Flexible pressure sensor performance offset compensation method based on neural network

Through the performance offset compensation method of flexible pressure sensors based on neural network, the performance offset of flexible pressure sensors is monitored and predicted in real time, which solves the problem of calibration of flexible pressure sensors and extends the service life of the equipment.

CN118980452APending Publication Date: 2024-11-19TONGJI UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411068239.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-06
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The performance of flexible pressure sensors is easily affected by ambient temperature and humidity, and long-term high-frequency pressures will lead to performance deviations. The existing technology lacks mature calibration solutions, resulting in a short service life of wearable devices and increases maintenance costs.

Method used

The performance offset compensation method of flexible pressure sensor based on neural network is adopted. By constructing a multi-layer multi-parameter sensor, performance offset is monitored and predicted in real time, calibration is performed using compensation algorithms, and calibration data is used for update optimization of prediction models.

Benefits of technology

High-precision calibration of flexible pressure sensors is achieved, extending the service time of the equipment and improving the normal service time after calibration.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118980452A_ABST
    Figure CN118980452A_ABST
Patent Text Reader

Abstract

The invention discloses a flexible pressure sensor performance offset compensation method based on a neural network, and the method comprises the steps: designing a three-layer multi-parameter flexible pressure sensor which can read temperature, humidity and pressure data at the same time, employing an MCU to control the storage and reading of sensor performance parameters in a server, and enabling the sensor MCU to monitor the performance offset condition in real time, when the offset is too large, the calibration step can be triggered, a performance offset value which can be generated in the future is predicted through a neural network performance prediction model of a server side, then a designed performance compensation algorithm is used for calibration, the calibration value is written into a sensor MCU, and the influence of temperature and humidity in the environment and long-time use working conditions on the sensor performance can be considered at the same time. According to the method and the device, the performance deviation condition of the sensor is monitored in real time, the deviation possibly generated in future use is predicted, high-precision calibration of the flexible pressure sensor can be realized, and meanwhile, the normal use duration after calibration is greatly prolonged.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of sensors and communications, and in particular to a method for compensating performance offset of a flexible pressure sensor based on a neural network. Background Art

[0002] With the development of wearable devices, various flexible sensors are used in the field of human biological signal detection. The performance of flexible pressure sensors is easily affected by the ambient temperature and humidity, and long-term high-frequency pressure can also cause sensor performance deviation. Currently, there is no mature flexible pressure sensor calibration solution on the market. Wearable devices based on flexible pressure sensors usually have a short service life, which increases the maintenance cost of device users.

[0003] Therefore, how to simultaneously calibrate the drift generated during the use of flexible pressure sensors with high precision and increase the usage time of wearable devices has become a crucial issue. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method for compensating for the performance offset of a flexible pressure sensor based on a neural network, which can monitor the performance offset of the sensor in real time, and predict the offset that may occur in future use, so as to achieve high-precision calibration of the flexible pressure sensor, and at the same time greatly increase the length of time that the sensor can be used normally after calibration. In order to achieve the above-mentioned purpose and other advantages according to the present invention, a method for compensating for the performance offset of a flexible pressure sensor based on a neural network is provided, comprising:

[0005] S1. Construction of multi-layer multi-parameter flexible sensor;

[0006] S2, read the temperature, humidity, usage time and pressure parameters in the sensor through the sensor MCU, and input the parameters into the prediction model based on the neural network;

[0007] S3, the influence curve of multiple parameters fitted by the prediction model on the pressure detection performance;

[0008] S4. Predict the future pressure detection performance offset of the sensor based on historical working conditions;

[0009] S5, output the predicted values ​​of three parameters: the offset of the effective area of ​​the sensor conductive layer, the offset of the distance between the capacitor plates, and the linear effect of temperature and humidity on the sensor performance;

[0010] S6. The output prediction value will be written into the sensor MCU and calculated using the compensation algorithm to check whether it is qualified. If it is unqualified, the next round of calibration process will be carried out; the data of each calibration will be used to update and optimize the prediction model.

[0011] Preferably, the sensor in step S1 includes a humidity detection layer, a temperature detection layer and a pressure detection layer, and the humidity detection layer, the temperature detection layer and the pressure detection layer are all connected to the sensor MCU, and the sensor MCU saves the data to the server after reading the parameters.

[0012] Preferably, when the sensor MCU reads the data in the sensor and the server respectively, a comparison is performed to determine whether the deviation is too large. If it is determined that the deviation is too large, a calibration request signal is sent to the mobile phone via Bluetooth; if it is determined that the deviation is not too large, the sensor MCU continues to read the sensor and the server respectively.

