Self-adjusting method for suppressing drift noise of a sensor signal

Abstract

The application provides a self-adjusting sensing signal drift noise suppression method, comprising the following steps: S1, designing a causal low-pass filter with a direct current gain of 1; S2, inputting a sensing signal; S3, low-pass filtering the sensing signal and first-order differentiating the low-pass filtered signal; S4, comparing the absolute value of the first-order differentiated quantity with the resolution of the sensing signal; if the absolute value of the first-order differentiated quantity is greater than or equal to the resolution of the sensing signal, a causal low-pass filter with a lower cut-off frequency is redesigned, and the step S3 is jumped to; if the absolute value of the first-order differentiated quantity is less than the resolution of the sensing signal, the current value of the drift noise is determined according to the differentiated quantity of the sensing system, the current value of the drift noise is subtracted from the sensing signal, and the sensing signal without the drift noise is output. The application does not depend on the input of the system, can adaptively estimate the drift noise regardless of the change of the input quantity, and cannot estimate the drift noise in the process of estimation due to the change of the useful signal.

Description

Technical Field

[0001] This invention relates to the field of sensor application technology, and more specifically, to a self-adjusting method for suppressing sensor signal drift noise. Background Technology

[0002] Due to fluctuations in ambient temperature and power supply, as well as 1 / f noise in devices and systems, drift noise in the sensor output signal is unavoidable. Suppressing drift noise is essential to ensure the stability, reliability, and repeatability of sensing and measurement. Drift noise exhibits significant low-frequency characteristics, and its time scale of change is often larger than that of the useful signal. However, its spectral characteristics, especially its bandwidth, are not deterministic, and the drift trend is time-varying and often unpredictable. Therefore, the time scale of drift noise change is also time-varying and difficult to predict.

[0003] In existing technologies, various frequency-selective filtering methods, based on the difference in time scale between drift noise and useful signal changes, often fail to guarantee consistently good results in suppressing drift noise. Furthermore, they may filter out useful signal components, distorting the useful signal and causing output errors. Existing technologies also employ methods that predict drift trends, establish a drift noise model, and subtract the drift model signal from the output signal to remove drift noise. However, to predict drift trends, it is necessary to acquire output signals with a sufficiently long period of unchanged input, requiring the input to remain unchanged during the drift prediction period. Since many factors contribute to drift, and drift trends are often time-varying, continuous prediction is needed during system operation to ensure the rationality and reliability of drift prediction. However, it is difficult to guarantee a sufficient period of unchanging input during system operation. In other words, in actual system operation, the conditions for predicting drift trends and establishing drift models are often unavailable. Even if a drift model is established, it cannot accurately reflect the real drift after a certain period of time because it cannot be predicted and corrected in a timely manner. Therefore, subtracting the drift model signal from the output signal cannot achieve the effect of removing drift, and may even cause signal distortion, further increasing the noise of the output signal and causing greater errors.

[0004] Patent document CN110031652A (application number: CN201910327397.8) discloses a method for suppressing accelerometer drift and noise. Addressing the problem that the low-frequency performance of accelerometers mounted on inertial stabilized platforms is affected by drift and noise, this invention systematically identifies the acceleration response characteristics of the inertial stabilized platform, then treats the platform's drive input and the system acceleration model as a whole as a virtual accelerometer, and fuses it with the actual accelerometer in the frequency domain. Specifically, the method involves first modeling the system's acceleration model based on physical principles; then using relatively accurate mid-to-high frequency signals from the accelerometer to obtain specific parameters; and finally fusing the low-frequency information from the virtual accelerometer with the mid-to-high frequency information from the actual accelerometer. However, this invention determines whether the acceleration response change is caused by input changes or drift based on the system model, so it is only suitable for sensor systems where models can be built. Different types of sensors require different models, and the performance in identifying and eliminating drift noise depends entirely on the reliability and accuracy of the built model. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this invention is to provide a self-adjusting sensor signal drift noise suppression method that does not require a sensor model, signal model, or drift model.

