Electronic cigarette microphone sensitivity self-adaptive calibration method

Abstract

The invention provides an electronic cigarette microphone sensitivity self-adaptive calibration method, which relates to the technical field of electronic cigarette intelligent control, and is characterized in that a main suction pressure signal and a bypass airflow signal are respectively acquired by arranging double sensing channels, and high-frequency oscillation airflow is injected based on a double-source excitation principle; and dynamically monitoring the high-frequency component attenuation degree of the main suction pressure signal, and generating a sensitivity drift trend. Thus, the accuracy and the real-time performance of sensitivity calibration can be improved, the method adapts to complex airflow environments and temperature and humidity changes, self-adaptive matching of the microphone sensitivity and the atomization effect is achieved, and therefore the airflow detection accuracy and the user experience of the electronic cigarette are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for electronic cigarettes, and more specifically, to an adaptive calibration method for the sensitivity of an electronic cigarette microphone. Background Technology

[0002] As a novel tobacco alternative, e-cigarettes have received widespread attention and rapid development in recent years. E-cigarettes heat an aerosol to generate an aerosol for users to inhale, simulating the smoking experience of traditional cigarettes. To enhance the user experience, e-cigarettes typically incorporate multiple sensors, with the microphone sensor being a key component. This sensor primarily detects the user's inhalation and triggers the atomizer. The sensitivity of the microphone in existing e-cigarettes is a crucial factor affecting airflow detection accuracy, atomizer response speed, and overall performance. However, due to variations in temperature and humidity in the e-cigarette's operating environment, complex airflow disturbances, and sensor aging or drift, sensitivity is prone to deviation, leading to decreased inhalation detection accuracy and uneven atomization. In recent years, the industry has conducted numerous studies on sensitivity calibration techniques for e-cigarette sensors, such as calibrating sensitivity using a single airflow sampling signal or periodically maintaining it with external calibration devices. However, these methods suffer from drawbacks such as poor real-time performance, weak adaptability, and complex structures, failing to meet the miniaturization and intelligent development trends of e-cigarettes.

[0003] The shortcomings of existing technologies in sensor sensitivity calibration are mainly reflected in the following aspects: First, traditional single-channel airflow detection schemes rely solely on the main inhalation pressure signal, making it difficult to comprehensively capture multi-dimensional information in complex airflow environments, resulting in poor accuracy and robustness of sensitivity calibration. Second, existing calibration techniques are mostly static calibrations, lacking dynamic response capabilities and struggling to adapt to rapid changes in airflow signals during inhalation. Third, they ignore the nonlinear effects of environmental factors (such as temperature and humidity) on sensor performance, failing to provide an effective sensitivity compensation mechanism. In summary, the shortcomings of existing technologies severely limit the improvement of the e-cigarette experience, especially in terms of human-computer interaction, energy consumption optimization, and reliability assurance. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides an adaptive calibration method for the sensitivity of an electronic cigarette microphone.

[0005] According to one aspect of the present invention, an adaptive calibration method for the sensitivity of an electronic cigarette microphone is provided, comprising: A dual-sensor channel is set in the electronic cigarette microphone. The first channel collects the main inhalation pressure signal and the second channel collects the bypass airflow signal. The ratio of the main inhalation pressure signal to the bypass airflow signal is used as the original feature ratio. Based on the dual-source excitation principle, a weak high-frequency oscillating airflow is injected into the microphone air intake end, and the attenuation degree of the high-frequency component in the main inhalation pressure signal is detected to generate a dynamic attenuation coefficient. The drift trend of sensor sensitivity is calculated based on the cross-correlation degree between the original characteristic ratio and the dynamic attenuation coefficient, and a compensation mapping curve is established based on the Pearson correlation coefficient. The compensation mapping curve adjusts non-linearly with increasing temperature. The controller adjusts the gain parameters of the sensor signal conditioning circuit according to the compensation mapping curve, and adjusts the power output curve of the atomizer in conjunction with the adjustment, so as to achieve adaptive matching between microphone sensitivity and atomization effect.

[0006] Furthermore, a main airway and a bypass airway are provided at the air intake end of the electronic cigarette microphone; The main airway is connected to the atomizing chamber, and the bypass airway is arranged in parallel with the main airway and separated from the main airway by an isolation structure. A first pressure sensor is installed in the main airway; a second pressure sensor is installed in the bypass airway; the first pressure sensor and the second pressure sensor are on the same cross-section and are both fixed to the sensor mounting base by elastic sealing rings; the sensor mounting base is detachably connected to the electronic cigarette shell by a threaded structure.

