Indicates the concentration value of the target ammonia gas; : The fixed fitting coefficients of the third-order polynomial under the corresponding temperature and humidity nodes are obtained by fitting the calibration dataset using the least squares method, and are constants after the fitting is completed; This represents the pure gas-sensitive resistor loss output in step 54; These are the square and cubic terms of the pure gas-sensitive resistor, used to fit the nonlinear characteristics between the gas-sensitive resistor and the gas concentration. The fitting models for all temperature and humidity nodes together form a pre-stored three-dimensional gas concentration calibration table, which is permanently stored in the non-volatile memory of the processing module. The goodness-of-fit is... The full-range fitting error is ≤ ±2%FS.

[0086] by , For dual-indexing, the nearest temperature and humidity node group is matched in the pre-stored calibration table, and the concentration mapping fitting coefficient group of the corresponding node is retrieved. If the current temperature and humidity are between four adjacent nodes, a bilinear interpolation algorithm is used to calculate the corresponding fitting coefficient group under the current temperature and humidity, ensuring that the mapping model completely matches the measured environment. The output of step 54 is... Substituting the values ​​into the concentration mapping formula under the current temperature and humidity, the unique target ammonia concentration value can be calculated. For the validity verification and final determination of concentration values, the concentration value must be within the physical effective range of 0~1000ppm; the pure gas-sensitive resistor corresponding to the concentration value must be consistent with the calibration trend (the higher the ammonia concentration, the higher the effective range). If any one of the conditions is not met, the solution is deemed invalid, triggering a full re-execution. The effective concentration values ​​obtained from three consecutive solutions are subjected to moving average filtering to suppress numerical jumps caused by small environmental fluctuations and noise, resulting in the final stable concentration value, which is officially determined as the target gas concentration.

[0087] In another preferred embodiment of the present invention, the degree of protonation of the polyaniline sensitive film is controlled, the pH value of its synthesis environment is 0.5, and the molar ratio of oxidant to monomer is 1.05:1; after the polyaniline sensitive film adsorbs the target gas, it undergoes a dedoping effect, which causes a change in the conductivity of the gas humidity sensitive LC sensing unit, resulting in a change in the amplitude of the first return loss.

[0088] This embodiment precisely controls the protonation degree of the polyaniline sensitive film. The core control parameters are: Synthesis environment pH=0.5: This pH value allows the polyaniline molecular chain to reach the optimal protonation state. The degree of protonation directly determines the response sensitivity of the film to the target ammonia gas. When pH=0.5, the proton concentration on the film surface is moderate, which can quickly react with ammonia gas to remove doping and avoid response lag or decreased sensitivity caused by too many / too few protons.

[0089] Oxidant to monomer molar ratio = 1.05:1: This ratio ensures that the polyaniline monomer is fully polymerized, while avoiding excessive oxidant which may lead to film structural defects and abnormal conductivity, or insufficient oxidant which may lead to incomplete polymerization and a reduction in gas-sensitive active sites. Ultimately, a sensitive film with uniform structure and stable protonation is formed, providing a good material basis for gas-sensitive response.

[0090] When the polyaniline sensitive film adsorbs the target gas (ammonia in this invention), a dedoping effect occurs. The specific process and its impact on the sensing unit are as follows: The protonated polyaniline molecular chain carries a large number of protons. When ammonia (an alkaline gas) is adsorbed by the thin film, the ammonia molecules combine with the protons, leading to a significant decrease in the concentration of charge carriers (protons) on the polyaniline molecular chain. This decrease in charge carrier concentration directly reduces the conductivity between the interdigitated electrodes of the humidity-sensitive LC sensing unit (conductivity is positively correlated with charge carrier concentration). The decrease in conductivity increases the total equivalent parallel resistance of the LC resonant circuit. According to the principle of radio frequency signal transmission, the change in resistance alters the reflection ratio of the incident signal, ultimately resulting in a positive change in the amplitude of the first return loss (a shallower resonance valley). This embodiment, by adjusting the thin film synthesis parameters, makes it more sensitive to the target gas and clarifies how the gas adsorption indirectly affects the radio frequency signal (amplitude of the first return loss) through changes in the thin film conductivity.

[0091] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0092] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-modal wireless radio frequency sensing method applied to a multi-modal wireless radio frequency sensing chip system, the system comprising a readout end and a sensing end, the sensing end comprising a gas humidity sensitive LC sensing unit and a temperature sensitive LC sensing unit, characterized in that, Includes the following steps: Step 1: The readout end transmits an radio frequency excitation signal to the sensing end, receives the echo signals reflected by the air-humidity sensitive LC sensing unit and the temperature sensitive LC sensing unit, and extracts the first resonant frequency and the first echo loss amplitude of the air-humidity sensitive LC sensing unit, as well as the second resonant frequency of the temperature sensitive LC sensing unit from them. Step 2: Based on the extracted second resonant frequency and combined with a preset temperature fitting relationship, determine the current ambient temperature; Step 3: Based on the determined current ambient temperature, query the preset reference matrix to obtain the reference resonant frequency and reference return loss amplitude of the zero gas concentration and zero humidity environment corresponding to the current ambient temperature. Step 4: Based on the deviation between the extracted first resonant frequency and the obtained reference resonant frequency, and combined with the humidity scaling relationship under the current ambient temperature, determine the current ambient relative humidity; Step 5: Based on the extracted first return loss amplitude, remove the high-frequency internal resistance loss caused by the current ambient temperature and the leakage current resistance loss caused by the current ambient relative humidity to obtain the pure gas-sensitive resistor loss, and determine the target gas concentration based on the pure gas-sensitive resistor loss.

