Small signal amplification and measurement device based on band-pass filter

Abstract

The invention relates to the technical field of signal processing, and discloses a small signal amplification and measurement device based on a band-pass filter. The multi-stage band-pass filter comprises a multi-stage band-pass filter module, a pre-amplification module, an analog-digital conversion module and the like. The center frequency of the multi-stage band-pass filter is distributed according to logarithm intervals, the bandwidth is dynamically adjusted, the pre-amplification module realizes low-noise amplification, and the analog-to-digital conversion module adopts an oversampling technology and a Sigma-Delta modulator to improve the quantization resolution. The dynamic gain control module adjusts gain in real time, and the noise suppression module eliminates periodic interference. In addition, the device also has the functions of multi-sensor fusion, dynamic range expansion, signal integrity detection, automatic calibration, temperature compensation, parallel processing and the like. The device can accurately amplify and measure small signals and effectively suppress noise, is suitable for the fields of biomedicine, communication, industrial monitoring and the like, and overcomes the defects of a traditional device in the aspect of small signal processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing, in particular to a small signal amplification and measurement device based on a band-pass filter. BACKGROUND

[0002] In the process of modern scientific and technological development, small signal amplification and measurement is a key technical requirement in many fields, covering communication, medical treatment, industrial detection, scientific research and other important fields. However, the existing small signal processing technology has many shortcomings, which limits its further development and application.

[0003] In the field of communication, with the advancement of 5G and even future 6G technology, the signal transmission environment is becoming more and more complex, and the processing of weak signals is becoming more and more difficult. On the one hand, the communication frequency band is constantly expanding, and the signal is easily disturbed by different frequency bands, and the out-of-band noise seriously affects the signal quality. The traditional fixed frequency band-pass filter is difficult to cope with the complex and variable signal environment, and cannot accurately filter out the target frequency band signal, resulting in signal loss or misjudgment. On the other hand, the communication equipment has increasingly stringent requirements on power consumption and size, and the existing small signal amplification circuit is often complex in structure and high in power consumption, which is difficult to meet the development trend of integration and miniaturization. For example, in some portable communication terminals, the limited battery endurance cannot support the long-term operation of high-power signal processing circuits.

[0004] In the field of industrial detection, in some high-precision production process monitoring, it is necessary to accurately measure the electrical signals generated by the tiny physical quantity changes. For example, in the semiconductor manufacturing process, accurate monitoring of signals such as tiny vibration and temperature change of the equipment is crucial to ensure product quality. However, the existing measurement devices have limitations in dynamic range and resolution, making it difficult to simultaneously consider both the accurate measurement of weak signals and the effective monitoring of large signal changes. Moreover, the industrial site environment is harsh, with strong electromagnetic interference, high temperature, humidity and other factors that can adversely affect small signal processing, and the anti-interference ability and environmental adaptability of existing devices need to be improved. SUMMARY

[0005] The purpose of the present application is to provide a small signal amplification and measurement device based on a band-pass filter to solve the problems raised in the background art.

[0006] To achieve the above purpose, the present application provides the following technical scheme: a small signal amplification and measurement device based on a band-pass filter, the device comprising:

[0007] A multi-stage band-pass filter module composed of at least three cascaded band-pass filters, the center frequency of each band-pass filter is distributed in logarithmic intervals, and the bandwidth of each band-pass filter is dynamically adjusted according to a pre-set frequency band division rule;

[0008] A preamplifier module is used to receive the original small signal and perform low-noise amplification, and the gain of the preamplifier module is adjusted by a transconductance amplifier in combination with a feedback resistor network;

[0009] An analog-to-digital conversion module is used to convert the filtered analog signal into a digital signal, and the analog-to-digital conversion module uses an oversampling technique in combination with a Sigma-Delta modulator to achieve high-resolution quantization;

[0010] A dynamic gain control module adjusts the gain coefficient of the post-amplifier in real time based on the signal amplitude detection result, and the gain coefficient is configured by a digital potentiometer in combination with a variable gain amplifier; a noise suppression module uses a time-domain averaging algorithm and a frequency-domain notch filter to jointly process the quantized digital signal, and eliminates periodic interference components.