[0013] Preferably, the compensation algorithm in step S6 comprehensively considers the ambient temperature and humidity and the historical performance changes of the sensor, and then writes the calculated performance parameters into the MCU of the sensor via Bluetooth.

[0014] Preferably, the compensation algorithm compensates for the performance drift of the capacitive flexible pressure sensor, and the specific compensation algorithm formula is as follows:

[0015]

[0016] Among them, ε 0 is the vacuum dielectric constant, ε r is the relative dielectric constant of the dielectric material, A is the effective area between the conductive layers, ΔA is the offset of the effective area between the conductive layers after the sensor has been used for a period of time, d is the distance between the capacitor plates,

[0017] Δd is the offset of the distance between the capacitor plates after the sensor has been used for a period of time, and O is the linear effect of temperature and humidity on the performance of the pressure sensor.

[0018] Compared with the prior art, the present invention has the following beneficial effects: BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of a multi-parameter flexible sensor structure supporting temperature, humidity and stress monitoring according to a performance offset compensation method of a flexible pressure sensor based on a neural network of the present invention;

[0020] Figure 2 A real-time monitoring flow chart of sensor performance deviation of the method for compensating performance deviation of a flexible pressure sensor based on a neural network according to the present invention;

[0021] Figure 3 A real-time monitoring data structure diagram of sensor performance deviation according to the flexible pressure sensor performance deviation compensation method based on neural network of the present invention;

[0022] Figure 4The present invention is a flow chart of a flexible pressure sensor compensation algorithm of a flexible pressure sensor performance offset compensation method based on a neural network. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Reference Figure 1-4 , a method for compensating performance offset of a flexible pressure sensor based on a neural network, comprising the following steps:

[0025] S1. Construct a multi-layer multi-parameter flexible sensor; the sensor includes a humidity detection layer, a temperature detection layer and a pressure detection layer, and the humidity detection layer, the temperature detection layer and the pressure detection layer are all connected to the sensor MCU. After the sensor MCU reads the parameters, it saves the data to the server. Each layer is responsible for detecting different parameters. These layers can independently respond to their respective parameter changes through different physical mechanisms including resistance changes and capacitance changes. The MCU can process electrical signals of three parameters: temperature, humidity and pressure at the same time. After reading the sensor parameters, the MCU transmits the data to the mobile phone via Bluetooth, and the mobile phone saves the data to the server through cloud services. Specifically, from bottom to top, the first layer contains piezoresistive materials for stress detection, the second layer contains materials sensitive to temperature changes for temperature detection, and the third layer contains materials sensitive to humidity changes for humidity detection.

[0026] S2. Read the temperature, humidity, usage time and pressure parameters in the sensor through the sensor MCU, and input the parameters into the prediction model based on the neural network; the sensor MCU is connected to the Bluetooth module to transmit the performance data to the server for storage, and can periodically read the latest sensor performance data and algorithm update push in the server. The Bluetooth protocol used is low-power Bluetooth. The performance parameters are mainly the sensor electrical signal data when there is no additional pressure. The data transmitted to the server includes the upload time and the performance parameters of the corresponding time.

[0027] S3, the influence curve of multiple parameters fitted by the prediction model on the pressure detection performance;

[0028] S4. Predict the future pressure detection performance offset of the sensor based on historical working conditions;

[0029] S5, output the predicted values ​​of three parameters: the offset of the effective area of ​​the sensor conductive layer, the offset of the distance between the capacitor plates, and the linear effect of temperature and humidity on the sensor performance;

[0030] S6. The output prediction value will be written into the sensor MCU and checked for qualification after calculation using the compensation algorithm. If it fails, the next round of calibration process will be carried out; the data of each calibration will be used to update and optimize the prediction model. This patent mainly compensates for the performance drift of capacitive flexible pressure sensors. The compensation algorithm is as follows:

[0031]

[0032] Among them, ε 0 is the vacuum dielectric constant, ε r is the relative dielectric constant of the dielectric material, A is the effective area between the conductive layers, ΔA is the offset of the effective area between the conductive layers after the sensor has been used for a period of time, d is the distance between the capacitor plates,

[0033] Δd is the offset of the distance between the capacitor plates after the sensor has been used for a period of time, and O is the linear effect of temperature and humidity on the performance of the pressure sensor. The values ​​of the three parameters ΔA, Δd, and O are all derived from the output of the performance offset prediction model based on the neural network.