[0006] A self-adjusting sensor signal drift noise suppression method according to the present invention includes:

[0007] Step S1: Design a causal low-pass filter with a DC gain of 1. ;

[0008] Step S2: Input sensor signal ;

[0009] Step S3: Use the designed causal low-pass filter For sensing signals Perform low-pass filtering and calculate the first-order difference of the output signal of the causal low-pass filter. ;

[0010] Step S4: Compare the absolute values ​​of the first-order difference components and sensor signal resolution size;

[0011] If the absolute value of the first-order difference component Greater than or equal to the sensor signal resolution Redesign the causal low-pass filter with a DC gain of 1 so that its cutoff frequency is lower than that of the causal low-pass filter, and then proceed to step S3.

[0012] If the absolute value of the first-order difference component Smaller than resolution Determine drift noise from the sensor signal. The drift noise is subtracted to output a drift-free sensing signal.

[0013] Preferably, in step S1:

[0014] Design a causal low-pass filter with a DC gain of 1. The cutoff frequency is less than the system bandwidth.

[0015] Preferably, in step S3:

[0016] Using a causal low-pass filter For sensing signals Perform low-pass filtering; the filtered signal is For low-pass filtered signals Perform a first-order difference to obtain the difference components. .

[0017] Preferably, in step S4:

[0018] when ≥0, redesign the causal low-pass filter The cutoff frequency of the new causal low-pass filter is lower than that of the existing causal low-pass filter, and the DC gain of the causal low-pass filter remains at 1. Proceed to step S3.

[0019] Preferably, when <0, calculate the current value of drift noise. Sensor signals The current value minus the drift noise is the current value of the sensor signal after removing the drift.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] 1. This invention does not rely on a sensing system. Regardless of whether the input quantity changes, it can adaptively estimate drift noise. During the estimation process, the drift will not be masked by changes in the useful signal, and there will be no blind spots in the estimation of drift noise.

[0022] 2. This invention can guarantee real-time and accurate estimation of drift noise under various conditions and reliably separate drift noise from the sensing signal. Attached Figure Description

[0023] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0024] Figure 1 This is a schematic diagram of the algorithm flow of the present invention;

[0025] Figure 2This is a schematic diagram of the algorithm flow for Example 2;

[0026] Figure 3 This is a schematic diagram of the algorithm flow in Example 3. Detailed Implementation

[0027] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0028] Example 1:

[0029] This invention proposes an algorithm that automatically and adaptively estimates drift noise in real time based on the drift noise scale, thereby subtracting the accurately estimated drift noise from the output signal to suppress drift.

[0030] According to the present invention, a self-adjusting drift noise suppression method is provided, such as... Figure 1-3 As shown, it includes:

[0031] Step S1: Design a causal low-pass filter;

[0032] Specifically, in step S1:

[0033] Design a causal low-pass filter The cutoff frequency is less than the system bandwidth, and the DC gain is 1.

[0034] Step S2: Acquire sensor signals;

[0035] Step S3: Perform low-pass filtering on the sensor signal, and perform first-order difference on the low-pass filtered sensor signal;

[0036] Specifically, in step S3:

[0037] Use a low-pass filter For signals Perform low-pass filtering; the filtered signal is For low-pass filtered signals Perform a first-order difference to obtain the difference components. .

[0038] Step S4: Determine the absolute value of the first-order difference component and the resolution of the sensing signal;

[0039] If the absolute value of the first-order difference component is greater than or equal to the resolution of the sensing signal, redesign the low-pass filter and proceed to step S3.

[0040] If the absolute value of the first-order difference component is less than the resolution of the sensing signal, the drift noise is determined. Subtracting the drift signal from the sensing signal yields the sensing signal with the drift noise removed.

[0041] Specifically, in step S4:

[0042] when If the value is ≥0, redesign the low-pass filter. The cutoff frequency of the new filter is lower than that of the current filter, and the DC gain is 1. Proceed to step S3.

[0043] Specifically, in step S4:

[0044] when <0, calculate the current value of drift noise. The current value of the sensor signal minus the current value of the drift noise is the sensor signal after removing the drift.