[0011] Preferably, the center frequency of each bandpass filter in the multi-stage bandpass filter module is determined by the following method:

[0012] A frequency band division model is constructed to divide the target frequency band into a plurality of sub-frequency bands, the number of the sub-frequency bands is equal to the number of the bandpass filters, and the center frequencies of adjacent sub-frequency bands satisfy a logarithmic relationship; the quality factor of each bandpass filter is optimized based on a gradient descent algorithm, so that the amplitude attenuation in the overlapping region of the passbands of adjacent filters does not exceed 3dB; the bandwidth of each bandpass filter is dynamically adjusted by a switched capacitor array, and the configuration parameters of the switched capacitor array are generated by a microcontroller according to the signal-to-noise ratio calculation result of the target frequency band.

[0013] Preferably, the dynamic gain control module includes:

[0014] A signal amplitude detection unit is used to calculate the root mean square value of the input signal in real time, and the root mean square value is obtained by a sliding window integrator in combination with a square root operation; a gain mapping unit converts the detected signal amplitude into a target gain value based on a pre-set gain-amplitude mapping table, and the mapping table is constructed by a piecewise linear interpolation method; a gain adjustment unit uses a digitally controlled variable gain amplifier and a digital-to-analog converter to achieve smooth switching of the gain, and the gain switching rate is limited by a first-order low-pass filter to avoid instantaneous mutation.

[0015] Preferably, the time-domain averaging algorithm of the noise suppression module includes:

[0016] The input signal is divided into a plurality of time windows, and the length of each time window is adaptively adjusted according to the period of the signal fundamental component; the signals in each time window are phase-aligned, the signal delay is calculated by a cross-correlation algorithm, and the phase deviation is compensated; the phase-aligned signals are weighted and averaged, and the weight coefficients are dynamically allocated according to the signal-to-noise ratio of the signals in the window.

[0017] Preferably, the device further comprises a multi-sensor fusion module:

[0018] At least two heterogeneous sensors are configured to collect different physical quantity characteristics of the target signal, including piezoelectric sensors, Hall sensors, and photoelectric sensors; a sensor data synchronization unit is constructed to eliminate transmission delay differences of multiple signals using a timestamp alignment method; a feature-level fusion unit is constructed to extract common features of multiple signals through principal component analysis and to eliminate abnormal data points based on Mahalanobis distance.

[0019] Preferably, the device further comprises a dynamic range expansion module:

[0020] A switchable attenuation network is configured, which is composed of a π-type resistance network controlled by relays, and the attenuation multiple is switched according to the input signal amplitude; an amplitude detection feedback loop is constructed, which triggers the attenuation network to insert and synchronously adjust the gain compensation coefficient of the post-stage amplifier when the input signal amplitude exceeds the preset threshold, and the compensation coefficient is realized by table lookup method for non-linear correction.

[0021] Preferably, the device further comprises a signal integrity detection module:

[0022] A distortion analysis unit is configured to calculate the harmonic distortion and intermodulation distortion of the signal through fast Fourier transform, and the calculation range of the harmonic distortion covers up to the fifth harmonic; a transient response detection unit is constructed to measure the rise time and overshoot of the system using step signal injection method, and a comparator array is used to determine whether the signal waveform distortion exceeds the tolerance threshold.

[0023] Preferably, the device further comprises an automatic calibration module:

[0024] A reference signal source is configured to generate standard sine wave and square wave signals as calibration input; a parameter adjustment unit is constructed to fit the actual frequency response curve of the bandpass filter through least squares method, and to adjust the filter center frequency and bandwidth to the target value; a gain error compensation unit is constructed to generate a compensation coefficient based on the difference between the reference signal amplitude and the measurement result, and to write the compensation coefficient into the non-volatile memory.

[0025] Preferably, the device further comprises a temperature compensation circuit:

[0026] A temperature sensor array is configured to monitor the temperature distribution of key nodes inside the device in real time; a parameter correction model is constructed, which establishes the relationship between temperature and filter center frequency drift through polynomial regression; a digital-to-analog converter is used to dynamically adjust the bias voltage of the bandpass filter to offset the frequency drift caused by temperature.