[0034] When the sensor MCU reads the data from the sensor and the server respectively, it compares and determines whether the deviation is too large. If it is determined that the deviation is too large, it sends a calibration request signal to the mobile phone via Bluetooth; if it is determined that the deviation is not too large, the sensor MCU continues to read the sensor and the server respectively. Specifically, when the collected sensor electrical signal data without additional pressure deviates too much from the sensor electrical signal data without additional pressure saved in the server the last time. The sensor MCU will periodically compare the pressure detection performance parameters during operation with the latest parameters read from the server. If the deviation is too large, the calibration process is triggered, realizing real-time monitoring of sensor performance deviation. Figure 3 As shown, the data structure sent is a series of values ​​of the sensor electrical signal as the pressure changes, and the starting value is the electrical signal value when there is no additional pressure.

[0035] In summary, this application designs a three-layer multi-parameter flexible pressure sensor that can read temperature, humidity and pressure data at the same time, and uses MCU to control the storage and reading of sensor performance parameters on the server. The sensor MCU can monitor the performance offset in real time. When the offset is too large, it will trigger the calibration step. The performance offset value that will be generated in the future is predicted by the neural network performance prediction model on the server, and then the designed performance compensation algorithm is used for calibration, and the calibration value is written into the sensor MCU. Using the method of the present invention, the influence of temperature and humidity in the environment and long-term use conditions on sensor performance can be considered at the same time, the performance offset of the sensor can be monitored in real time, and the offset that may be generated in future use can be predicted, which can achieve high-precision calibration of the flexible pressure sensor, and at the same time greatly increase the length of time that can be used normally after calibration.

[0036] The number of devices and processing scales described here are used to simplify the description of the present invention, and the application, modification and variation of the present invention will be obvious to those skilled in the art.

[0037] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A method for compensating performance offset of a flexible pressure sensor based on a neural network, characterized in that: The following steps are involved: S1. Construction of multi-layer multi-parameter flexible sensor; S2, read the temperature, humidity, usage time and pressure parameters in the sensor through the sensor MCU, and input the parameters into the prediction model based on the neural network; S3, the influence curve of multiple parameters fitted by the prediction model on the pressure detection performance; S4. Predict the future pressure detection performance offset of the sensor based on historical working conditions; S5, output the predicted values ​​of three parameters: the offset of the effective area of ​​the sensor conductive layer, the offset of the distance between the capacitor plates, and the linear effect of temperature and humidity on the sensor performance; S6. The output predicted value will be written into the sensor MCU and then checked for qualification after calculation using the compensation algorithm. If it fails, the next round of calibration process will be carried out. The data from each calibration will be used to update and optimize the prediction model.

2. A method for compensating performance offset of a flexible pressure sensor based on a neural network as claimed in claim 1, characterized in that: In step S1, the sensor includes a humidity detection layer, a temperature detection layer and a pressure detection layer, and the humidity detection layer, the temperature detection layer and the pressure detection layer are all connected to the sensor MCU. After the sensor MCU reads the parameters, it saves the data to the server.

3. A method for compensating performance offset of a flexible pressure sensor based on a neural network as claimed in claim 2, characterized in that: When the sensor MCU reads the data from the sensor and the server respectively, it compares and determines whether the deviation is too large. If it is determined that the deviation is too large, it sends a calibration request signal to the mobile phone via Bluetooth; if it is determined that the deviation is not too large, the sensor MCU continues to read the sensor and the server respectively.

4. A method for compensating performance deviation of a flexible pressure sensor based on a neural network as claimed in claim 3, characterized in that: In step S6, the compensation algorithm comprehensively considers the ambient temperature and humidity and the historical performance changes of the sensor, and then writes the calculated performance parameters into the MCU of the sensor via Bluetooth.

5. A method for compensating performance deviation of a flexible pressure sensor based on a neural network as claimed in claim 4, characterized in that: The compensation algorithm compensates for the performance drift of the capacitive flexible pressure sensor. The specific compensation algorithm formula is as follows: Where ε0 is the dielectric constant of vacuum, ε r is the relative dielectric constant of the dielectric material, A is the effective area between the conductive layers, ΔA is the offset of the effective area between the conductive layers after the sensor has been used for a period of time, d is the distance between the capacitor plates, Δd is the offset of the distance between the capacitor plates after the sensor has been used for a period of time, and O is the linear effect of temperature and humidity on the performance of the pressure sensor.

Citation Information

Cited By

  • Pressure sensor calibration method based on environmental interference compensation

    CN119245924A

  • Error analysis driven pressure gauge metering calibration method and system

    CN119334534A