[0007] Furthermore, the first pressure sensor acquires a main inhalation pressure signal mixed with a high-frequency oscillation component, and the main inhalation pressure signal is processed by a signal conditioning circuit to obtain a pure high-frequency component A1; After the high-frequency component A1 passes through the rectifier circuit and the RC integrator circuit, the effective value A2 reflecting the high-frequency energy is obtained; The dynamic attenuation coefficient K is obtained by calculating the ratio of the effective value A2 to the original amplitude value A0 of the high-frequency oscillating airflow.

[0008] Furthermore, the drift trend of sensor sensitivity is calculated by performing an FFT transform on the complex product of the original characteristic ratio R and the dynamic attenuation coefficient K sequence, followed by an inverse Fourier transform to obtain the time-domain cross-correlation function, and then using a sliding window to calculate the dynamic change of the Pearson correlation coefficient r to reflect the drift trend of sensor sensitivity.

[0009] Furthermore, the time-domain cross-correlation function is obtained by multiplying the FFT result FR of the two sequences with the conjugate complex number of FK to obtain the frequency-domain cross-correlation function S(f), which is then frequency normalized and subjected to a Hanning window before IFFT operation to obtain the time-domain cross-correlation function.

[0010] Furthermore, the time-domain cross-correlation function is shown in the following equation: in, It is the frequency domain cross-correlation function. The maximum magnitude of S(f) is used for normalization. For high-frequency component attenuation function, These are weighting coefficients. The number of sampling points. For time variables, For frequency variables, The imaginary unit, Pi This represents the sequence number of the discrete sampling point.

[0011] Furthermore, the Pearson correlation coefficient is calculated based on the time-domain cross-correlation function by calculating the mean and standard deviation of the original feature ratio R sequence and the dynamic decay coefficient K sequence respectively, and then calculating the covariance and dividing the covariance by the product of the two standard deviations to obtain the Pearson correlation coefficient.

[0012] Furthermore, based on the domain of the Pearson correlation coefficient, sub-intervals are divided and control points are selected. Then, cubic spline interpolation is performed on each interval and the intervals are spliced ​​to obtain the nonlinear compensation mapping curve.

[0013] Furthermore, the gain parameter of the sensor signal conditioning circuit is adjusted according to the compensation mapping curve. The corresponding gain adjustment value is found from the compensation mapping curve based on the ratio of the sensor output signal to the reference level, converted into a digital control word of the PGA, and then the gain parameter is adjusted in real time through the SPI interface.

[0014] Furthermore, the adaptive matching calculates the reference power based on the gain parameters by the controller, determines the target power by combining the adjustment coefficient with the microphone sound pressure signal, and controls the atomizer output power in real time through PWM duty cycle modulation and PI algorithm, and dynamically establishes the mapping relationship between microphone sensitivity and atomization power.

[0015] Compared with existing technologies, the adaptive calibration method for electronic cigarette microphone sensitivity provided by this invention uses dual sensing channels to collect the main inhalation pressure signal and the bypass airflow signal respectively. Based on the dual-source excitation principle, a high-frequency oscillating airflow is injected to dynamically monitor the attenuation of the high-frequency component of the main inhalation pressure signal, generating a sensitivity drift trend. This improves the accuracy and real-time performance of sensitivity calibration, adapts to complex airflow environments and temperature and humidity changes, and achieves adaptive matching between microphone sensitivity and atomization effect, thereby significantly improving the airflow detection accuracy and user experience of electronic cigarettes. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a system block diagram of an adaptive calibration method for electronic cigarette microphone sensitivity according to an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of the compensation mapping curve in the adaptive calibration method for electronic cigarette microphone sensitivity according to an embodiment of the present invention. Detailed Implementation