[0027] Preferably, the device further comprises a parallel processing architecture:

[0028] The multi-core processor is configured, each processor core independently processes a signal of a sub-band; a data distribution unit is constructed, the input signal is divided into multiple sub-bands based on a frequency domain energy detection result and is distributed to the corresponding processor core; and a result synthesis unit is constructed, a weighted superposition algorithm is used to fuse the processing results of the sub-bands, and a weight coefficient is dynamically distributed according to a sub-band signal-to-noise ratio.

[0029] Compared with the prior art, the present application has the following beneficial effects:

[0030] The multi-stage band-pass filter module is composed of at least three cascaded band-pass filters, the center frequencies are distributed in logarithmic intervals, and the bandwidth is dynamically adjusted. By constructing a frequency band division model and optimizing the quality factor, the amplitude attenuation in the overlapping area of adjacent filter passbands is not more than 3dB, the target frequency band can be accurately covered, small signals of different frequencies can be effectively separated, the specificity and accuracy of signal processing are improved, and it is suitable for various complex signal environments. In communication signal processing, weak signals of different frequency bands can be accurately screened out, and signal aliasing and interference can be avoided.

[0031] The preamplifier module uses a transconductance amplifier and a feedback resistor network to jointly adjust the gain and achieve low-noise amplification. While amplifying weak signals, the noise introduced is minimized, and the signal-to-noise ratio is improved. The dynamic gain control module adjusts the gain coefficient of the post-amplifier in real time based on signal amplitude detection, and through the cooperative configuration of a digital potentiometer and a variable gain amplifier, it ensures that the appropriate amplification factor can be achieved under different signal amplitudes, avoids signal distortion, and ensures the stability and reliability of signal processing. In medical electrocardiogram signal monitoring, it can stably amplify weak electrocardiogram signals and improve diagnostic accuracy.

[0032] The analog-to-digital conversion module uses oversampling technology combined with a Sigma-Delta modulator to achieve high-resolution quantization and improve signal quantization accuracy. The noise suppression module uses a time domain averaging algorithm and a frequency domain notch filter to jointly process the quantized digital signal. The time domain averaging algorithm adjusts the time window length, phase alignment and dynamically allocates weight coefficients, and the frequency domain notch filter specifically eliminates interference at specific frequencies, effectively eliminating periodic interference components and further improving signal quality, providing reliable data for subsequent accurate analysis and processing. In industrial field signal measurement, electromagnetic interference can be effectively removed to obtain accurate measurement data.

[0033] The multi-sensor fusion module configures at least two heterogeneous sensors to collect different physical quantity characteristics of the target signal, eliminates transmission delay differences through timestamp alignment, and extracts common features and removes abnormal data points using principal component analysis. It realizes the complementation of multi-source information, improves the comprehensiveness and accuracy of signal measurement, and enhances the perception ability of complex environments and targets. In intelligent security monitoring, combined with multiple sensor information, it can more accurately identify abnormal situations.

[0034] The switchable attenuation network of the dynamic range extension module consists of a relay-controlled π-type resistor network, which switches the attenuation factor in stages according to the input signal amplitude. The amplitude detection feedback loop triggers the attenuation network and adjusts the gain compensation coefficient of the subsequent amplifier when the signal amplitude exceeds a threshold. Nonlinear correction is achieved through a lookup table method, effectively extending the dynamic range of signal measurement and ensuring the device can adapt to the measurement requirements of signals of different intensities, avoiding signal saturation or loss. In power system monitoring, it can accurately measure electrical signals of different amplitudes.

[0035] The signal integrity detection module calculates the harmonic distortion and intermodulation distortion of the signal using Fast Fourier Transform, measures the system's rise time and overshoot, and determines whether the signal waveform distortion exceeds the tolerance threshold. It can monitor signal quality in real time, promptly detect problems in the signal processing process, provide a basis for system optimization and fault diagnosis, and ensure the reliability and stability of the device's output signal. In audio signal processing, it can guarantee audio quality and avoid distortion.

[0036] The automatic calibration module generates a standard signal using a reference signal source, fits the bandpass filter's frequency response curve using the least squares method, adjusts the filter parameters, and generates gain error compensation coefficients based on the amplitude difference of the reference signal, writing them into non-volatile memory to ensure long-term stable operation of the device. The temperature compensation circuit monitors temperature through a temperature sensor array, establishes the relationship between temperature and the filter's center frequency drift, and dynamically adjusts the bandpass filter's bias voltage using a digital-to-analog converter to offset temperature-induced frequency drift, improving the device's adaptability and accuracy under different temperature environments. In demanding fields such as aerospace, this ensures the device operates normally under complex temperature conditions.

