Decoupled interference mitigation regulation method and system

Abstract

The application discloses a decoupling predicted interference suppression regulation method and system, and belongs to the technical field of sensors, and comprises the following steps: in the array offline calibration stage, a single gradient variable is applied to the sensor, cross-sensitivity characteristics and electromagnetic crosstalk characteristics are obtained, and a joint feature library is constructed; based on original resistance signals and a real-time regulation voltage matrix, predicted real pressure and temperature are generated through two-stage decoupling; a deviation between actual measurement values and target values is obtained, a cooperative regulation factor is calculated in combination with the joint feature library; and two pre-compensations are performed according to the cooperative regulation factor, so that a cooperative regulation voltage is applied, the double linkage of force-temperature cross-sensitivity and electromagnetic crosstalk is broken, and the overall measurement accuracy under the array cooperative scene is effectively ensured.

Description

Technical Field

[0001] This invention relates to a method and system for interference suppression and control based on decoupling prediction, belonging to the field of sensor technology. Background Technology

[0002] Flexible force-temperature dual-mode sensors are widely used in fields such as human-computer interaction, health monitoring, and intelligent robots. They can respond to pressure and temperature stimuli simultaneously. However, the electrical signals of these sensors are often affected by the cross-influence of pressure and temperature. That is, pressure changes can interfere with the accurate identification of temperature signals, and temperature fluctuations can also affect the measurement accuracy of pressure signals. This cross-sensitivity phenomenon restricts their application in complex environments. In this case, it is necessary to analyze the coupling mechanism of pressure and temperature on the sensor's electrical signals and distinguish their independent effects in order to achieve separate identification of pressure and temperature.

[0003] However, existing technologies do not consider the collaborative differentiation and control chain effects of multidimensional composite interference in multi-sensor arrays, especially the cross-sensitivity of pressure and temperature and electromagnetic crosstalk. Specifically, in an array composed of multiple force-temperature dual-mode sensors, the resistance change caused by the cross-action of pressure and temperature is difficult to distinguish from the resistance fluctuation caused by electromagnetic crosstalk of adjacent sensors. When a single sensor is controlled, electromagnetic crosstalk will further amplify the force-temperature signal coupling error of other sensors, forming a double chain problem of force-temperature cross-sensitivity and electromagnetic crosstalk, which cannot guarantee the overall measurement accuracy in array collaborative scenarios. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for interference suppression and control based on decoupling prediction. By constructing a joint feature library through offline calibration, the cross-sensitivity features of pressure and temperature and electromagnetic crosstalk are extracted respectively. In online cloud operation, the two types of interference are distinguished through two-stage decoupling. The collaborative control factor is calculated to avoid amplified crosstalk. After double pre-compensation, the control voltage is applied to break the double interlock and ensure the accuracy of array collaborative measurement.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] Interference suppression and regulation methods for decoupling prediction include:

[0007] During the offline array calibration phase, individual gradient variables are applied to the sensors to obtain cross-sensitivity features and electromagnetic crosstalk features, and a joint feature library is constructed.

[0008] Based on the original resistance signal and the real-time controlled voltage matrix, the predicted real pressure and temperature are generated through two-stage decoupling.

[0009] Obtain the deviation between the actual measured value and the target value, and calculate the synergistic regulation factor by combining the joint feature library;

[0010] Based on the synergistic regulation factor, double pre-compensation is performed to apply the synergistic regulation voltage.

[0011] Specifically, the steps for constructing a joint feature library include:

[0012] Each force-temperature dual-mode sensor in the sensor array is numbered to generate a sensor sequence;

[0013] For sensor Q x Apply a stepped pressure gradient and maintain t at each pressure level. awa.1 For each time period, calculate the average resistance value at the corresponding pressure level to generate the sensor Q. x Pressure data pairs under different pressures;

[0014] For sensor Q x Apply a stepped temperature gradient and maintain t at each temperature level. awa.2 Time period;

[0015] Calculate the average resistance value at the corresponding temperature and pressure levels to generate sensor Q. x Temperature data pairs at different temperatures;

[0016] Calculate sensor Q x Pressure sensitivity S Px Temperature coefficient of resistance S Tx And generate an array mapping table.

[0017] Specifically, the steps for constructing a joint feature library also include:

[0018] Define a target sensor and adjacent sensors, apply a stepped gradient control voltage to the target sensor, and maintain t at each control voltage level. awa.3 Time period;

[0019] Calculate the average resistance value and resistance fluctuation of each adjacent sensor to construct a four-dimensional data pair;

[0020] Based on the ratio of resistance fluctuation to control voltage, linear fitting is performed to calculate the crosstalk transmission coefficient of the target sensor in order to construct a crosstalk correlation table.

[0021] A three-level structured joint feature library is constructed using the sensor number as the core index;

[0022] Randomly select experimental data that were not used for fitting, calculate resistance deviation and fluctuation deviation. Once the resistance deviation is not less than the resistance deviation threshold or the fluctuation deviation is not less than the fluctuation deviation threshold, re-collect data and perform feature extraction until the verification is successful.

[0023] Specifically, the steps of two-stage decoupling include:

[0024] Collect actual data from the online operation of the array, and construct an offline benchmark set based on the data from the offline calibration phase and the joint feature library;

[0025] Obtain the actual temperature from the actual data, calculate the calibration temperature difference, and thus obtain the dynamic reference resistor;

[0026] Calculate the actual total resistance change and obtain the sensitivity correction coefficient to calculate the real-time pressure sensitivity;

[0027] Dynamically allocate decoupling weights and use the least squares weighted iterative method to calculate pressure and temperature changes;

[0028] By combining the real-time initial pressure and real-time ambient temperature from the actual data, the preliminary actual pressure and preliminary actual temperature are calculated.

[0029] Specifically, the two-stage decoupling process also includes:

[0030] Based on the online real-time position offset between the current sensor and adjacent sensors, configure the corresponding position weight factor and calculate the actual crosstalk resistance change.

[0031] Obtain the pressure change rate and temperature change rate under actual online working conditions, configure the distribution ratio of crosstalk deviation on pressure and temperature, and calculate the pressure deviation and temperature deviation caused by crosstalk.

[0032] Based on the difference between the initial actual physical quantities and the crosstalk deviation, the true pressure and true temperature are obtained.

[0033] Specifically, the steps for calculating the co-regulatory factor include:

[0034] Acquire real-time physical quantities, target physical quantities, and deviation benchmarks; calculate cross-interference deviation based on the difference between the target physical quantity and the real-time physical quantity.

