High-precision real-time filtering control system

Abstract

The invention discloses a high-precision real-time filtering control system, which comprises an STM32 single chip microcomputer, an FPGA (Field Programmable Gate Array) and a digital-to-analog conversion module, and is characterized in that the STM32 single chip microcomputer is connected with the FPGA and the digital-to-analog conversion module; a band elimination filter transfer function stored in an STM32 single-chip microcomputer is used for representing the frequency response characteristic of a band elimination filter corresponding to a formula, and in the band elimination filter corresponding to the band elimination filter transfer function, the form of the transfer function reflects the suppression or passing of a signal with a specific frequency; high-speed processing is carried out on data by utilizing the FPGA, so that the data transmission capability of the high-precision real-time filtering control system and the received data processing speed are improved; different filters are simulated by using a plurality of formulas, so that the advantages of the plurality of filters are integrated, and the filtering performance of the high-precision real-time filtering control system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a high-precision real-time filtering control system. Background Art

[0002] In signal processing, in order to eliminate noise, a single filter is usually used to filter the signal, limiting the wave transmitted by the signal within a certain range. In this way, the single filter filtering method is not effective when dealing with complex noise environments. For example, a single Kalman filter may result in reduced filtering accuracy when dealing with nonlinear systems and high-noise environments; a single sliding integral filter is prone to produce large delays in rapidly changing signals; a single sliding integral filter is prone to produce large delays in rapidly changing signals; for this purpose, a high-precision real-time filtering control system is provided. Summary of the invention

[0003] To overcome the above-mentioned shortcomings, the object of the present invention is to provide a high-precision real-time filtering control system, by using the band-stop filter transfer function stored in the STM32 single-chip microcomputer to represent the frequency response characteristics of the band-stop filter corresponding to the formula, in the band-stop filter corresponding to the band-stop filter transfer function, the form of the transfer function embodies the suppression or passing of the signal of a specific frequency; by using the state prediction formula and the error covariance prediction formula stored in the STM32 single-chip microcomputer, the predicted value of the noise at different times is evaluated, so as to facilitate the corresponding processing of the noise; by using the Kalman gain, the state update formula and the error covariance update formula, according to the predicted error, real-time correction is performed to reduce the noise; by using the output formula to calculate the best output signal, the high-frequency fluctuation and noise caused by the signal transmission when the data is operated in the motor are effectively reduced; by using FPGA to process the data at high speed, the data transmission capacity of the high-precision real-time filtering control system and the data processing speed received are improved; by using multiple formulas to simulate different filters, the advantages of multiple filters are combined to improve the filtering performance of the high-precision real-time filtering control system.

[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is: a high-precision real-time filtering control system, including: an STM32 single-chip microcomputer, an FPGA, and a digital-to-analog conversion module, wherein the STM32 single-chip microcomputer includes an analog band-stop filter module, and the analog band-stop filter module is provided with a band-stop filter transfer function:

[0005]

[0006] Where: b 0 , b 1 , b 2 is the numerator coefficient of the filter; a0 、a 1 、a 2 is the denominator coefficient of the filter; z -1 Indicates a delay of one sampling period; z -2 Indicates a delay of two sampling periods.

[0007] In this technical solution, the STM32 single-chip microcomputer is used to control the operation of the entire system, and the band-stop filter transfer function stored in the STM32 single-chip microcomputer is used to represent the frequency response characteristics of the band-stop filter corresponding to the formula. In the band-stop filter corresponding to the band-stop filter transfer function, the form of the transfer function reflects the suppression or passage of signals of specific frequencies; FPGA is used to process data at high speed, thereby improving the data transmission capacity and the received data processing speed of the high-precision real-time filtering control system; the digital-to-analog conversion module is used to convert and process the signal, which is beneficial to the processing of the signal by the high-precision real-time filtering control system.