[0037] The parallel processing architecture is configured with a multi-core processor, with each processor core independently processing a sub-band signal. The data distribution unit divides the signal into sub-bands based on frequency domain energy detection and allocates them to the corresponding processor cores. The result synthesis unit uses a weighted superposition algorithm to fuse the processing results, with weight coefficients dynamically allocated according to the sub-band signal-to-noise ratio. This effectively improves signal processing speed and efficiency, meeting the needs of applications with high real-time requirements. In high-speed communication data processing, it can quickly process large amounts of data, improving communication efficiency. Attached Figure Description

[0038] Fig. 1 This is a flowchart illustrating the working process of the small-signal amplification and measurement device for the bandpass filter described in this invention.

[0039] Fig. 2 This is a schematic diagram of the working principle of a multi-stage bandpass filter module.

[0040] Fig. 3 This is a schematic diagram of the working principle of a multi-sensor fusion module. Detailed Implementation

[0041] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0042] Please refer to Figs. 1-3 The present application provides a technical solution: a small signal amplification and measurement device based on a band-pass filter, which comprises:

[0043] A multi-stage band-pass filter module: composed of at least three cascaded band-pass filters. The center frequencies of these band-pass filters are distributed in logarithmic intervals, and the bandwidth of each band-pass filter is dynamically adjusted according to a pre-set frequency band division rule. This design can effectively filter small signals of different frequency ranges, improving the specificity and accuracy of signal processing.

[0044] A pre-amplification module: used to receive the original small signal and perform low-noise amplification. Its gain is adjusted by a transconductance amplifier and a feedback resistor network, which can reduce the noise introduced as much as possible while amplifying the signal, ensuring the quality of the signal.

[0045] An analog-to-digital conversion module: converts the filtered analog signal into a digital signal. This module uses oversampling technology combined with a Sigma-Delta modulator to achieve high-resolution quantization, which can improve the quantization accuracy of the signal and provide more accurate data for subsequent digital signal processing.

[0046] A dynamic gain control module: based on the signal amplitude detection result, it adjusts the gain coefficient of the post-amplifier in real time. The gain coefficient is configured by a digital potentiometer and a variable gain amplifier to ensure appropriate amplification under different signal amplitudes and avoid signal distortion.

[0047] A noise suppression module: uses time domain averaging algorithm and frequency domain notch filter to jointly process the quantized digital signal, effectively eliminating periodic interference components and further improving the quality of the signal.

[0048] The present application will be further described below in conjunction with Examples 1 to 5:

[0049] Example 1:

[0050] This embodiment describes in detail the determination method of the center frequency of each band-pass filter in the multi-stage band-pass filter module and the dynamic adjustment method of the bandwidth, to realize accurate filtering of small signals of different frequency bands.

[0051] In practical applications, for multi-stage bandpass filter modules, a frequency band division model is first constructed. For example, assuming the target frequency band is [10Hz-10kHz], if five bandpass filters are required, then the target frequency band is divided into five sub-bands. The center frequency of each sub-band must satisfy the logarithmic relationship between the center frequencies of adjacent sub-bands.

[0052] When determining the center frequency, you can first set the center frequency of the first bandpass filter. The frequency is 10Hz (adjustable according to actual needs). Based on logarithmic relationships, if the logarithmic interval between the center frequencies of adjacent sub-bands is... The center frequency of the second bandpass filter is... The center frequency of the third bandpass filter And so on. The quality factor of each bandpass filter is optimized using the gradient descent algorithm. Quality factor The frequency response of the filter is closely related to optimization. The goal is to ensure that the amplitude attenuation in the passband overlap region of adjacent filters does not exceed 3dB. During the optimization process, adjustments are continuously made... The value is used to calculate the amplitude attenuation in the overlapping area of ​​the passbands of adjacent filters until the requirement of not exceeding 3dB is met.