[0035] Classify the deviation levels and construct a deviation information table;

[0036] Obtain real-time pressure sensitivity and real-time temperature sensitivity, and calculate the target resistance change required to compensate for cross-interference deviation;

[0037] Call the crosstalk correlation table to obtain the base pressure control voltage and base temperature control voltage corresponding to the target resistance change, and calculate the total base voltage adjustment.

[0038] Based on the aforementioned deviation level, the corresponding deviation level correction coefficient is invoked to calculate the preliminary cross-interference control factor.

[0039] Specifically, the steps for calculating the co-regulatory factor also include:

[0040] Calculate the change in crosstalk resistance of adjacent sensors, combine it with the real-time pressure sensitivity of adjacent sensors, calculate the pressure deviation of adjacent sensors, and combine it with the background crosstalk deviation to obtain the predicted total crosstalk deviation.

[0041] If the total crosstalk deviation is not greater than the upper limit of crosstalk deviation, there is no need to limit the voltage, and the crosstalk constraint factor is 1.

[0042] If the total crosstalk deviation is greater than the upper limit of crosstalk deviation, calculate the voltage limiting coefficient and the auxiliary correction coefficient, and generate a set of crosstalk constraint factors.

[0043] Calculate the indirect crosstalk deviation, and once the indirect crosstalk deviation exceeds the indirect ratio of the upper limit of the crosstalk deviation, reduce the voltage limit factor;

[0044] Calculate the adjustment amount of the control voltage. Once the adjustment amount of the control voltage is less than the cross-interference control factor, calculate the compensation deviation corresponding to the adjustment amount of the control voltage and calculate the cross-compensation gap.

[0045] Generate a table of synergistic regulation factors and construct a regulation correlation table that includes the target sensor and adjacent sensors.

[0046] Specifically, the steps of dual pre-compensation include:

[0047] Configure and initialize the pre-compensation parameters, and calculate the total cross-interference pre-compensation deviation based on the pressure and temperature coupling resistance deviation.

[0048] The pre-compensated decoupled input signal is obtained by measuring the difference between the original resistance signal of the target sensor and the total cross-interference pre-compensation deviation, and the deviation is verified.

[0049] Obtain the reference voltage of the target sensor from actual data, and calculate the final control voltage to be applied;

[0050] The regulated voltage is applied to the target sensor via the voltage drive terminal of the array.

[0051] Specifically, the steps of dual pre-compensation also include:

[0052] The real-time resistance signal after the target sensor is regulated is collected, and two-stage decoupling is performed to obtain the real pressure and real temperature after regulation.

[0053] Collect real-time resistance signals after adjustment of adjacent sensors, and obtain the actual physical quantities of adjacent sensors;

[0054] Verify whether the deviation between the actual physical quantity of the target sensor and the preset target value has been reduced to within the cross-interference deviation threshold to confirm that the cross-interference compensation is effective;

[0055] Verify whether the crosstalk deviation between adjacent sensors remains within the crosstalk deviation threshold to confirm that crosstalk suppression is effective;

[0056] If the deviation does not meet the standard, return to readjust the pre-compensation deviation calculation logic until the standard is met.

[0057] Decoupled prediction interference suppression control systems include:

[0058] The system includes an offline calibration module, an online simulation module, a verification module, and a data storage module.

[0059] The offline calibration module is used to apply gradient values ​​individually to each force-temperature dual-mode sensor to construct a joint feature library;

[0060] The online simulation module includes a preprocessing unit, a two-stage decoupling unit, a collaborative control unit, and a dual pre-compensation unit;

[0061] The preprocessing unit is used to collect data from the online operation of the array, correct the offline initial resistance drift, and generate a pure resistance signal and the actual total resistance change.

[0062] The two-stage decoupling unit is used to construct a sensitivity matrix and calculate the initial actual physical quantities and crosstalk resistance changes to obtain the true pressure and true temperature.

[0063] The coordinated control unit is used to calculate the cross-interference deviation and classify the deviation level, obtain the basic control voltage adjustment amount, and generate a coordinated control factor table.

[0064] The dual pre-compensation unit is used to generate a pre-compensated decoupled input signal based on the resistance deviation;

[0065] The verification module is used to apply the adjustment amount of the control voltage, collect the real-time resistance signal, and verify the compliance of the physical quantities after the two-stage decoupled output control.

[0066] The data storage module is used to store relevant data on interference suppression and regulation.

[0067] The beneficial effects of this invention are:

[0068] In the offline phase, a joint feature library is constructed to extract pressure / temperature cross-sensitivity features and electromagnetic crosstalk features separately, providing a benchmark for interference differentiation and solving the fundamental problem of confusion between the two types of interference features. In the online phase, a two-stage decoupling is adopted. Based on the feature library, the signal mixing caused by cross-sensitivity is first removed, and then the influence of electromagnetic crosstalk is eliminated to solve the problem of difficulty in distinguishing the resistance change caused by pressure and temperature cross-effects and the fluctuation of crosstalk resistance. When calculating the collaborative control factor, the cross-interference compensation voltage is determined by mapping relationship, and the crosstalk intensity of voltage on adjacent sensors is predicted and constrained to avoid the control amplification coupling error. Finally, the cross-interference and crosstalk deviation caused by control are deducted in advance through double pre-compensation before the control voltage is applied, breaking the double chain of force-temperature cross-sensitivity and electromagnetic crosstalk, effectively ensuring the overall measurement accuracy in array collaborative scenarios. Attached Figure Description

[0069] Figure 1 A schematic diagram of the interference suppression and control method for decoupling prediction;

[0070] Figure 2 This is a flowchart of the process for constructing the joint feature library in this invention;

[0071] Figure 3 This is a flowchart of the two-stage decoupling process in this invention;

[0072] Figure 4 The diagram shows the structure of the interference suppression and control system for decoupling prediction. Detailed Implementation

[0073] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0074] Example 1

[0075] refer to Figures 1 to 3 As shown in the figure, this embodiment introduces a method for interference suppression and control based on decoupling prediction, including the following steps:

[0076] Step S1: During the array offline calibration stage, single physical quantity response simulation is performed, and environmental interference is kept constant, such as electromagnetic shielding, constant temperature and pressure. Gradient pressure and gradient temperature are applied separately to each force-temperature dual-mode sensor. The resistance change of each sensor is collected simultaneously, and the pressure sensitivity and temperature coefficient are calculated. The exclusive mapping relationship between pressure and resistance, and temperature and resistance is established to obtain cross-sensitivity characteristics and avoid misjudging cross-sensitivity as other interference in the future. At the same time, pure electromagnetic crosstalk simulation is performed to ensure that there is no pressure or temperature physical stimulation. Electromagnetic crosstalk characteristics are extracted, the two types of characteristics are integrated, a joint feature library is constructed, and the resistance response law and crosstalk correlation of pressure and temperature acting alone are stored.