[0018] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0019] Figure 1 This is a system block diagram of an adaptive calibration method for electronic cigarette microphone sensitivity according to an embodiment of the present invention. Figure 1 As shown, the adaptive calibration method for electronic cigarette microphone sensitivity includes: S1: A dual-sensor channel is set in the electronic cigarette microphone. The first channel collects the main inhalation pressure signal, and the second channel collects the bypass airflow signal. The ratio of the main inhalation pressure signal to the bypass airflow signal is used as the original feature ratio. The bypass airflow signal is used to eliminate interference from environmental factors. A main airway and a bypass airway are provided at the air intake end of the electronic cigarette microphone. The main airway is connected to the atomizing chamber, and the bypass airway is arranged in parallel with the main airway and separated from the main airway by an isolation structure. A first pressure sensor is installed in the main airway. The first pressure sensor is a MEMS pressure sensor with a sensitivity of not less than 0.1 kPa and a response time of less than 5 ms. A second pressure sensor is installed in the bypass airway. The second pressure sensor is of the same model as the first pressure sensor and is installed on the same cross-section as the first pressure sensor. The cross-sectional area of ​​the main airway is 5-8 times that of the bypass airway to ensure sufficient airflow to the main airway. The main inhalation pressure signal collected by the first pressure sensor is denoted as P1. The bypass airflow signal collected by the second pressure sensor is denoted as P2, and the ratio of P1 to P2 is denoted as R. The R value is used as the original characteristic ratio to characterize the user's actual inhalation intensity. Since the first and second pressure sensors are affected by the same ambient temperature and humidity, when environmental factors cause sensor performance drift, the interference caused by environmental factors can be effectively eliminated by calculating the ratio R. The isolation structure is made of porous ceramic material, which has good heat insulation performance to ensure that there is no temperature interference between the two sensing channels. Both the first and second pressure sensors are fixed to the sensor mounting base by elastic sealing rings. The sensor mounting base is detachably connected to the electronic cigarette shell by a threaded structure, which facilitates sensor replacement and maintenance.

[0020] S2: Based on the dual-source excitation principle, a weak high-frequency oscillating airflow is injected into the microphone air inlet, the attenuation degree of the high-frequency component in the main inhalation pressure signal is detected, and a dynamic attenuation coefficient is generated. The dynamic attenuation coefficient reflects the real-time impedance characteristics of the microphone cavity. When the piezoelectric ceramic resonator at the microphone's air intake is working, its controller outputs a 3.3kHz drive signal. Under a 30% duty cycle, the resonator generates a reciprocating vibration with a peak intensity of 0.2mm, thus forming a weak high-frequency oscillating airflow at the air intake. The original amplitude value A0 of this high-frequency oscillating airflow is acquired and stored by a reference sensor located at the air intake. When the high-frequency oscillating airflow propagates within the microphone cavity, the first pressure sensor acquires a main inhalation pressure signal mixed with a high-frequency oscillation component. This main inhalation pressure signal passes through a high-pass filter with a cutoff frequency of 2kHz in the signal conditioning circuit, filtering out low-frequency inhalation signals to obtain a pure high-frequency component A1. The high-frequency component A1 undergoes dynamic gain adjustment in the signal conditioning circuit. During adjustment, the gain coefficient is inversely proportional to the amplitude of A1, maintaining the signal strength at an appropriate level. After the high-frequency component A1 passes through a rectifier circuit and an RC integrator circuit, an effective high-frequency energy is obtained. The effective value A2 is used as the reference value. At the end of each sampling period, the controller calculates the dynamic attenuation coefficient K by the ratio of the effective value A2 to the original amplitude value A0, where K = A2 / A0. The dynamic attenuation coefficient K is calculated using a moving average method, with each calculation based on the data of the first 50 sampling points. The influence of instantaneous fluctuations is eliminated through smoothing. After the new dynamic attenuation coefficient K is calculated, the controller performs curve fitting with historical data to generate a trend curve of K value change. The change state of the microphone characteristics is judged based on the slope and curvature characteristics of the curve. During the calculation of the dynamic attenuation coefficient K, the controller eliminates the influence of ambient temperature changes and signal intensity fluctuations through real-time temperature compensation and signal amplitude normalization, ensuring that the K value can accurately reflect the changes in the acoustic impedance characteristics of the microphone cavity. During continuous sampling, the controller writes the filtered K value into the historical data buffer and predicts the dynamic change trend of microphone performance by analyzing the change pattern of K value.