[0053] The bandwidth of each bandpass filter is dynamically adjusted via a switched capacitor array. The switched capacitor array consists of multiple capacitors and switches; the capacitor values ​​are changed by controlling the opening and closing states of the switches, thereby adjusting the filter's bandwidth. The configuration parameters of the switched capacitor array are generated by a microcontroller based on the signal-to-noise ratio (SNR) of the target frequency band. The microcontroller monitors the SNR of the target frequency band in real time and calculates appropriate capacitor configuration parameters according to a pre-set algorithm, thus achieving dynamic adjustment of the bandpass filter bandwidth.

[0054] Example 2:

[0055] The dynamic gain control module includes a signal amplitude detection unit, a gain mapping unit, and a gain adjustment unit.

[0056] The signal amplitude detection unit is used to calculate the root mean square (RMS) value of the input signal in real time. It obtains the RMS value through a sliding window integrator combined with square root calculation. The sliding window integrator integrates the input signal within a set time window, for example, a window time of... Within this window, the signal Points Then, the square root operation is performed on the integral result to obtain the root mean square value.

[0057] The gain mapping unit converts the detected signal amplitude into a target gain value based on a preset gain-amplitude mapping table. The mapping table is constructed by a piecewise linear interpolation method. For example, according to actual application requirements, the signal amplitude range is divided into multiple intervals, and a linear gain-amplitude relationship is determined in each interval. When the detected signal amplitude is in a certain interval, the linear relationship in the interval is used for interpolation calculation to obtain the corresponding target gain value.

[0058] The gain adjustment unit uses a digitally controlled variable gain amplifier and a digital-to-analog converter to realize smooth switching of the gain. During the switching process, the gain switching rate is limited by a first-order low-pass filter to avoid instantaneous mutation. The transfer function of the first-order low-pass filter can be expressed as where is the cutoff frequency, is the complex frequency variable. By reasonably setting the cutoff frequency , the rate of gain switching can be controlled so that the gain can be smoothly adjusted to the target value, avoiding interference to the signal.

[0059] Embodiment 3:

[0060] This embodiment describes the specific steps of the time domain averaging algorithm in the noise suppression module. By adaptively adjusting the time window length, phase alignment processing, and dynamically allocating weight coefficients, periodic interference components are effectively eliminated.

[0061] In the time domain averaging algorithm of the noise suppression module, the input signal is first divided into multiple time windows. The length of each time window is adaptively adjusted according to the period of the signal fundamental component. For example, by performing spectral analysis on the input signal, the fundamental frequency is determined, and then the time window length can be set to ( is a positive integer, which can be adjusted according to actual conditions), which can ensure that each time window contains a complete signal period, which is beneficial to subsequent processing.

[0062] The signals in each time window are subjected to phase alignment processing. The signal delay is calculated by the cross-correlation algorithm and the phase deviation is compensated. Assuming that there are two signals and in a time window, the cross-correlation function is calculated, and the value of corresponding to the maximum value of the cross-correlation function is found. This value is the delay between the two signals. The signals are phase compensated according to the delay to make the signals in different time windows have consistent phases.

[0063] The phase-aligned signals are weighted and averaged, and the weight coefficients are dynamically assigned according to the signal-to-noise ratio of the signals within the window. The signal-to-noise ratio of the signals within the window is high, and the weight is large. The signal-to-noise ratio of the signals within the window is low, and the weight is small. For example, by calculating the signal-to-noise ratio of each window signal , the weight coefficient can be set as ( is the total number of time windows), so that the contribution of high-quality signals can be highlighted and the influence of noise can be suppressed.

[0064] Embodiment 4:

[0065] This embodiment details the configuration, data synchronization and feature-level fusion method of heterogeneous sensors in the multi-sensor fusion module, and explains the working principle of the switchable attenuation network and amplitude detection feedback loop in the dynamic range expansion module, improving the accuracy and adaptability of signal measurement.

[0066] ① Multi-sensor fusion module:

[0067] At least two heterogeneous sensors are configured, such as piezoelectric sensors, Hall sensors and photoelectric sensors. Piezoelectric sensors can be used to detect physical quantities such as pressure and vibration, Hall sensors can be used to detect magnetic field changes, and photoelectric sensors can be used to detect light signals. These sensors collect different physical characteristics of the target signal.