[0077] Step S2: To address the difficulty in distinguishing between resistance changes caused by pressure / temperature cross-effects and resistance fluctuations caused by electromagnetic crosstalk, the original resistance signals and real-time control voltage matrix of all sensors in the current array are obtained by combining the joint feature library. Signal separation is achieved through two-stage decoupling. Based on the cross-sensitivity features, a sensitivity matrix is ​​constructed to decompose the original resistance signal into the resistance change caused by pressure alone and the resistance change caused by temperature alone. The pressure and temperature values ​​of the initial decoupling are then deduced. The signal mixing caused by cross-sensitivity is first removed. Based on the real-time control voltage matrix, the electromagnetic crosstalk features in the joint feature library are matched to calculate the crosstalk amplitude of adjacent sensors under the current control voltage. The interference of crosstalk on the pressure and temperature values ​​is further eliminated, and the final predicted real pressure and temperature are output, thus achieving decoupling prediction.

[0078] Step S3: Obtain the deviation between the actual measured value and the target value, determine the cross-interference amplitude that needs to be compensated, and calculate the synergistic control factor that takes into account both cross-interference compensation and crosstalk suppression based on the joint feature library. This includes: calculating the required control voltage adjustment based on the mapping relationship between pressure / temperature and resistance to ensure that the adjustment can offset the measurement deviation caused by cross sensitivity; predicting the crosstalk intensity of the control voltage adjustment to adjacent sensors based on the relationship between control voltage and crosstalk amplitude; if the predicted crosstalk exceeds a preset threshold, adding a constraint factor to limit the control voltage amplitude; and supplementing an auxiliary compensation algorithm to ensure that the control compensates for cross-interference without amplifying crosstalk in adjacent sensors, thus achieving synergistic and non-chaining.

[0079] Step S4: Based on the coordinated control factor, perform dual pre-compensation. For cross-interference pre-compensation, predict the resistance deviation caused by pressure / temperature coupling based on the sensitivity matrix, and subtract it from the input of the decoupling prediction model to eliminate cross-interference caused by control from the source. For electromagnetic crosstalk pre-compensation, predict the impact of control on adjacent sensors based on the crosstalk transfer function, and correct the input parameters of the decoupling prediction model of adjacent sensors in advance to avoid crosstalk entering the decoupling process. Based on the real signal after dual pre-compensation, apply the coordinated control voltage to suppress cross-sensitivity and electromagnetic crosstalk chain errors, and complete the interference suppression of the target sensor.

[0080] Specifically, the steps for constructing a joint feature library include:

[0081] Each force-temperature dual-mode sensor in the sensor array is numbered, and the initial resistance of each sensor is recorded to generate a sensor sequence Q = {Q1, ..., Q...}. x ,…,Q a}; where Q x There are x force-temperature dual-mode sensors, x = 1, ..., a, where a is the number of force-temperature dual-mode sensors in the sensor array;

[0082] In scenarios with constant temperature and no additional electromagnetic interference, such as an electromagnetically shielded constant temperature and pressure chamber, a standard pressure source is used to control the pressure of a single sensor Q in the array. x Apply stepped pressure gradients sequentially according to their numbers to avoid mutual interference caused by multiple sensors being stressed simultaneously. Wait for t after each pressure level is applied. awa.1 During a specific time period, to ensure stable force on the sensor and avoid transient response interference, the sensor's resistance value is continuously collected via a data acquisition terminal. The average resistance value under the current pressure level is calculated and recorded as sensor Q. x Pressure data pairs under different pressures (P) i ,R Pi ); where i is the pressure level number, i = 1, ..., n, n is the total number of pressure levels, P i R is the pressure value corresponding to the i-th pressure level. Pi This represents the average resistance value corresponding to the i-th pressure level;

[0083] Based on the sensor number, sequentially process individual sensors Q in the array. x Apply a stepped temperature gradient, waiting for t after each temperature increment. awa.2 During the specified time period, ensure the sensor temperature is consistent with the environment to avoid temperature lag, continuously collect the sensor resistance value, calculate the average resistance value under the current temperature and pressure level, and record the sensor Q value. x Temperature data pairs at different temperatures (T) j ,R Tj ); where j is the temperature level number, j = 1, ..., m, m is the total number of temperature levels, T j R represents the temperature value corresponding to the j-th temperature level. Tj This represents the average resistance value corresponding to the j-th temperature level;

[0084] For sensor Q x Each was fitted separately to generate sensor Q. x Pressure sensitivity S Px Temperature coefficient of resistance S TxFurthermore, a mapping relationship between pressure and resistance, and temperature and resistance is constructed for each sensor, thereby generating an array mapping table to clarify the specific laws governing the change of resistance in a single physical quantity.

[0085] To ensure that the sensors are only affected by electromagnetic crosstalk caused by the control voltage and to avoid physical interference from entering the crosstalk data, in a scenario without pressure stimulation and with constant temperature, target sensors are selected sequentially based on their numbers to apply the control voltage. The remaining sensors are treated as adjacent sensors to monitor crosstalk. A stepped gradient control voltage is applied to the target sensor, covering the control voltage range of online operation. After each control voltage level is applied, a waiting period of t is observed. awa.3 During the time period, to ensure the stability of the crosstalk signal, the resistance values ​​of all adjacent sensors are continuously collected, the average resistance value of each adjacent sensor is calculated, and the resistance fluctuation caused by crosstalk is calculated based on the difference between the average resistance value and the initial resistance of the sensor.

[0086] Replace the target sensor and repeat the above process to cover the scenario where all sensors in the array are used as target sensors. Record the four-dimensional data pairs of target sensor number, control voltage, adjacent sensor number, and resistance fluctuation. Define the crosstalk propagation coefficient by the ratio of resistance fluctuation to control voltage. Perform linear fitting on the control voltage and resistance fluctuation of each target-adjacent sensor pair to obtain the crosstalk propagation coefficient of the corresponding target sensor, so as to construct a crosstalk association table.

[0087] Using the sensor number as the core index, a three-level structured joint feature library is constructed. The first-level directory is the sensor number, the second-level directory is the feature type, including single physical quantity response features and electromagnetic crosstalk features, and the third-level content includes pressure sensitivity, resistance temperature coefficient, pressure-resistance mapping relationship, and temperature-resistance mapping relationship stored under the single physical quantity response features, and crosstalk transmission coefficient and crosstalk association table stored under the electromagnetic crosstalk features.