[0008] In some embodiments, the calculation formula of the numerator coefficient is as follows:

[0009] b 0 =1;

[0010] b 1 =-2cos(w 0 );

[0011] b 2 =1;

[0012] The calculation formula of the denominator coefficient is as follows:

[0013] a 0 =1+α;

[0014] a 1 =-2cos(w 0 );

[0015] a 2 =1-α;

[0016] Among them, w 0 is the normalized angular frequency, whose value is α is the broadband parameter of the band-stop filter, and its value is f 0 is the center frequency of the band-stop filter; f s is the sampling frequency of the band-stop filter, and Q is the quality factor of the band-stop filter.

[0017] In this technical solution, the numerator coefficient is used to determine the weighting of the band-stop filter corresponding to the band-stop filter transfer function to the input signal, and the grave coefficient is used to adjust the corresponding characteristics of the band-stop filter corresponding to the band-stop filter transfer function.

[0018] In some embodiments, the STM32 single chip microcomputer further includes an analog Kalman filter module, and the analog Kalman filter module includes a state prediction formula, and the state prediction formula is:

[0019] x k |k-1=Ax k -1|k-1+Buk;

[0020] Among them, k represents the time, x k |k-1 is the state prediction value of the current time k; A is the state transfer matrix; x k -1|k-1 is the state estimate at the previous moment k-1; B is the control input matrix; uk is the control input at the current moment k.

[0021] In this technical solution, the state prediction formula is used to represent the height state data of the current system relative to the zero position obtained by the high-precision real-time filtering control system through the capacitive sensor at this moment, which is beneficial to the processing of signal data by the high-precision real-time filtering control system.

[0022] In some implementations, the analog Kalman filter module further includes an error covariance prediction formula, which is:

[0023] P k∣k-1 =AP k-1∣k-1 AT+Q;

[0024] Among them, P k∣k-1 is the prediction error covariance matrix at the current time k; P k-1∣k-1 is the error covariance matrix of the previous moment k-1; Q is the process noise covariance matrix.

[0025] In this technical solution, the error covariance prediction formula predicts the covariance matrix P at the current moment k∣k-1 , the error range of the system model prediction can be quantified to provide a basis for subsequent filtering updates, thereby achieving optimal processing of noise signals.

[0026] In some implementations, the analog Kalman filter module further includes a Kalman gain calculation formula, which is:

[0027] K k =P k∣k-1 H T (HP k∣k-1 H T +R) -1 ;

[0028] Among them, K k is the Kalman gain matrix; Pk∣k-1 is the prediction error covariance matrix; H is the observation matrix; R is the observation noise covariance matrix.

[0029] In this technical solution, the Kalman gain calculation formula is used to obtain the Kalman gain matrix, thereby adjusting the weight of the state prediction, making the work of the high-precision real-time filtering control system more meticulous, thereby improving the performance of the high-precision real-time filtering control system in filtering noise.

[0030] In some implementations, the analog Kalman filter module further includes a state update formula, which is:

[0031] x k|k =x k|k-1 +K k (z k -Hx k|k-1 );

[0032] Among them, x k|k is the updated state estimate at the current time k; x k|k-1 is the state prediction value of the current analog filter; k is the actual measured value of k at the current moment; z k -Hx k|k-1 is the difference between the actual measurement and the predicted value.

[0033] In this technical solution, the state update formula is used to combine the actual measured value of the noise with the state prediction value obtained by using the state prediction formula, so as to optimize the high-precision real-time filtering control system.

[0034] In some implementations, the analog Kalman filter module further includes an error covariance update formula, and the error covariance update formula is:

[0035] P k∣k =(IK k H)P k∣k-1 ;

[0036] Among them, P k∣k is the updated error covariance matrix at the current time k; I is the identity matrix.

[0037] In this technical solution, the error covariance update formula is used to obtain the difference between the state estimate obtained by the state update formula and the state estimate that does not pass the state update formula.

[0038] In some implementations, the STM32 single chip microcomputer further includes an analog sliding integral filter, and the analog sliding integral filter includes an output formula:

[0039]

[0040] Among them, y[n] is the filter output at time point n; x[ni] is the input signal of N sampling points forward from the current time point n; N is the size of the window.