[0021] The specific implementation of dynamic gain adjustment in the signal conditioning circuit is as follows: The dynamic gain adjustment circuit employs a programmable gain amplifier (PGA). The PPA's gain range covers multiple orders of magnitude and can be continuously adjusted via a digital interface. When the high-frequency component A1 enters the PPA, the controller first reads the peak-to-peak value of A1 and converts it into logarithmic form to obtain the logarithmic amplitude P. The gain G of the PPA is calculated according to the formula G=k / P, where k is a calibration coefficient used to adjust the signal to a suitable range. The calculation of gain G uses a fixed-point arithmetic method. The controller updates the gain value at fixed time intervals, and the gain change is gradual. The change in gain between adjacent intervals is proportionally limited to ensure a smooth transition and avoid abrupt gain changes. The signal distortion caused by this is addressed in the signal conditioning process. During this process, a feedback detection circuit is installed at the output of the programmable gain amplifier. This circuit includes a peak detector and an operational amplifier, used to monitor the amplitude characteristics of the amplified signal in real time. When a change in the amplitude characteristics of the output signal is detected, the controller adjusts the k value according to the trend, ensuring the output signal remains within the linear amplification range. During dynamic gain adjustment, the controller obtains the gain control word using a lookup table and writes it into the control register of the programmable gain amplifier via a digital communication interface, achieving real-time gain adjustment. Simultaneously with gain adjustment, the controller records the gain value for each adjustment and uses it as a calibration coefficient for subsequent signal amplitude restoration calculations.

[0022] The specific methods for limiting the gradual transition during the dynamic gain adjustment process are as follows: The gradual transition of the gain G employs a slope-limited algorithm. The controller calculates the difference between the target gain value at the current moment and the actual gain value at the previous moment to obtain the gain change ΔG. When the gain change ΔG is positive, it indicates that the gain needs to be increased, and the controller uses the product of ΔG and the actual gain value at the previous moment as the dynamic limiting reference value M1. When the gain change ΔG is negative, it indicates that the gain needs to be decreased, and the controller uses the product of ΔG and the actual gain value at the previous moment as the dynamic limiting reference value M2. The controller calculates the maximum allowable step size based on the dynamic limiting reference value. If the current gain change requirement exceeds the maximum... The step size divides the gain change into multiple consecutive adjustment cycles, with the gain change amount in each cycle determined by the dynamic limiting reference value. Within each consecutive adjustment cycle, the controller employs a weighted smoothing algorithm to perform a weighted average of the gain values ​​of adjacent cycles, with the weighting coefficient decreasing as the time interval between adjacent cycles increases. At the end of each adjustment cycle, the controller records the actual gain change trajectory in the historical data area for optimizing subsequent gradual adjustment strategies. By combining dynamic limiting and weighted smoothing, a smooth transition in gain adjustment is achieved, avoiding signal distortion caused by sudden gain changes and improving the dynamic response quality of the system.

[0023] It should be noted that the weak high-frequency oscillating airflow generated by the piezoelectric ceramic oscillator has the following mechanism of action: When the oscillating airflow propagates in the microphone cavity at a frequency of 3.3 kHz, this frequency is much higher than the user's normal inhalation frequency, allowing the oscillating airflow to propagate in the microphone cavity as an independent acoustic detection signal; during the propagation process, the oscillating airflow interacts acoustically with the microphone inner wall, atomizing core, and airflow channel, where the high-frequency oscillation energy is partially absorbed and attenuated by the microphone structure. When subtle changes occur in the internal structure of the microphone, such as changes in the wetting state of the atomizing core, or the appearance of condensation or foreign objects in the airway, the propagation characteristics of the oscillating airflow will change accordingly; due to the different propagation speed and attenuation characteristics of sound waves in different media, when liquid accumulates inside the microphone or the atomizing liquid level changes, the energy loss of the oscillating airflow during propagation will change accordingly, and this change is directly reflected in the value of the dynamic attenuation coefficient K; the detection process of the oscillating airflow is similar to that of medical ultrasound. The principle of acoustic imaging is to analyze the attenuation characteristics of high-frequency signals during propagation to obtain real-time information on changes in the internal state of the microphone. When a user inhales, the oscillating airflow and the main inhalation airflow propagate together. Due to the significant difference in frequency characteristics, the two airflows can be completely separated using frequency domain analysis methods, without interfering with each other. The amplitude of the oscillating airflow is set on the order of 0.2 mm, which is sufficient to generate a detectable acoustic signal while being small enough not to affect the user's smoking experience or significantly impact the atomization process. During the microphone's lifespan, continuous monitoring of the propagation characteristics of the oscillating airflow can detect early signs of microphone performance degradation in a timely manner, providing a basis for preventative maintenance. When abnormal fluid accumulation or blockage occurs inside the microphone, the propagation path of the oscillating airflow changes, leading to a significant change in the attenuation mode of acoustic energy. This change can be quickly captured by the dynamic attenuation coefficient K, thereby enabling timely detection of abnormal microphone conditions.