[0088] Experimental data not used in the fitting were randomly selected and substituted into the joint feature library for verification. This included: calculating the theoretical resistance difference corresponding to the physical quantity based on the pressure sensitivity and resistance temperature coefficient in the joint feature library, comparing it with the actual resistance difference collected, and calculating the resistance deviation. When the resistance deviation was less than the preset resistance deviation threshold, the physical quantity verification passed; otherwise, data was re-collected, and cross-sensitive features were extracted until the verification passed. Simultaneously, the theoretical resistance fluctuation corresponding to the control voltage was calculated based on the crosstalk transmission coefficient in the joint feature library, and compared with the actual resistance fluctuation collected to calculate the fluctuation deviation. When the fluctuation deviation was less than the preset fluctuation deviation threshold, the crosstalk verification passed; otherwise, data was re-collected, and electromagnetic crosstalk features were extracted until the verification passed.

[0089] Specifically, the steps of two-stage decoupling include:

[0090] The system collects actual data from the online operation of the array, including real-time raw resistance signals, real-time control voltage matrix, and real-time environmental interference monitoring values. These data are categorized into steady-state and transient signals based on signal fluctuation intensity. An offline reference set is constructed based on the clean data collected in the offline controlled environment and a joint feature library. Based on the inherent noise characteristics of the sensors in the offline reference set, dynamic window filtering is applied to different types of real-time signals. A smaller filtering window is used when the steady-state signal or noise is low to avoid excessive smoothing and loss of signal details. A larger window is used when the transient signal or noise is high, and wavelet thresholding is superimposed for noise reduction to eliminate online random interference.

[0091] Since the initial resistance of offline calibration may drift due to changes in online ambient temperature and sensor aging, real-time benchmark calibration is performed using the offline reference set as a reference. The actual temperature in the actual data is obtained, and the initial temperature calibrated in the offline reference set is called synchronously. The difference between the actual temperature and the initial temperature is calculated as the calibration temperature difference. The dynamic reference resistance is calculated by multiplying the sensor's initial resistance by the resistance temperature coefficient and the calibration temperature difference, and adding the initial resistance to it. This dynamic reference resistance replaces the fixed historical initial resistance, ensuring the accuracy of subsequent resistance change calculations. At the same time, the actual total resistance change is calculated by the difference between the filtered actual resistance and the dynamic reference resistance.

[0092] Based on sensor number and real-time timing, the offline reference set is bound to the processed actual data, such as the filtered resistance signal, dynamic reference resistor, and real-time control voltage matrix, to generate a dynamic matching table, ensuring that the actual data of each frame accurately matches the offline reference parameters.

[0093] To eliminate force-temperature cross-sensitivity interference in the actual signal, a sensitivity correction coefficient is obtained by taking the fluctuation ratio of pressure sensitivity for every 1°C change in temperature. The corrected real-time pressure sensitivity is calculated by multiplying the reference pressure sensitivity calibrated by the sensor, the sensitivity correction coefficient, and the calibration temperature difference, and then adding the reference pressure sensitivity. Based on the linear decoupling model of the sensitivity matrix, a solution formula for the sensitivity matrix is ​​constructed, as shown below:

[0094]

[0095] In the formula, ΔR total_real This represents the actual change in total resistance. R represents the real-time pressure sensitivity at the current temperature. 0_real S is the dynamic reference resistance of the current sensor. T R is the temperature coefficient of resistance of the current sensor. 0_real ×S T For real-time temperature sensitivity, ΔP pre Let ΔT be the initial pressure change to be solved, representing the difference between the initial decoupled pressure and the initial online pressure.pre The initial temperature change to be solved represents the difference between the initial decoupling temperature and the initial online temperature.

[0096] Based on the signal fluctuation intensity and pressure and temperature change rate in the online actual data, the decoupling weights are dynamically allocated. If the real-time signal is a steady-state signal and the pressure and temperature change rates are balanced, the pressure and temperature decoupling weights are equal. If the real-time signal is a transient signal or a certain physical quantity has a higher change rate, the decoupling weight of that physical quantity is increased to ensure its decoupling accuracy. Then, the least squares weighted iteration method is used to solve the sensitivity matrix to obtain the preliminary pressure and temperature changes.

[0097] Based on the real-time initial pressure and real-time ambient temperature in the actual data, combined with the corresponding pressure and temperature changes, the preliminary actual pressure and preliminary actual temperature are calculated, and the reasonableness range is verified.

[0098] To eliminate electromagnetic crosstalk deviations in the initial actual physical quantities, based on the online real-time position offset of the current sensor and adjacent sensors, corresponding position weight factors are configured for each adjacent sensor and the current sensor. Once the online real-time position offset exceeds the historical calibration deviation threshold, the crosstalk is corrected according to the historical attenuation coefficient. The product of the offset and the attenuation coefficient is subtracted from the original position weight factor to avoid inaccurate crosstalk calculation due to position changes. The crosstalk transmission coefficient of adjacent sensors relative to the current sensor, the online real-time control voltage of adjacent sensors, and the corresponding position weight factors are combined to perform weighted summation to calculate the actual crosstalk resistance change.

[0099] The pressure change rate and temperature change rate under actual online operating conditions are obtained. Combined with the crosstalk distribution ratio under different operating conditions in offline calibration, the distribution ratio of crosstalk deviation in pressure and temperature is obtained by comparing the pressure change rate and temperature change rate. Thus, the pressure deviation and temperature deviation caused by crosstalk are calculated. If the pressure change rate is greater than the temperature change rate, the larger distribution ratio is assigned to the pressure change rate; otherwise, the larger distribution ratio is assigned to the temperature change rate to avoid deviation caused by uniform distribution. The pressure deviation is obtained by multiplying the ratio of the actual crosstalk resistance change to the real-time pressure sensitivity at the current temperature by the corresponding distribution ratio. The temperature deviation is obtained by multiplying the ratio of the actual crosstalk resistance change to the real-time temperature sensitivity by the corresponding distribution ratio.

[0100] The true pressure and true temperature are obtained by subtracting the corresponding crosstalk deviation from the preliminary actual physical quantities.