[0041] In this technical solution, the output formula is used to obtain the filtered output, and the output is smoothed by calculating the average value within the time window of the input signal, thereby outputting the signal to control the motor, effectively reducing the high-frequency fluctuations and noise caused by signal transmission when the data is running in the motor.

[0042] In some implementations, the high-precision real-time filtering control system further includes a capacitive sensor connected to the FPGA.

[0043] In this technical solution, the capacitive sensor is used to collect external signals and transmit the external signals to the FPGA, which is conducive to the high-precision real-time filtering control system to control the motor according to the external signals.

[0044] In some implementations, the high-precision real-time filtering control system further includes a servo driver, and the digital-to-analog conversion module is connected to the servo driver.

[0045] In this technical solution, the servo driver is used to drive the motor. The output signal is converted and transmitted to the servo driver by using the digital-to-analog conversion module, and the servo driver is used to drive the motor.

[0046] The invention has the beneficial effects of representing the frequency response characteristic of the band-stop filter corresponding to the formula by using the band-stop filter transfer function stored in the STM32 single-chip microcomputer, and in the band-stop filter corresponding to the band-stop filter transfer function, the form of the transfer function reflects the suppression or passing of the signal of a specific frequency; by using the state prediction formula and the error covariance prediction formula stored in the STM32 single-chip microcomputer, the predicted value of the noise at different times is evaluated, so as to facilitate the corresponding processing of the noise; by using the Kalman gain, the state update formula and the error covariance update formula, according to the predicted error, real-time correction is performed to reduce the noise; by using the output formula to calculate the best output signal, the high-frequency fluctuation and noise caused by the signal transmission when the data is operated in the motor are effectively reduced; by using FPGA to process the data at high speed, the data transmission capacity of the high-precision real-time filtering control system and the received data processing speed are improved; by using multiple formulas to simulate different filters, the advantages of multiple filters are combined to improve the filtering performance of the high-precision real-time filtering control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a schematic diagram of the overall structural connection of a high-precision real-time filtering control system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more definite definition of the protection scope of the present invention.

[0049] Combined with Figure 1 As shown, the present invention provides a high-precision real-time filtering control system, including: an STM32 single-chip microcomputer, an FPGA (field programmable gate array), and a digital-to-analog conversion module. The STM32 single-chip microcomputer is connected to the FPGA and the digital-to-analog conversion module. In the STM32 single-chip microcomputer, an analog filter can be used to filter an input signal. The STM32 single-chip microcomputer includes an analog band-stop filter module, and the analog band-stop filter module is provided with a band-stop filter transfer function, and the band-stop filter transfer function is:

[0050]

[0051] Where: b 0 , b 1 , b 2 is the numerator coefficient of the filter, which is used to determine the weighting of the input signal by the band-stop filter corresponding to the transfer function of the band-stop filter. 0 、a 1 、a 2 is the denominator coefficient of the filter, which is used to adjust the response characteristics of the band-stop filter corresponding to the transfer function of the band-stop filter. -1 Indicates a delay of one sampling period, z -2 Indicates a delay of two sampling periods.

[0052] The STM32 single-chip microcomputer is used to control the operation of the entire system, and the band-stop filter transfer function stored in the STM32 single-chip microcomputer is used to express the frequency response characteristics of the band-stop filter corresponding to the formula. In the band-stop filter corresponding to the band-stop filter transfer function, the form of the transfer function reflects the suppression or passage of signals of specific frequencies; FPGA is used to process data at high speed, thereby improving the data transmission capacity and the processing speed of received data of the high-precision real-time filtering control system; the digital-to-analog conversion module is used to convert and process the signal, which is beneficial to the processing of the signal by the high-precision real-time filtering control system.