[0024] S3: Calculate the drift trend of sensor sensitivity based on the cross-correlation degree between the original characteristic ratio and the dynamic attenuation coefficient, and establish a compensation mapping curve based on the Pearson correlation coefficient. The compensation mapping curve adjusts non-linearly with increasing temperature. After the electronic cigarette controller receives continuous sampling data of the original characteristic ratio R and the dynamic attenuation coefficient K, it first performs time alignment between the data sequence of the original characteristic ratio R and the data sequence of the dynamic attenuation coefficient K. The alignment process adopts a sliding window mechanism with a window length of 200 sampling points. The aligned data sequence is then subjected to a fast Fourier transform to obtain frequency domain features. The frequency domain features of the R sequence are denoted as FR, and the frequency domain features of the K sequence are denoted as FK. The FR and FK are then multiplied by a complex number. The result is then subjected to an inverse Fourier transform to obtain a time-domain cross-correlation function C(t). The peak position of C(t) reflects the time delay characteristics between R and K, and the peak size reflects the degree of correlation between the two sequences.

[0025] The controller calculates the cross-correlation degree ρ between R and K based on the cross-correlation function; when calculating the sensor sensitivity drift trend, the controller standardizes the R and K sequences to eliminate the influence of dimensions and then calculates the Pearson correlation coefficient r between the two sequences; the Pearson correlation coefficient r is calculated in a sliding manner, using data from the most recent 1000 sampling points for each calculation, and by analyzing the dynamic change characteristics of the r value, the rate and direction of sensitivity drift are obtained.

[0026] The controller constructs a nonlinear compensation mapping curve M(r) based on the Pearson correlation coefficient r. The compensation mapping curve is generated using a cubic spline interpolation method, and the shape parameters of the curve are dynamically adjusted according to the distribution characteristics of the r value. When the Pearson correlation coefficient r exhibits periodic changes, the controller will periodically update the compensation mapping curve according to the change period. The output value of the compensation mapping curve M(r) serves as a correction factor for the sensor signal, used for real-time calibration of the original signal.

[0027] In this process, after the controller acquires the frequency domain features FR and FK of the original feature ratio R sequence and dynamic attenuation coefficient K sequence, it multiplies the complex conjugates of FR and FK to obtain the frequency domain cross-correlation function S(f), where FR represents the complex form of the R sequence after Fast Fourier Transform (FFT), and FK represents the complex form of the K sequence after FFT. The frequency domain cross-correlation function S(f) is frequency normalized so that its frequency components are distributed in the range of -π to π. After normalization, the controller performs a discrete inverse Fourier transform (IFT) on S(f), using the Cooley-Tukey algorithm to decompose the frequency domain data of S(f) into even and odd terms, and recursively calculates the time-domain contribution of each frequency component. During the IFT calculation, the controller applies a Hanning window function to the high-frequency components to suppress spectral leakage. After the IFT is completed, the real-valued time-domain cross-correlation function C(t) is obtained, as shown in the following equation: in, It is the frequency domain cross-correlation function. The maximum magnitude of S(f) is used for normalization. Let H(f) be the attenuation function for high-frequency components and H(f) = 0.5[1 - cos(2πf / fs)], where fs is the sampling frequency. Let W_n = 0.5[1-cos(2πn / (N-1))] be the weighting coefficients, representing the Hanning window function. The number of sampling points. For time variables, For frequency variables, The imaginary unit, Pi This represents the sequence number of the discrete sampling point.