2. The multimodal wireless radio frequency sensing method according to claim 1, characterized in that, The readout end transmits an radio frequency excitation signal to the sensing end, receives the echo signals reflected by the humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit, and extracts the first resonant frequency and the first return loss amplitude of the humidity-sensitive LC sensing unit, as well as the second resonant frequency of the temperature-sensitive LC sensing unit, including: The vector network analyzer in the readout end sends swept electromagnetic waves to the sensing end through the radio frequency transceiver antenna. The vector network analyzer receives the reflection and scattering parameter signals reflected by the air-humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit; The lowest valley of the curve of the reflection and scattering parameter signal is traced in the frequency domain. The lowest valley frequency of the curve corresponding to the air humidity sensitive LC sensing unit is marked as the first resonant frequency, and the lowest valley amplitude of the curve corresponding to the air humidity sensitive LC sensing unit is marked as the first return loss amplitude. The lowest valley frequency of the curve corresponding to the temperature-sensitive LC sensing unit is marked as the second resonant frequency.

3. The multimodal wireless radio frequency sensing method according to claim 2, characterized in that, The step of determining the current ambient temperature based on the extracted second resonant frequency and a preset temperature fitting relationship includes: The processing module at the readout end pre-stores the temperature frequency fitting curve of the temperature-sensitive LC sensing unit. The processing module at the readout end will substitute the extracted second resonant frequency into the temperature frequency fitting curve; Based on the frequency-temperature mapping relationship in the temperature-frequency fitting curve, a unique temperature value matching the second resonant frequency is calculated, and the unique temperature value is determined as the current ambient temperature.

4. The multimodal wireless radio frequency sensing method according to claim 3, characterized in that, The step of querying a preset reference matrix based on the determined current ambient temperature to obtain the reference resonant frequency and reference return loss amplitude for a zero-gas-concentration and zero-humidity environment at the current ambient temperature includes: The processing module at the readout end stores the reference matrix, which records the standard resonant frequency and standard return loss amplitude at different temperature nodes, without target gas and without humidity interference. The processing module of the readout end uses the current ambient temperature as the index key to perform a search in the reference matrix; The standard resonant frequency corresponding to the temperature node closest to the current ambient temperature is retrieved from the reference matrix and used as the reference resonant frequency; The standard return loss amplitude corresponding to the temperature node closest to the current ambient temperature is retrieved from the reference matrix and used as the reference return loss amplitude.

5. The multimodal wireless radio frequency sensing method according to claim 4, characterized in that, The determination of the current relative humidity based on the deviation between the extracted first resonant frequency and the acquired reference resonant frequency, combined with the humidity scaling relationship at the current ambient temperature, includes: The processing module at the readout end calculates the frequency deviation between the first resonant frequency and the reference resonant frequency; Based on the equivalent circuit model, the frequency deviation is identified; Obtain the pre-stored humidity scaling relationship at the current ambient temperature, wherein the humidity scaling relationship establishes a mapping between frequency deviation and relative humidity value; The frequency deviation is input into the humidity scaling relation to calculate the corresponding current ambient relative humidity.

6. The multimodal wireless radio frequency sensing method according to claim 5, characterized in that, Based on the extracted first return loss amplitude, the high-frequency internal resistance loss caused by the determined current ambient temperature and the leakage current resistance loss caused by the determined current ambient relative humidity are removed to obtain the pure gas-sensitive resistor loss. The target gas concentration is then determined based on the pure gas-sensitive resistor loss, including: The processing module at the readout end obtains the total equivalent parallel impedance of the sensing end by inverse solving based on the first return loss amplitude. The resistive component caused by the skin effect of the metal conductor due to the current ambient temperature is separated from the total equivalent parallel impedance and used as the high-frequency internal resistance loss. The resistive component caused by the water film leakage current due to the current relative humidity is separated from the total equivalent parallel impedance and used as the leakage current resistance loss. Subtracting the high-frequency internal resistance loss and the leakage current resistance loss from the total equivalent parallel impedance yields the resistance component caused solely by the adsorption of the target gas, which is used as the pure gas-sensitive resistor loss. The loss of the pure gas-sensitive resistor is compared with the pre-stored gas concentration calibration table to determine the target gas concentration.

7. The multimodal wireless radio frequency sensing method according to claim 1, characterized in that, The degree of protonation of the polyaniline sensitive film is controlled, the pH value of its synthesis environment is 0.5, and the molar ratio of oxidant to monomer is 1.05:

1. The polyaniline sensitive film undergoes a dedoping effect after adsorbing the target gas, resulting in a change in the conductivity of the gas humidity-sensitive LC sensing unit, which in turn causes a change in the amplitude of the first return loss.

8. A multimodal wireless radio frequency sensing system, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, Including the readout end and the sensing end; The sensing end includes a humidity-sensitive LC sensing unit and a temperature-sensitive LC sensing unit. The surface of the interdigitated capacitor of the humidity-sensitive LC sensing unit is coated with a polyaniline sensitive film. The readout unit includes an RF transceiver antenna, a vector network analyzer, and a processing module; The vector network analyzer transmits radio frequency excitation signals to the sensing end through the radio frequency transceiver antenna, receives the echo signals reflected by the air humidity-sensitive LC sensing unit and the temperature-sensitive LC sensing unit, and extracts the first resonant frequency and the first return loss amplitude of the air humidity-sensitive LC sensing unit, as well as the second resonant frequency of the temperature-sensitive LC sensing unit.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images