[0053] In some embodiments, the calculation formula of the numerator coefficient is as follows:

[0054] b 0 =1, indicating that the initial gain of the band-stop filter corresponding to the band-stop filter transfer function is 1;

[0055] b1 =-2cos(w 0 );

[0056] b 2 =1, used to match b 0 and b 1 Complete the function of blocking signals of specific frequencies;

[0057] The calculation formula of the denominator coefficient is as follows:

[0058] a 0 =1+α;

[0059] a 1 =-2cos(w 0 ), used to set the frequency response characteristics of the filter;

[0060] a 2 =1-α, used for 0 Cooperate to further adjust the filter response;

[0061] Among them, w 0 is the normalized angular frequency, and its value is It is used to adjust the center frequency of the filter. α is the broadband parameter of the band-stop filter. The wider the bandwidth, the larger the frequency range that the filter can block. Its value is f 0 is the center frequency of the band-stop filter, f s is the sampling frequency of the band-stop filter, and Q is the quality factor of the band-stop filter.

[0062] The numerator coefficient is used to determine the weighting of the input signal by the band-stop filter corresponding to the band-stop filter transfer function, and the grave coefficient is used to adjust the corresponding characteristics of the band-stop filter corresponding to the band-stop filter transfer function.

[0063] In some embodiments, the STM32 single chip microcomputer further includes an analog Kalman filter module, and the analog Kalman filter module includes a state prediction formula, and the state prediction formula is:

[0064] x k |k-1=Ax k -1|k-1+Buk;

[0065] Among them, k represents the time, x k |k-1 is the state prediction value at the current time k; A is the state transfer matrix, which describes how the system state transfers from time k-1 to time k; x k -1|k-1 is the state estimate at the previous moment k-1; B is the control input matrix, which converts the control input uk into the contribution of the state space; uk is the control input at the current moment k.

[0066] The state prediction formula is used to express the height state of the high-precision real-time filtering control system at this moment, which is beneficial to the processing of signal data by the high-precision real-time filtering control system.

[0067] In some implementations, the analog Kalman filter module further includes an error covariance prediction formula, which is:

[0068] P k∣k-1 =AP k-1∣k-1 AT+Q;

[0069] Among them, P k∣k-1 is the prediction error covariance matrix at the current time k; P k-1∣k-1 is the error covariance matrix of the previous moment k-1; Q is the process noise covariance matrix, which represents the uncertainty of the model; A is the state transfer matrix.

[0070] The error covariance prediction formula is used to obtain the noise difference between two moments, which is beneficial for the high-precision real-time filtering control system to filter the noise.

[0071] In some implementations, the analog Kalman filter module further includes a Kalman gain calculation formula, which is:

[0072] K k =P k∣k-1 H T (HP k∣k-1 H T +R) -1 ;

[0073] Among them, K k is the Kalman gain matrix, which is used to balance the weights of the predicted value and the measured value in the state update; k∣k-1 is the prediction error covariance matrix, which indicates the prediction error of the system state at the current moment; H is the observation matrix, which is used to map the system state to the observation space; R is the observation noise covariance matrix, which reflects the uncertainty of the measurement noise, (HP k∣k- 1 H T +R) -1 is the inverse matrix of the measurement uncertainty matrix, which represents the uncertainty after the combination of prediction error and measurement noise. The inverse matrix (HP k∣k-1 H T +R) -1 Used to weigh the importance of predicted values ​​and measured values.

[0074] The Kalman gain matrix represents the weight distribution of the measured value and the predicted value during the update process. Its size depends on the covariance matrix of the measurement error and the prediction error. The Kalman gain calculation formula is used to obtain the Kalman gain matrix, thereby adjusting the weight of the state prediction, making the work of the high-precision real-time filtering control system more meticulous, thereby improving the performance of the high-precision real-time filtering control system in filtering noise.

[0075] In some implementations, the analog Kalman filter module further includes a state update formula, which is:

[0076] x k|k =x k|k-1 +K k (z k -Hx k|k-1 );

[0077] Among them, x k|k : The updated state estimate at the current time k, that is, the optimal estimate at time k; x k|k-1 is the state prediction value of the current analog filter; k : The actual measured value of k at the current moment; z k -Hx k|k-1 : Measurement residuals, that is, the difference between the actual measurement and the predicted value.