[0101] Specifically, the steps for calculating the co-regulatory factor include:

[0102] The actual pressure and temperature of the two-stage decoupling are retrieved as real-time physical quantities to ensure that the data is free from cross-sensitivity and crosstalk interference. The target pressure and target temperature of the array during online operation are obtained as target physical quantities. The cross-interference deviation threshold and crosstalk correlation deviation threshold are retrieved from the joint feature library as deviation benchmarks. Based on the difference between the target physical quantity and the real-time physical quantity, the cross-interference deviation is calculated, including pressure cross-interference deviation and temperature cross-interference deviation. The positive and negative values ​​of the deviation represent whether the physical quantity needs to be increased or decreased, respectively.

[0103] Based on the ratio of the cross-interference deviation value to the deviation benchmark, the deviation levels are divided into slight deviation, moderate deviation, and severe deviation. A deviation information table is constructed according to the sensor number and deviation level, including sensor number, real-time physical quantity, target physical quantity, and deviation level.

[0104] The real-time pressure sensitivity and real-time temperature sensitivity are obtained. Based on the product of the cross-interference deviation and the real-time sensitivity, the target resistance change required to compensate for the cross-interference deviation is calculated, including the resistance change required to compensate for the pressure deviation and the resistance change required to compensate for the temperature deviation. Specifically, the pressure cross-interference deviation is multiplied by the real-time pressure sensitivity to obtain the target resistance change required to compensate for the pressure deviation; the temperature cross-interference deviation is multiplied by the real-time temperature sensitivity to obtain the target resistance change required to compensate for the temperature deviation.

[0105] The crosstalk correlation table is called, and the base pressure control voltage and base temperature control voltage corresponding to the target resistance change are found through the interpolation algorithm. If the target resistance change of pressure and temperature are in the same direction, such as both requiring an increase in resistance, the total base voltage adjustment is the sum of the base pressure control voltage and the base temperature control voltage. If the directions are opposite, the control voltage with the larger absolute value is taken as the dominant one, and a portion of the smaller one is added to avoid voltage cancellation leading to insufficient compensation.

[0106] Based on the deviation level in the deviation information table, the corresponding deviation level correction coefficient is called, and combined with the basic total voltage adjustment, the preliminary cross-interference control factor is obtained, which is the voltage adjustment to be applied, in order to compensate for the deviation caused by cross sensitivity or aging.

[0107] The sensor's safe voltage range is retrieved to avoid damage to the hardware due to excessive voltage. If the cross-interference control factor causes the final voltage to exceed the safe range, the verification ratio of the safe threshold is used as the corrected cross-interference control factor. At the same time, the cross-interference control factor is compared with the historical voltage adjustment. If the deviation is too large, the cross-interference deviation is recalculated. Finally, the cross-interference control factor and the corresponding calculation report are output.

[0108] Based on the product of the crosstalk adjustment factor and the crosstalk transmission coefficient, the change in crosstalk resistance of adjacent sensors is calculated. Combined with the real-time pressure sensitivity of adjacent sensors, the pressure deviation of adjacent sensors caused by crosstalk is calculated based on the ratio of the change in crosstalk resistance to the real-time pressure sensitivity. Adding the background crosstalk deviation of adjacent sensors, the predicted total crosstalk deviation is obtained.

[0109] If the total crosstalk deviation is not greater than the upper limit of crosstalk deviation, there is no need to limit the voltage, and the crosstalk constraint factor is 1. If the total crosstalk deviation is greater than the upper limit of crosstalk deviation, the voltage limit coefficient is obtained based on the deviation rate between the upper limit of crosstalk deviation and the background crosstalk deviation. Based on the correspondence between the crosstalk deviation recorded offline and the decoupling correction ratio, the auxiliary correction coefficient is obtained by interpolation. This is used to correct the decoupling model of adjacent sensors, make up for the compensation gap caused by the voltage limit, and finally form a set of crosstalk constraint factors including the voltage limit coefficient, the auxiliary correction coefficient, and the total crosstalk deviation.

[0110] In addition to directly adjacent sensors, the crosstalk of indirectly adjacent sensors is predicted. For example, if the target sensor affects sensor A, and sensor A then affects sensor B, the indirect crosstalk deviation is calculated based on the product of the crosstalk pressure deviation of the target sensor and the crosstalk transmission coefficient of the two adjacent sensors. If the indirect crosstalk deviation exceeds the indirect ratio of the upper limit of the crosstalk deviation, the voltage limit coefficient is further reduced to ensure that the entire crosstalk path is within the safe range.

[0111] Based on the principle of prioritizing cross-interference compensation, the adjustment amount of the control voltage is obtained by multiplying the cross-interference control factor and the voltage limiting coefficient.

[0112] If the adjustment amount of the control voltage is less than the initial cross-interference control factor, the compensation deviation corresponding to the adjustment amount of the control voltage is obtained by multiplying the adjustment amount of the control voltage and the real-time pressure sensitivity and then dividing by the deviation level correction coefficient. The cross-compensation gap is calculated by the difference between the pressure cross-interference deviation and the compensation deviation.

[0113] The parameters are labeled according to priority, with the adjustment amount of the control voltage and the auxiliary correction coefficient as high priority, and the compensation gap and total crosstalk deviation as medium priority. A collaborative control factor table is generated, and a control correlation table containing the target sensor and adjacent sensors is constructed to clarify the impact of the control of each target sensor on adjacent sensors and the corresponding countermeasures.

[0114] Specifically, the steps of dual pre-compensation include:

[0115] Retrieve relevant data on the coordinated control factors, including the control voltage adjustment, auxiliary correction coefficient, cross-compensation gap, and control correlation table; retrieve real-time decoupling basic data, including real-time pressure sensitivity, real-time temperature sensitivity, and background crosstalk between adjacent sensors; retrieve historical benchmark data from the joint feature library, including array mapping table and crosstalk transfer function; and perform data verification to confirm that the control voltage adjustment is within the sensor's safe voltage range and that the drift between the real-time sensitivity and the offline calibrated benchmark sensitivity is within acceptable limits, thus avoiding pre-compensation errors caused by abnormal data.

[0116] Based on the sensor number and control timing, a pre-compensation data association table is constructed to correspond the control parameters, decoupling parameters, correction parameters of affected adjacent sensors, and historical characteristics of the target sensor one by one, ensuring that the acquisition timing of all data is synchronized, such as real-time data at the same moment, to avoid the pre-compensation and control actions being out of sync due to timing misalignment, such as using the sensitivity of the previous moment to calculate the pre-compensation deviation at the current moment.

[0117] The pre-compensation parameters are cross-interference pre-compensation deviation, electromagnetic crosstalk pre-compensation deviation, and decoupled input signal after pre-compensation, and will be initialized.