[0068] A sensor data synchronization unit is constructed, and a timestamp alignment method is used to eliminate the transmission delay difference of multiple signals. When each sensor collects data, a timestamp is added to the data. For example, when the piezoelectric sensor collects data, the time is recorded as the timestamp; when the Hall sensor collects data, the time is recorded as the timestamp. By comparing the timestamps, the data of different sensors are time-aligned to ensure that the time reference of multiple signals is consistent in subsequent processing.

[0069] A feature-level fusion unit is constructed, and the common features of multiple signals are extracted by principal component analysis (PCA) method, and the abnormal data points are removed based on Mahalanobis distance. Principal component analysis is a common dimensionality reduction technique that can convert multiple related features into a few independent principal components. For example, for the pressure data collected by the piezoelectric sensor, the magnetic field data collected by the Hall sensor and the light intensity data collected by the photoelectric sensor, the principal components , and so on. Then, the Mahalanobis distance is used to calculate the distance between each data point and the center of the data distribution. If the Mahalanobis distance of a certain data point exceeds a set threshold, the data point is determined to be an abnormal point and is removed.

[0070] ② Dynamic range expansion module:

[0071] A switchable attenuation network is configured, which is composed of a relay-controlled π-type resistance network. The π-type resistance network is composed of three resistances. By controlling the connection and disconnection of the resistances through the relay, different attenuation multiples are realized. The attenuation multiple is switched according to the input signal amplitude. For example, when the input signal amplitude is small, the relay does not act, and the attenuation network is not connected; when the input signal amplitude exceeds a certain threshold, the relay acts to connect a specific resistance to the π-type network, realizing a certain multiple of attenuation, such as attenuation by 2 times, 4 times, etc.

[0072] A magnitude detection feedback loop is constructed. When the input signal amplitude exceeds a preset threshold, the attenuation network is inserted and the gain compensation coefficient of the subsequent amplifier is adjusted synchronously. The compensation coefficient is realized by table lookup method for non-linear correction. In practical application, a gain compensation coefficient table is measured and prepared in advance, which records the gain compensation coefficients corresponding to different input signal amplitudes. When the attenuation network is inserted, the gain compensation coefficient corresponding to the input signal amplitude is looked up in the table, and the gain of the subsequent amplifier is adjusted to ensure accurate measurement of the signal in the entire dynamic range.

[0073] Embodiment 5:

[0074] This embodiment describes in detail the methods of distortion analysis and transient response detection in the signal integrity detection module, the working principles of the reference signal source, parameter adjustment and gain error compensation in the automatic calibration module, the implementation methods of temperature monitoring and frequency drift compensation in the temperature compensation circuit, and the specific operation mechanisms of the multi-core processor, data distribution and result synthesis in the parallel processing architecture, ensuring the stable and reliable performance of the device.

[0075] ① Signal integrity detection module:

[0076] A distortion analysis unit is configured to calculate the harmonic distortion and intermodulation distortion of the signal by fast Fourier transform (FFT). Fast Fourier transform can convert time domain signals into frequency domain signals, which is convenient for analyzing the frequency components of the signal. When calculating the harmonic distortion, the calculation range covers the fifth harmonic. For example, for an input signal , its frequency spectrum is obtained by FFT, and then the ratio of the amplitude of each harmonic (such as second harmonic, third harmonic, fourth harmonic, and fifth harmonic) to the fundamental amplitude is calculated to obtain the harmonic distortion.

[0077] A transient response detection unit is constructed to measure the rise time and overshoot of the system using the step signal injection method. A step signal is injected into the system , and the response of the system output signal is recorded. The rise time is the time required for the output signal to rise from 10% to 90% of the steady-state value, and the overshoot is the ratio of the maximum amplitude of the output signal exceeding the steady-state value to the steady-state value. The comparator array determines whether the signal waveform distortion exceeds the tolerance threshold. The output signal is compared with the preset tolerance threshold, and if it exceeds the threshold, it is determined that the signal waveform has been distorted.

[0078] ② Automatic calibration module:

[0079] A reference signal source is configured to generate standard sine and square wave signals as calibration inputs. The standard sine and square wave signals have clear frequency and amplitude characteristics and can be used to calibrate the performance of the device. For example, a standard sine wave with a frequency of 1 kHz and an amplitude of 1 V and a square wave signal with a peak-to-peak value of 2 V are generated.