[0078] The state update formula is used to combine the actual measured value of the noise with the state prediction value obtained by using the state prediction formula, so as to optimize the high-precision real-time filtering control system.

[0079] In some implementations, the analog Kalman filter module further includes an error covariance update formula, and the error covariance update formula is:

[0080] P k∣k =(IK k H)P k∣k-1 ;

[0081] Among them, P k∣k is the updated error covariance matrix at the current time k; I is the identity matrix.

[0082] The larger the Kalman gain matrix is, the greater the influence of the measurement value on the update. The error covariance update formula is used to obtain the difference between the state estimate obtained by the state update formula and the state estimate that does not pass the state update formula.

[0083] The state prediction formula, error covariance prediction formula, Kalman gain calculation formula, state update formula and error covariance update formula are used to simulate the Kalman filter.

[0084] In some embodiments, the STM32 microcontroller further includes an analog sliding integral filter, and the analog sliding integral filter includes an output formula, and the output formula is:

[0085]

[0086] where y[n] is the filtered output at time point n; x[n - i] is the input signal looking N sampling points forward from the current time point n; N is the size of the window, which determines the smoothness and delay of the filter. The larger the window, the smoother the filtering effect, but the slower the response speed.

[0087] The output formula is used to obtain the filtered output, smooth the output by calculating the average value within the time window of the input signal, and thus output the signal to control the motor, effectively reducing the high-frequency fluctuations and noise caused by signal transmission when the data operates in the motor.

[0088] Continue to combine with the attached Figure 1 As shown, in some embodiments, the high-precision real-time filtering control system further includes a capacitance sensor. The capacitance sensor is connected to the FPGA. The capacitance sensor is used to collect external signals, so as to obtain various data of the external motor during operation, and thus use the external signal as the input signal and transmit the external signal to the FPGA, which is beneficial for the high-precision real-time filtering control system to control the motor according to the external signal.

[0089] Continue to combine with the attached Figure 1 As shown, in some embodiments, the high-precision real-time filtering control system further includes a servo driver. The digital-to-analog conversion module is connected to the servo driver, and the servo driver is connected to the external motor. The servo driver is used to drive the motor. The output signal is converted and transmitted to the servo driver by using the digital-to-analog conversion module, and the servo driver is used to drive the motor.

[0090] The operation process of the present invention is as follows: in the STM32 microcontroller, initialize the formula corresponding to the analog band-stop filter, initialize the formula corresponding to the analog Kalman filter, initialize the formula corresponding to the analog sliding filter, and by using the formula corresponding to the analog band-stop filter, the formula corresponding to the analog Kalman filter, and the formula corresponding to the analog sliding filter, obtain the ideal sampling points. Collect external signals through the FPGA, use the external signals as input signals and input them into the STM32 microcontroller, and use the changes in the input signals to adjust the coefficients of the formula corresponding to the analog band-stop filter, the formula corresponding to the analog Kalman filter, and the formula corresponding to the analog sliding filter in real time, so as to obtain the input signals.

[0091] In summary, the present invention provides a high-precision real-time filtering control system, which is used to represent the frequency response characteristics of the band-stop filter corresponding to the formula by using the band-stop filter transfer function stored in the STM32 single-chip microcomputer, and in the band-stop filter corresponding to the band-stop filter transfer function, the form of the transfer function reflects the suppression or passing of the signal of a specific frequency; by using the state prediction formula and the error covariance prediction formula stored in the STM32 single-chip microcomputer, the predicted value of the noise at different times is evaluated, so as to facilitate the corresponding processing of the noise; by using the Kalman gain, the state update formula and the error covariance update formula, according to the predicted error, real-time correction is performed to reduce the noise; by using the output formula to calculate the best output signal, the high-frequency fluctuation and noise caused by the signal transmission when the data is operated in the motor are effectively reduced; by using FPGA to process the data at high speed, the data transmission capacity of the high-precision real-time filtering control system and the received data processing speed are improved; by using multiple formulas to simulate different filters, the advantages of multiple filters are combined to improve the filtering performance of the high-precision real-time filtering control system.