[0028] Furthermore, after the controller acquires 1000 sampling points of the original feature ratio R sequence and the dynamic decay coefficient K sequence, it first calculates the mean of each sequence. The mean of the R sequence is obtained by summing all sampling point values ​​and dividing by the number of sampling points, denoted as Rm. The mean of the K sequence is calculated in the same way and denoted as Km. Next, the standard deviation of the two sequences is calculated. For the R sequence, the square of the difference between each sampling point value and Rm is summed, divided by the number of sampling points, and the square root is taken to obtain the standard deviation Sr. The same method is used to obtain the standard deviation Sk for the K sequence. After the mean and standard deviation are calculated, the controller calculates... The covariance of the R and K sequences is calculated by subtracting Rm from each sampling point in the R sequence and Km from each sampling point in the K sequence. The differences at corresponding positions are then multiplied and summed, and finally divided by the number of sampling points to obtain the covariance Cov. After obtaining the covariance, the controller divides it by the product of the standard deviations of the two sequences, i.e., Cov / (Sr×Sk), to obtain the Pearson correlation coefficient r. The correlation coefficient is calculated using a sliding window method. When new sampling data arrives, the controller removes the oldest data point, adds the new data point, and recalculates the correlation coefficient, thereby achieving dynamic monitoring of the correlation. The formula is as follows: in, This represents the value of the i-th sampling point in the original feature ratio sequence. This represents the value of the i-th sampling point in the dynamic attenuation coefficient sequence. The mean of the R sequence and , The mean of the K-series and , The number of sampling points, molecule part Let Covariance (Cov) be the first term in the denominator. Sr represents the standard deviation of the R series, and the second term in the denominator. Sk represents the standard deviation of the K-series.

[0029] After the controller obtains the Pearson correlation coefficient r, it first divides the domain of r into multiple sub-intervals based on the correlation strength, with the interval division based on a progressive relationship from weak to strong correlation. Three control points are selected within each sub-interval, with the endpoints being mandatory and the location of the intermediate point dynamically determined based on the distribution density of r values ​​within that interval. Once the control points are determined, a mapping function is constructed for each sub-interval using cubic spline interpolation. The interpolation process employs natural boundary conditions to ensure the continuity of function values ​​and first derivatives at the boundaries of adjacent sub-intervals. The cubic spline interpolation uses the r values ​​within the interval as the independent variable and the correction coefficient at the historical optimal compensation effect as the dependent variable. After constructing the interpolation functions for all sub-intervals, they are spliced ​​together to form a complete nonlinear compensation mapping curve M(r).

[0030] S4: The controller adjusts the gain parameters of the sensor signal conditioning circuit according to the compensation mapping curve, and adjusts the power output curve of the atomizer in conjunction with it to achieve adaptive matching between microphone sensitivity and atomization effect.

[0031] After the controller acquires the output value of the compensation mapping curve M(r), it converts it into a digital gain control signal, which is transmitted to the programmable gain amplifier (PGA) via the SPI interface. When a change in the amplitude of the sensor output signal is detected, the controller calculates the ratio of the signal to the reference level and looks up the corresponding gain adjustment value from the compensation mapping curve based on this ratio. Then, it converts this value into a digital control word for the PGA through a mapping table. The digital control word directly determines the gain factor of the PGA. The controller monitors the adjusted signal amplitude in real time through sampling feedback and continuously updates the gain parameters according to the compensation mapping curve to achieve dynamic adjustment of the signal conditioning circuit gain.

[0032] The process of converting the data into a digital gain control signal is as follows: After the controller obtains the output value of the compensation mapping curve M(r), it substitutes this value into the mapping function, which is a piecewise function, with each segment corresponding to a different gain range. Then, the controller generates a 16-bit digital control word based on the mapping result. The high 8 bits of this control word represent the integer part of the gain, and the low 8 bits represent the fractional part of the gain. The controller then packages the 16-bit digital control word according to the data frame format of the SPI communication protocol, adding a frame header identifier at the start of the data frame and a check bit at the end. After the data frame is packaged, the controller sends the data bit by bit to the PGA via the MOSI pin of the SPI bus, while simultaneously outputting a synchronization clock signal via the SCLK pin. During data transmission, the controller receives confirmation information returned by the PGA via the MISO pin, completing bidirectional communication. After confirming that the PGA has correctly received the data, the controller updates the gain parameter value in its internal register, completing the conversion process of the digital gain control signal.