[0118] The pressure coupling resistor deviation to be supplemented is calculated based on the product of the cross-compensation gap and the real-time pressure sensitivity. The temperature coupling resistor deviation is calculated based on the product of the calibration temperature difference, the real-time temperature sensitivity, and the electrical temperature influence coefficient. The total cross-interference pre-compensation deviation is calculated by summing the pressure coupling resistor deviation and the temperature coupling resistor deviation. Among them, the real-time temperature sensitivity is the product of the dynamic reference resistor and the resistance temperature coefficient, and the electrical temperature influence coefficient is the proportion of the influence of the unit voltage on the temperature coefficient measured offline.

[0119] The original resistance signal of the target sensor after filtering is retrieved. The pre-compensated decoupled input signal is obtained by the difference between the original resistance signal and the total cross-interference pre-compensation deviation. This ensures that the force-temperature coupling error caused by the control is no longer included during decoupling. At the same time, deviation verification is performed by comparing the total cross-interference pre-compensation deviation with the cross-interference deviation threshold to ensure that it does not exceed the deviation ratio of the threshold. This avoids the distortion of the decoupled signal due to over-compensation. A cross-interference pre-compensation report is generated, which records the calculation process of various deviations, the value of the pre-compensated input signal and the verification results. The pre-compensated input signal is then pushed to the input interface of the two-stage decoupling to replace the original signal and provide a coupling-free input for subsequent decoupling.

[0120] For directly adjacent sensors, the direct crosstalk resistance deviation is obtained by multiplying the voltage adjustment amount, crosstalk transmission coefficient, and auxiliary correction coefficient. For indirectly adjacent sensors, such as when the target sensor affects sensor A, and sensor A then affects sensor B, the indirect crosstalk resistance deviation is obtained by multiplying the crosstalk deviation of the directly adjacent sensors with the crosstalk transmission coefficient between the indirectly adjacent sensors. The direct crosstalk resistance deviation and the indirect crosstalk resistance deviation are then added together to obtain the total crosstalk pre-compensation deviation for each adjacent sensor.

[0121] The filtered original resistance signal of each adjacent sensor is retrieved and the corresponding total crosstalk pre-compensation deviation is subtracted to obtain the pre-compensated decoupled input signal of the adjacent sensor. This ensures that the crosstalk caused by the target sensor control is no longer included when the adjacent sensors are decoupled. At the same time, deviation verification is performed. Based on the ratio of the total crosstalk pre-compensation deviation to the real-time pressure sensitivity of the adjacent sensors, the adjacent crosstalk deviation is calculated and compared with the crosstalk deviation threshold to ensure that the adjacent crosstalk deviation does not exceed the threshold deviation ratio. The background crosstalk of the adjacent sensors after pre-compensation is compared to ensure that the residual crosstalk after pre-compensation is within an acceptable range, avoiding the crosstalk residue from affecting the measurement of adjacent sensors. Thus, an electromagnetic crosstalk pre-compensation report is generated, recording the calculation process of various crosstalk deviations of each adjacent sensor, the pre-compensated input signal and the verification results. The pre-compensated input signal is pushed to the input interface of the corresponding adjacent sensor's two-stage decoupling to replace the original signal, completing the crosstalk pre-compensation for all affected sensors.

[0122] The reference voltage of the target sensor is obtained from the actual data. The final control voltage to be applied is calculated by summing the reference voltage and the control voltage adjustment. The control voltage is applied to the target sensor through the voltage drive terminal of the array, ensuring that the timing of voltage application is synchronized with the decoupling cycle and avoiding control lag.

[0123] The system acquires the real-time resistance signal of the target sensor after adjustment, performs two-stage decoupling, and outputs the actual pressure and temperature after adjustment. It also acquires the real-time resistance signal of adjacent sensors after adjustment, inputs it into the two-stage decoupling operation after crosstalk pre-compensation, and outputs the actual physical quantities of adjacent sensors. This verifies whether the deviation between the actual physical quantity of the target sensor and the preset target value has been reduced to within the crosstalk deviation threshold, confirming the effectiveness of crosstalk compensation. Furthermore, it verifies whether the crosstalk deviation of adjacent sensors remains within the crosstalk deviation threshold, confirming the effectiveness of crosstalk suppression. If the deviation does not meet the standard, the system returns to readjust the pre-compensation deviation calculation logic, such as correcting the sensitivity parameter or crosstalk transmission coefficient, until the standard is met.

[0124] Generate a dual pre-compensation and control execution report, recording the calculation process of dual pre-compensation, the value of the control voltage, the changes in physical quantity deviations before and after control, and the changes in crosstalk between adjacent sensors; store the report, along with all pre-compensation parameters and decoupling data, into the control log database.

[0125] Example 2

[0126] Please see Figure 4 Another embodiment of the present invention provides: a decoupled prediction interference suppression and control system, comprising: an offline calibration module, an online simulation module, a verification module, and a data storage module;

[0127] The offline calibration module provides a constant interference environment through an electromagnetically shielded constant temperature and pressure chamber. Combined with a standard pressure source, standard temperature control equipment, and data acquisition terminal, it applies gradient pressure and gradient temperature to each force-temperature dual-mode sensor to collect resistance changes, calculates pressure sensitivity and temperature coefficient, and establishes a physical quantity-resistance mapping relationship. At the same time, it applies gradient control voltage to the sensor without physical stimulation to collect crosstalk data from adjacent sensors and calculates the crosstalk transmission coefficient. Finally, it integrates the two types of features to construct a three-level structured joint feature library, providing an offline benchmark for subsequent interference differentiation and control, and avoiding confusion between cross-sensitivity and crosstalk features.

[0128] The online simulation module includes a preprocessing unit, a two-stage decoupling unit, a coordinated control unit, and a dual pre-compensation unit;

[0129] The preprocessing unit is used to acquire the real-time raw resistance signal, real-time control voltage matrix and real-time environmental interference monitoring value of the array online through the data acquisition terminal. Combining the inherent noise characteristics of the sensor in the joint feature library, it uses small window filtering for steady-state signals and large window + wavelet threshold denoising for transient / strong noise signals. At the same time, it calculates the dynamic reference resistance based on the real-time ambient temperature and the initial temperature and resistance temperature coefficient in the joint feature library to correct the offline initial resistance drift. Finally, it outputs the filtered pure resistance signal and the actual total resistance change, eliminating the influence of online random interference and reference drift on decoupling.