[0092] The above implementation modes are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with this technology to understand the content of the present invention and implement it, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A high-precision real-time filtering control system, characterized in that: include: STM32 single-chip microcomputer, FPGA, digital-to-analog conversion module, the STM32 single-chip microcomputer is connected to the FPGA and the digital-to-analog conversion module, the STM32 single-chip microcomputer includes an analog band-stop filter module, and the analog band-stop filter module is provided with a band-stop filter transfer function: Among them: b0, b1, b2 are the numerator coefficients of the filter; a0, a1, a2 are the denominator coefficients of the filter; z-1 means delaying one sampling period; z-2 means delaying two sampling periods.

2. The high-precision real-time filtering control system according to claim 1, characterized in that: The calculation formula of the numerator coefficient is as follows: b0＝1； b1=-2cos(w0); b2＝1； The calculation formula of the denominator coefficient is as follows: a0＝1+α； a1=-2cos(w0); a2＝1-α； Among them, w0 is the normalized angular frequency, and its value is α is the broadband parameter of the band-stop filter, and its value is f0 is the center frequency of the band-stop filter; f s is the sampling frequency of the band-stop filter, and Q is the quality factor of the band-stop filter.

3. The high-precision real-time filtering control system according to claim 1, characterized in that: The STM32 single-chip microcomputer further includes an analog Kalman filter module, and the analog Kalman filter module includes a state prediction formula, which is: x k ∣k-1=Ax k -1∣k-1+Buk; Among them, k represents the time, x k |k-1 is the state prediction value of the current time k; A is the state transfer matrix; x k -1|k-1 is the state estimate at the previous moment k-1; B is the control input matrix; uk is the control input at the current moment k.

4. The high-precision real-time filtering control system according to claim 3, characterized in that: The analog Kalman filter module also includes an error covariance prediction formula, which is: P k∣k-1 =AP k-1∣k-1 AT+Q; Among them, P k∣k-1 is the prediction error covariance matrix at the current time k; P k-1∣k-1 is the error covariance matrix of the previous moment k-1; Q is the process noise covariance matrix.

5. The high-precision real-time filtering control system according to claim 3, characterized in that: The analog Kalman filter module also includes a Kalman gain calculation formula, which is: K k =P k∣k-1 H T (HP k∣k-1 H T +R) -1 ; Among them, K k is the Kalman gain matrix; P k∣k-1 is the prediction error covariance matrix; H is the observation matrix; R is the observation noise covariance matrix.

6. The high-precision real-time filtering control system according to claim 5, characterized in that: The analog Kalman filter module also includes a state update formula, which is: x k|k =x k|k-1 +K k (z k -Hx k|k-1 ); Among them, x k|k is the updated state estimate at the current time k; xk|k-1 is the state prediction value of the current analog filter; z k is the actual measured value of k at the current moment; z k -Hx k|k-1 is the difference between the actual measurement and the predicted value.

7. The high-precision real-time filtering control system according to claim 5, characterized in that: The analog Kalman filter module also includes an error covariance update formula, which is: P k∣k =(I-K k H)P k∣k-1 ; Among them, P k∣k is the updated error covariance matrix at the current time k; I is the identity matrix.

8. The high-precision real-time filtering control system according to claim 1, characterized in that: The STM32 single chip microcomputer further includes an analog sliding integral filter, and the analog sliding integral filter includes an output formula, which is: Among them, y[n] is the filter output at time point n; x[ni] is the input signal of N sampling points forward from the current time point n; N is the size of the window.

9. The high-precision real-time filtering control system according to claim 1, characterized in that: A capacitive sensor is also included, and the capacitive sensor is connected to the FPGA.

10. The high-precision real-time filtering control system according to claim 1, characterized in that: It also includes a servo driver, and the digital-to-analog conversion module is connected to the servo driver.

Images