[0080] A parameter adjustment unit is constructed to fit the actual frequency response curve of the bandpass filter by the least squares method, and adjust the filter center frequency and bandwidth to the target value. The least squares method is a commonly used curve fitting method that minimizes the sum of the squares of the errors between the actual measurement data and the theoretical model to obtain the best fitting parameters. In this embodiment, the signal output by the reference signal source is used to measure the actual frequency response of the bandpass filter, and then the adjustment value of the filter center frequency and bandwidth is calculated by the least squares method to adjust it to the target value.

[0081] A gain error compensation unit is constructed to generate a compensation coefficient based on the difference between the reference signal amplitude and the measurement result, and write the compensation coefficient into the non-volatile memory. For example, the reference signal amplitude is , the measured signal amplitude is , and the gain error is . The compensation coefficient is calculated according to the gain error, and then written into the non-volatile memory. In subsequent measurements, the measurement result is corrected according to the compensation coefficient.

[0082] ③ Temperature compensation circuit:

[0083] A temperature sensor array is configured to monitor the temperature distribution of key nodes inside the device in real time. The temperature sensor array can be composed of multiple temperature sensors distributed near key components such as bandpass filters and amplifiers to collect temperature data in real time.

[0084] A parameter correction model is constructed to establish the relationship between temperature and filter center frequency drift by polynomial regression. For example, assuming the temperature is , the filter center frequency drift is By collecting a large amount of temperature and frequency drift data, a polynomial regression is used to obtain the frequency drift amount is a regression coefficient.

[0085] The bias voltage of the band-pass filter is dynamically adjusted by the digital-to-analog converter to offset the temperature-induced frequency drift. According to the frequency drift amount calculated by the parameter correction model, the digital-to-analog converter converts it into a corresponding voltage value, and adjusts the bias voltage of the band-pass filter, so that the center frequency of the filter remains stable.

[0086] 4. Parallel processing architecture:

[0087] Configure a multi-core processor, and each processor core independently processes the signal of a sub-band. For example, if the target frequency band is divided into 4 sub-bands, a 4-core processor is used, and each core processes the signal of a sub-band, improving processing efficiency.

[0088] A data distribution unit is constructed, and the input signal is divided into multiple sub-bands based on the frequency energy detection result and distributed to the corresponding processor core. By analyzing the frequency energy of the input signal, the energy distribution of different frequency bands is determined, and then the signal is divided into multiple sub-bands according to the energy distribution. For example, the frequency band with higher energy is divided into one sub-band, and the frequency band with lower energy is divided into another sub-band, and then these sub-bands are distributed to the corresponding processor core for processing.

[0089] A result synthesis unit is constructed, and a weighted superposition algorithm is used to fuse the processing results of each sub-band. The weight coefficient is dynamically distributed according to the signal-to-noise ratio of the sub-band. The weight coefficient is the total number of sub-bands), and the processing results of each sub-band are superimposed according to the weight coefficient to obtain the final processing result.

[0090] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0091] ​​​​​​While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A small signal amplification and measurement device based on a bandpass filter, characterized by, Comprise: Multi-stage band-pass filter module, composed of at least three cascaded band-pass filters, the center frequency of each band-pass filter is distributed in logarithmic intervals, and the bandwidth of each band-pass filter is dynamically adjusted according to the preset frequency band division rule; Pre-amplification module for receiving original small signal and performing low-noise amplification, the gain of the pre-amplification module is adjusted by a transconductance amplifier and a feedback resistor network; Analog-to-digital conversion module for converting the filtered analog signal into a digital signal, the analog-to-digital conversion module uses oversampling technology combined with Sigma-Delta modulator to realize high-resolution quantization; Dynamic gain control module, based on the signal amplitude detection result, the gain coefficient of the post-amplifier is adjusted in real time, and the gain coefficient is configured by a digital potentiometer and a variable gain amplifier; Noise suppression module, using time domain average algorithm and frequency domain notch filter to jointly process the quantized digital signal, and eliminating periodic interference components.

2. The apparatus of claim 1, wherein, The center frequency of each band-pass filter in the multi-stage band-pass filter module is determined by the following method: Construct a frequency band division model to divide the target frequency band into several sub-bands, the number of sub-bands is equal to the number of band-pass filters, and the center frequencies of adjacent sub-bands satisfy the logarithmic relationship; based on the gradient descent algorithm, the quality factor of each band-pass filter is optimized, so that the amplitude attenuation of the overlapping region of adjacent filters does not exceed 3dB; the bandwidth of each band-pass filter is dynamically adjusted by a switch capacitor array, and the configuration parameters of the switch capacitor array are generated by a microcontroller according to the signal-to-noise ratio calculation result of the target frequency band.