[0130] The dual-stage decoupling unit is used to construct a sensitivity matrix containing real-time pressure sensitivity and real-time temperature sensitivity based on cross-sensitive features in the joint feature library and real-time data. It separates the resistance changes caused by pressure and temperature through the least squares weighted iterative method to obtain the preliminary actual physical quantities. Then, it calculates the actual crosstalk resistance change by combining the crosstalk transmission coefficient in the joint feature library and the real-time position offset of the sensor. It allocates the crosstalk deviation according to the rate of change of the physical quantity and removes it from the preliminary physical quantities. Finally, it outputs the real pressure and real temperature, which solves the problem of difficulty in distinguishing the resistance change caused by pressure / temperature cross-action and the electromagnetic crosstalk resistance fluctuation.

[0131] The collaborative control unit is used to retrieve the real physical quantities and array preset target physical quantities of the dual-stage decoupling output, calculate the cross-interference deviation and classify the deviation level, combine the physical quantity-resistance mapping relationship in the joint feature library and the interpolation algorithm to obtain the basic control voltage adjustment amount, and then predict the total cross-interference deviation of the adjustment amount on the adjacent sensors based on the cross-interference transmission coefficient. When the threshold is exceeded, the voltage limit coefficient and auxiliary correction coefficient are calculated. Finally, a collaborative control factor table containing the control voltage adjustment amount, auxiliary correction coefficient, cross-compensation gap and target-adjacent sensor correlation relationship is generated to ensure that the control both compensates for cross-interference and does not amplify the cross-interference of adjacent sensors.

[0132] The dual pre-compensation unit is used to calculate the pressure coupling resistance deviation and temperature coupling resistance deviation caused by the adjustment of the control voltage based on the collaborative control factor and the crosstalk transfer function in the joint feature library, and deduct them from the decoupling input of the target sensor. At the same time, it calculates the direct crosstalk and indirect crosstalk resistance deviation of adjacent sensors and deducts them from their decoupling input. Then, the pre-compensated decoupling input signal is pushed to the dual-stage decoupling module to eliminate the crosstalk and crosstalk caused by the control from the source and enter the decoupling process to avoid interference chain amplification.

[0133] The verification module is used to superimpose the adjustment amount of the control voltage in the collaborative control factor onto the reference voltage of the target sensor through the voltage drive terminal and apply it to ensure that the voltage application timing is synchronized with the two-stage decoupling cycle. At the same time, it collects the real-time resistance signals of the target sensor and adjacent sensors after control, outputs the controlled physical quantity after two-stage decoupling, and verifies whether the deviation between the physical quantity of the target sensor and the target value has been reduced to within the cross-interference threshold and whether the crosstalk of adjacent sensors has been maintained within the crosstalk threshold. If the standard is not met, it returns to the adjustment pre-compensation calculation logic to ultimately ensure the control and pre-compensation effect.

[0134] The data storage module is used to store the joint feature library, online acquired raw and preprocessed data, two-stage decoupling results, collaborative control factors, dual pre-compensation parameters and control verification data. It also generates cross-interference pre-compensation reports, electromagnetic crosstalk pre-compensation reports, dual pre-compensation and control execution reports to support subsequent system optimization and data traceability.

[0135] Working principle and effects:

[0136] A joint feature library is constructed through offline calibration. Cross-sensitivity features are extracted by simulating pressure and temperature effects separately, and crosstalk features are extracted by simulating pure electromagnetic crosstalk, thus solving the problem of confusion between the two types of interference features and providing a benchmark for subsequent differentiation. During online operation, sensor signals are acquired and dynamically filtered for noise reduction and reference resistors are calibrated. Then, through two-stage decoupling, a sensitivity matrix is ​​first constructed based on the feature library to remove cross-sensitivity interference, and then the crosstalk influence is eliminated by combining crosstalk features to output the true physical quantity. Based on the deviation between the true value and the target value, a collaborative control factor is calculated in conjunction with the feature library. This determines the cross-interference compensation voltage and predicts and constrains the voltage's crosstalk to adjacent sensors to avoid control amplification coupling error. Through double pre-compensation, the cross-interference and crosstalk deviation caused by control are deducted in advance before the control voltage is applied, completely breaking the double chain of force-temperature cross-sensitivity and electromagnetic crosstalk, effectively ensuring the overall measurement accuracy in array collaborative scenarios.

[0137] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for controlling interference suppression based on decoupling prediction, characterized in that, include: During the offline calibration phase of the array, gradient pressure and gradient temperature are applied separately to each force-temperature dual-mode sensor to establish a unique mapping relationship between pressure and resistance, and temperature and resistance, in order to obtain cross-sensitivity features. At the same time, pure electromagnetic crosstalk simulation is performed to extract electromagnetic crosstalk features and construct a joint feature library. Based on the joint feature library, the original resistance signals and real-time control voltage matrix of all sensors in the current array are obtained. Through two-stage decoupling, the signal mixing caused by cross sensitivity is removed, and the interference of crosstalk on pressure and temperature values ​​is eliminated to generate the predicted real pressure and temperature. The deviation between the actual measured value and the target value is obtained. Combined with the joint feature library and the addition of constraint factors to limit the control voltage amplitude, the synergistic control factor that takes into account both crosstalk compensation and crosstalk suppression is calculated. Based on the coordinated control factor, dual pre-compensation is performed. For cross-interference pre-compensation, the resistance deviation caused by pressure / temperature coupling is predicted based on the sensitivity matrix and subtracted from the input of the decoupling prediction model. For electromagnetic crosstalk pre-compensation, the influence of control on adjacent sensors is predicted based on the crosstalk transfer function, and the input parameters of the decoupling prediction model of adjacent sensors are corrected. Based on the real signal after dual pre-compensation, a coordinated control voltage is applied.

2. The interference suppression and control method for decoupling prediction according to claim 1, characterized in that, The steps to build a joint feature library include: Each force-temperature dual-mode sensor in the sensor array is numbered to generate a sensor sequence; For sensors Apply a stepped pressure gradient and maintain it at each pressure level. For each time period, calculate the average resistance value at the corresponding pressure level to generate the sensor. Pressure data pairs under different pressures; For sensors Apply a stepped temperature gradient and maintain it at each temperature level. Time period; Calculate the average resistance value at the corresponding temperature and pressure levels to generate the sensor. Temperature data pairs at different temperatures; Computational Sensors Pressure sensitivity Temperature coefficient of resistance And generate an array mapping table.