[0033] After the controller completes the gain adjustment of the signal conditioning circuit, it first calculates the atomizer's reference power value based on the current gain parameters. Next, the controller acquires the sound pressure level (SPL) signal output from the microphone in real time, filters it to obtain the effective SPL value, and correlates this value with the gain parameters to obtain a real-time adjustment coefficient. Then, the controller multiplies the reference power value by the adjustment coefficient to calculate the atomizer's target output power. Once the target power is determined, the controller outputs a control signal using PWM modulation, where the PWM signal frequency is fixed and the duty cycle varies with the target power value. The controller uses a proportional-integral (PI) algorithm to adjust the PWM signal's duty cycle in real time, gradually bringing the atomizer's actual output power closer to the target value. During the adjustment process, the controller continuously acquires the microphone's SPL signal and the atomizer's actual power, dynamically optimizing the adjustment coefficient by comparing their changes, thereby establishing a mapping relationship between microphone sensitivity and atomization power. Once the mapping relationship is established, the controller adjusts the PWM signal parameters in real time according to this relationship to ensure that the atomization effect always matches the microphone's sensitivity.

[0034] It should be noted that after the controller obtains the current gain parameter, it first substitutes the gain parameter into the power mapping function, which maps the gain range to the power range. Then, it obtains the initial reference power value based on the mapping result. Next, the controller reads the amplitude of the sound pressure signal output by the microphone in real time, and normalizes the amplitude with the gain parameter to obtain a correction coefficient. After the correction coefficient is determined, the initial reference power value is multiplied by the correction coefficient to obtain the corrected reference power value. Subsequently, the controller performs compensation processing on the corrected reference power value according to the power curve characteristics to ensure the continuity and stability of the power output. Finally, the controller uses the processed power value as the final reference power value of the atomizer.

[0035] It's worth noting that while the above description outlines the overall process of adaptive matching between microphone sensitivity and atomization effect, a concrete example can better illustrate its working principle. For instance, when a user begins vaping, they initially use a gentle inhalation. At this point, the microphone detects a weak sound pressure level (SPL) signal, and the controller calculates a lower gain value and accordingly sets a lower atomization power output. Subsequently, as the user gradually increases their inhalation intensity, the SPL signal detected by the microphone strengthens, and the controller dynamically increases the gain value and simultaneously increases the atomization power, resulting in a smooth increase in vapor production. This process mirrors the natural correspondence between inhalation intensity and vapor production when using traditional cigarettes, achieving intelligent adjustment of the e-cigarette's user experience. Ultimately, the goal is to ensure that the e-cigarette responds to the user's inhalation actions in real time, providing the most suitable atomization effect.

[0036] In summary, the adaptive calibration method for electronic cigarette microphone sensitivity based on embodiments of the present invention has been clarified. It employs a dual-sensor channel to acquire the main inhalation pressure signal and the bypass airflow signal, respectively, and injects a high-frequency oscillating airflow based on the dual-source excitation principle. The method dynamically monitors the attenuation of the high-frequency component of the main inhalation pressure signal to generate a sensitivity drift trend. This improves the accuracy and real-time performance of sensitivity calibration, adapts to complex airflow environments and temperature and humidity changes, and achieves adaptive matching between microphone sensitivity and atomization effect, thereby significantly enhancing the airflow detection accuracy and user experience of electronic cigarettes.

[0037] Here, those skilled in the art will understand that the specific operations of each step in the above-described adaptive calibration method for electronic cigarette microphone sensitivity have been referenced above. Figure 1 and Figure 2 The adaptive calibration method for electronic cigarette microphone sensitivity has been described in detail, and therefore, its repeated description will be omitted.

[0038] In summary, the adaptive calibration method for electronic cigarette microphone sensitivity based on embodiments of the present invention has been clarified. It employs a dual-sensor channel to acquire the main inhalation pressure signal and the bypass airflow signal, respectively, and injects a high-frequency oscillating airflow based on the dual-source excitation principle. The method dynamically monitors the attenuation of the high-frequency component of the main inhalation pressure signal to generate a sensitivity drift trend. This improves the accuracy and real-time performance of sensitivity calibration, adapts to complex airflow environments and temperature and humidity changes, and achieves adaptive matching between microphone sensitivity and atomization effect, thereby significantly enhancing the airflow detection accuracy and user experience of electronic cigarettes.