[0045] Because drift noise has the following two characteristics: (1) the gradient of the useful signal is greater than the gradient of the drift noise, and the minimum gradient of the useful signal that can be detected is not less than the resolution of the sensing signal. And (2) the time scale of drift noise change is larger than the time scale of useful signal change, so the present invention proposes the following technical approach:

[0046] 1) Design a causal low-pass filter Its cutoff frequency is less than the system bandwidth, that is, the digital frequency is less than 0.5, and the DC gain is 1.

[0047] 2) Acquisition system output signal ;

[0048] 3) Use a low-pass filter right Perform low-pass filtering; the filtered signal is ;

[0049] 4) Low-pass filtered signal Performing the first-order difference, we get ;

[0050] 5) Judgment If it is less than 0, proceed to step 6); otherwise, redesign the low-pass filter. The cutoff frequency of the new filter is less than the cutoff frequency of the current filter, and the DC gain is still 1. Proceed to step 3) and filter again using the filter.

[0051] 6) Calculate the current value of drift noise. ;

[0052] 7) The current value of the output signal minus the current value of the drift noise equals the output value after drift removal. The algorithm flow based on this technical approach is as follows: Figure 1 .

[0053] Example 2:

[0054] Example 2 is a preferred embodiment of Example 1, and is used to illustrate the present invention in more detail.

[0055] The resolution of the sensing signal to the input quantity is Let the initial timescale of the drift noise analysis be... If at the initial time scale In this process, if the absolute value of the gradient of the sensor signal is greater than the resolution, the time scale is increased by an increment m (m>1), and the gradient of the signal in that time scale is observed. If the absolute value of the gradient of the signal is still greater than the resolution, the increment is continued. Increase the time scale to observe the gradient of the sensor signal until the absolute value of the gradient is less than the resolution. Figure 2 The algorithm for programming the process.

[0056] Example 3:

[0057] Example 3 is a preferred example of Example 1, and is used to illustrate the present invention in more detail.

[0058] A first-order IIR (Infinite Impulse Response) filter is used as the low-pass filter, and the cutoff frequency of the filter is changed by adjusting a single parameter. Let the resolution of the sensing signal be... Design the initial filter , Self-adjusting drift noise suppression. Figure 3 The algorithm for programming the process.

[0059] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0060] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A self-adjusting method for suppressing drift noise of a sensing signal, characterized in that, Comprising: Step S1: Design a causal low-pass filter with a direct current gain of 1 ; Step S2: inputting the sensing signal ; Step S3: Use the designed causal low-pass filter The sensor signal is low-pass filtered, the first difference of the causal low-pass filter output signal is calculated ; Step S4: comparing the absolute values of the first order difference quantities and the resolution of the sensor signal size; If the absolute value of the first-order difference component Greater than or equal to the sensor signal resolution Redesign the causal low-pass filter with a DC gain of 1 so that its cutoff frequency is lower than that of the causal low-pass filter, and then proceed to step S3. If the absolute value of the first-order difference component Smaller than the resolution of the sensing signal Determine drift noise from the sensor signal. The drift noise is subtracted to output a drift-free sensing signal.

2. The self-adjusting, sensed signal drift noise suppression method of claim 1, wherein, In said step S1: Designing causal low-pass filters with dc gain 1 with a cutoff frequency less than the system bandwidth.

3. The self-adjusting, sensed signal drift noise suppression method of claim 1, wherein, In said step S3: With a causal low-pass filter On the sensor signal Low-pass filtered, the filtered signal is ; On the low-pass filtered signal Take the first-order difference, and get the difference .

4. The self-adjusting, sensed signal drift noise suppression method of claim 1, wherein, In said step S4: When ≥ 0, the causal low-pass filter is redesigned , the new causal low-pass filter cutoff frequency is less than the cutoff frequency of the existing causal low-pass filter, the causal low-pass filter DC gain remains 1, and step S3 is entered.

5. The self-adjusting method for suppressing noise of a sensing signal drift according to claim 4, characterized in that: When <0, the current value of the drift noise ; the sensor signal The current value of the sensor signal minus the current value of the drift noise is the drift-removed sensor signal.

Images