3. The interference suppression and control method for decoupling prediction according to claim 2, characterized in that, The steps for building a joint feature library also include: Define a target sensor and adjacent sensors, apply a stepped gradient control voltage to the target sensor, and maintain the voltage at each control voltage level. Time period; Calculate the average resistance value and resistance fluctuation of each adjacent sensor to construct a four-dimensional data pair; Based on the ratio of resistance fluctuation to control voltage, linear fitting is performed to calculate the crosstalk transmission coefficient of the target sensor in order to construct a crosstalk correlation table. A three-level structured joint feature library is constructed using the sensor number as the core index; Randomly select experimental data that were not used for fitting, calculate resistance deviation and fluctuation deviation. Once the resistance deviation is not less than the resistance deviation threshold or the fluctuation deviation is not less than the fluctuation deviation threshold, re-collect data and perform feature extraction until the verification is successful.

4. The interference suppression and control method for decoupling prediction according to claim 3, characterized in that, The steps of two-stage decoupling include: Collect actual data from the online operation of the array, and construct an offline benchmark set based on the data from the offline calibration phase and the joint feature library; Obtain the actual temperature from the actual data, calculate the calibration temperature difference, and thus obtain the dynamic reference resistor; Calculate the actual total resistance change and obtain the sensitivity correction coefficient to calculate the real-time pressure sensitivity; Dynamically allocate decoupling weights and use the least squares weighted iterative method to calculate pressure and temperature changes; By combining the real-time initial pressure and real-time ambient temperature from the actual data, the preliminary actual pressure and preliminary actual temperature are calculated.

5. The interference suppression and control method for decoupling prediction according to claim 4, characterized in that, The two-stage decoupling process also includes: Based on the online real-time position offset between the current sensor and adjacent sensors, configure the corresponding position weight factor and calculate the actual crosstalk resistance change. Obtain the pressure change rate and temperature change rate under actual online working conditions, configure the distribution ratio of crosstalk deviation on pressure and temperature, and calculate the pressure deviation and temperature deviation caused by crosstalk. Based on the difference between the initial actual physical quantities and the crosstalk deviation, the true pressure and true temperature are obtained.

6. The interference suppression and control method for decoupling prediction according to claim 5, characterized in that, The steps for calculating the co-regulatory factor include: Acquire real-time physical quantities, target physical quantities, and deviation benchmarks; calculate cross-interference deviation based on the difference between the target physical quantity and the real-time physical quantity. Classify the deviation levels and construct a deviation information table; Obtain real-time pressure sensitivity and real-time temperature sensitivity, and calculate the target resistance change required to compensate for cross-interference deviation; Call the crosstalk correlation table to obtain the base pressure control voltage and base temperature control voltage corresponding to the target resistance change, and calculate the total base voltage adjustment. Based on the aforementioned deviation level, the corresponding deviation level correction coefficient is invoked to calculate the preliminary cross-interference control factor.

7. The interference suppression and control method for decoupling prediction according to claim 6, characterized in that, The steps for calculating the co-regulatory factor also include: Calculate the change in crosstalk resistance of adjacent sensors, combine it with the real-time pressure sensitivity of adjacent sensors, calculate the pressure deviation of adjacent sensors, and combine it with the background crosstalk deviation to obtain the predicted total crosstalk deviation. If the total crosstalk deviation is not greater than the upper limit of crosstalk deviation, there is no need to limit the voltage, and the crosstalk constraint factor is 1. If the total crosstalk deviation is greater than the upper limit of crosstalk deviation, calculate the voltage limiting coefficient and the auxiliary correction coefficient, and generate a set of crosstalk constraint factors. Calculate the indirect crosstalk deviation, and once the indirect crosstalk deviation exceeds the indirect ratio of the upper limit of the crosstalk deviation, reduce the voltage limit factor; Calculate the adjustment amount of the control voltage. Once the adjustment amount of the control voltage is less than the cross-interference control factor, calculate the compensation deviation corresponding to the adjustment amount of the control voltage and calculate the cross-compensation gap. Generate a table of synergistic regulation factors and construct a regulation correlation table that includes the target sensor and adjacent sensors.

8. The interference suppression and control method for decoupling prediction according to claim 7, characterized in that, The steps of dual pre-compensation include: Configure and initialize the pre-compensation parameters, and calculate the total cross-interference pre-compensation deviation based on the pressure and temperature coupling resistance deviation. The pre-compensated decoupled input signal is obtained by measuring the difference between the original resistance signal of the target sensor and the total cross-interference pre-compensation deviation, and the deviation is verified. Obtain the reference voltage of the target sensor from actual data, and calculate the final control voltage to be applied; The regulated voltage is applied to the target sensor via the voltage drive terminal of the array.

9. The interference suppression and control method for decoupling prediction according to claim 8, characterized in that, The steps of dual pre-compensation also include: The real-time resistance signal after the target sensor is regulated is collected, and two-stage decoupling is performed to obtain the real pressure and real temperature after regulation. Collect real-time resistance signals after adjustment of adjacent sensors, and obtain the actual physical quantities of adjacent sensors; Verify whether the deviation between the actual physical quantity of the target sensor and the preset target value has been reduced to within the cross-interference deviation threshold to confirm that the cross-interference compensation is effective; Verify whether the crosstalk deviation between adjacent sensors remains within the crosstalk deviation threshold to confirm that crosstalk suppression is effective; If the deviation does not meet the standard, return to readjust the pre-compensation deviation calculation logic until the standard is met.

10. A decoupled prediction interference suppression control system, used to implement the decoupled prediction interference suppression control method as described in any one of claims 1-9, characterized in that, include: The system includes an offline calibration module, an online simulation module, a verification module, and a data storage module. The offline calibration module is used to apply gradient values ​​individually to each force-temperature dual-mode sensor to construct a joint feature library; The online simulation module includes a preprocessing unit, a two-stage decoupling unit, a collaborative control unit, and a dual pre-compensation unit; The preprocessing unit is used to collect data from the online operation of the array, correct the offline initial resistance drift, and generate a pure resistance signal and the actual total resistance change. The two-stage decoupling unit is used to construct a sensitivity matrix and calculate the initial actual physical quantities and crosstalk resistance changes to obtain the true pressure and true temperature. The coordinated control unit is used to calculate the cross-interference deviation and classify the deviation level, obtain the basic control voltage adjustment amount, and generate a coordinated control factor table. The dual pre-compensation unit is used to generate a pre-compensated decoupled input signal based on the resistance deviation; The verification module is used to apply the adjustment amount of the control voltage, collect the real-time resistance signal, and verify the compliance of the physical quantities after the two-stage decoupled output control. The data storage module is used to store relevant data on interference suppression and regulation.

Images