3. The apparatus of claim 1, wherein, The dynamic gain control module comprises: Signal amplitude detection unit for real-time calculation of the root mean square value of the input signal, the root mean square value is obtained by a sliding window integrator combined with a square root operation; gain mapping unit, based on the preset gain-amplitude mapping table, converts the detected signal amplitude into a target gain value, the mapping table is constructed by piecewise linear interpolation method; gain adjustment unit, using digitally controlled variable gain amplifier and digital-to-analog converter to realize smooth switching of gain, wherein the gain switching rate is limited by a first-order low-pass filter to avoid instantaneous mutation.

4. The apparatus of claim 1, wherein, The time domain average algorithm of the noise suppression module includes: Divide the input signal into multiple time windows, the length of each time window is adaptively adjusted according to the period of the signal fundamental component; phase alignment processing is performed on the signal in each time window, the signal delay is calculated by cross-correlation algorithm and the phase deviation is compensated; weighted average is performed on the phase-aligned signal, and the weight coefficient is dynamically allocated according to the signal-to-noise ratio of the signal in the window.

5. The apparatus of claim 1, wherein, Also includes a multi-sensor fusion module: Configure at least two heterogeneous sensors to collect different physical quantity characteristics of the target signal, the heterogeneous sensors include piezoelectric sensors, Hall sensors and photoelectric sensors; construct a sensor data synchronization unit to eliminate the transmission delay difference of multiple signals using timestamp alignment method; Construct a feature-level fusion unit to extract common features of multiple signals by principal component analysis method, and remove abnormal data points based on Mahalanobis distance.

6. The apparatus of claim 1, wherein, Also includes a dynamic range expansion module: A switchable attenuation network is configured, which is composed of a π-type resistor network controlled by a relay, and the attenuation multiple is switched according to the amplitude of the input signal; an amplitude detection feedback loop is constructed, when the amplitude of the input signal exceeds a preset threshold, the attenuation network is triggered to be inserted and the gain compensation coefficient of the subsequent amplifier is adjusted synchronously, and the compensation coefficient is realized by a table lookup method to realize nonlinear correction.

7. The apparatus of claim 1, wherein, A signal integrity detection module is also included: A distortion analysis unit is configured to calculate the harmonic distortion and intermodulation distortion of the signal by fast Fourier transform, and the calculation range of the harmonic distortion covers the fifth harmonic; a transient response detection unit is constructed, a step signal injection method is used to measure the rise time and overshoot of the system, and a comparator array is used to judge whether the signal waveform distortion exceeds the tolerance threshold.

8. The apparatus of claim 1, wherein, An automatic calibration module is also included: A reference signal source is configured to generate standard sine wave and square wave signals as calibration input; a parameter adjustment unit is constructed to fit the actual frequency response curve of the band-pass filter by the least square method, and the center frequency and bandwidth of the filter are adjusted to the target value; A gain error compensation unit is constructed, a compensation coefficient is generated based on the difference between the reference signal amplitude and the measurement result, and the compensation coefficient is written into the non-volatile memory.

9. The apparatus of claim 1, wherein, A temperature compensation circuit is also included: A temperature sensor array is configured to monitor the temperature distribution of the key nodes inside the device in real time; a parameter correction model is constructed, which establishes the relationship between temperature and filter center frequency drift by polynomial regression; a digital-to-analog converter is used to dynamically adjust the bias voltage of the band-pass filter to offset the frequency drift caused by temperature.

10. The apparatus of claim 1, wherein, A parallel processing architecture is also included: A multi-core processor is configured, each processor core independently processes the signal of a sub-band; a data distribution unit is constructed, based on the frequency energy detection result, the input signal is divided into multiple sub-bands and distributed to the corresponding processor core; a result synthesis unit is constructed, a weighted superposition algorithm is used to fuse the processing results of each sub-band, and the weight coefficient is dynamically allocated according to the signal-to-noise ratio of the sub-band.

Images