Claims

1. A method for adaptive calibration of electronic cigarette microphone sensitivity, characterized in that, include: A dual-sensor channel is set in the electronic cigarette microphone. The first channel collects the main inhalation pressure signal and the second channel collects the bypass airflow signal. The ratio of the main inhalation pressure signal to the bypass airflow signal is used as the original feature ratio. Based on the dual-source excitation principle, a weak high-frequency oscillating airflow is injected into the microphone air intake end, and the attenuation degree of the high-frequency component in the main inhalation pressure signal is detected to generate a dynamic attenuation coefficient. The drift trend of sensor sensitivity is calculated based on the cross-correlation degree between the original characteristic ratio and the dynamic attenuation coefficient, and a compensation mapping curve is established based on the Pearson correlation coefficient. The controller adjusts the gain parameters of the sensor signal conditioning circuit according to the compensation mapping curve, and adjusts the power output curve of the atomizer in conjunction with the adjustment, so as to achieve adaptive matching between microphone sensitivity and atomization effect.

2. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 1, characterized in that, A main airway and a bypass airway are provided at the air intake end of the electronic cigarette microphone; The main airway is connected to the atomizing chamber, and the bypass airway is arranged in parallel with the main airway and separated from the main airway by an isolation structure. A first pressure sensor is installed in the main airway; a second pressure sensor is installed in the bypass airway; the first pressure sensor and the second pressure sensor are on the same cross-section and are both fixed to the sensor mounting base by elastic sealing rings; the sensor mounting base is detachably connected to the electronic cigarette shell by a threaded structure.

3. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 2, characterized in that, The first pressure sensor acquires a main inhalation pressure signal mixed with a high-frequency oscillation component, and the main inhalation pressure signal is processed by a signal conditioning circuit to obtain a pure high-frequency component A1; After the high-frequency component A1 passes through the rectifier circuit and the RC integrator circuit, the effective value A2 reflecting the high-frequency energy is obtained; The dynamic attenuation coefficient K is obtained by calculating the ratio of the effective value A2 to the original amplitude value A0 of the high-frequency oscillating airflow.

4. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 3, characterized in that, The drift trend of sensor sensitivity is calculated by performing an FFT transform on the complex product of the original characteristic ratio R and the dynamic attenuation coefficient K sequence, followed by an inverse Fourier transform to obtain the time-domain cross-correlation function, and then using a sliding window to calculate the dynamic change of the Pearson correlation coefficient r to reflect the drift trend of sensor sensitivity.

5. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 4, characterized in that, The time-domain cross-correlation function is obtained by multiplying the FFT result FR of the two sequences by the conjugate complex number of FK to obtain the frequency-domain cross-correlation function S(f). After frequency normalization and applying a Hanning window, the time-domain cross-correlation function is obtained by performing IFFT operation.

6. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 5, characterized in that, The time-domain cross-correlation function is shown in the following equation: in, It is the frequency domain cross-correlation function. The maximum magnitude of S(f) is used for normalization. For high-frequency component attenuation function, These are weighting coefficients. The number of sampling points. For time variables, For frequency variables, The imaginary unit, Pi This represents the sequence number of the discrete sampling point.

7. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 6, characterized in that, The Pearson correlation coefficient is calculated based on the time-domain cross-correlation function by calculating the mean and standard deviation of the original characteristic ratio R sequence and the dynamic decay coefficient K sequence, respectively, and then calculating the covariance and dividing the covariance by the product of the two standard deviations to obtain the Pearson correlation coefficient.

8. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 1, characterized in that, include: The domain of the Pearson correlation coefficient is divided into sub-intervals and control points are selected. Then, cubic spline interpolation is performed on each interval and the intervals are spliced ​​to obtain the nonlinear compensation mapping curve.

9. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 8, characterized in that, The gain parameter of the sensor signal conditioning circuit is adjusted according to the compensation mapping curve. The corresponding gain adjustment value is found from the compensation mapping curve based on the ratio of the sensor output signal to the reference level, converted into the digital control word of the PGA, and then adjusted in real time through the SPI interface.

10. The adaptive calibration method for electronic cigarette microphone sensitivity according to claim 9, characterized in that, The adaptive matching is achieved by the controller calculating the reference power based on the gain parameters, obtaining the adjustment coefficient by combining it with the microphone sound pressure signal, and determining the target power. The atomizer output power is controlled in real time through PWM duty cycle modulation and PI algorithm, and a mapping relationship between microphone sensitivity and atomization power is dynamically established.

Images