A system and method for a capacitive sensor

Abstract

The application relates to the technical field of capacitive sensor measurement, and specifically discloses a system and method for a capacitive sensor, which first applies a detection signal to excite and collect an initial response of the system, extracts damping and frequency stability characteristics from the initial response; then, based on the characteristics, generates adaptive excitation shaping control parameters through a pre-trained model decision; controls a periodic excitation signal with adjustable edge slope to be generated according to the parameters and applied to the sensor; finally, synchronously collects and compensates an optimized output signal of the sensor to obtain a final measurement value; the application suppresses noise from the source by active sensing and adaptive excitation, and improves dynamic measurement precision and the adaptability of the system to different environments.

Description

Technical Field

[0001] This invention relates to the field of capacitive sensor measurement technology, and more specifically to a system and method for capacitive sensors. Background Technology

[0002] In the field of precision measurement based on capacitive sensors, especially in applications requiring dynamic interaction with the external environment or objects, such as robotic haptics and force feedback for minimally invasive surgical instruments, a long-standing technical challenge is the significant "ringing" noise introduced by the composite resonant system formed by the coupling between the sensor's mechanical structure and the interacting object. Traditional measurement methods typically employ excitation signals with fixed waveforms (such as standard square waves), whose abundant harmonic components easily excite this composite system to produce continuous, damped oscillations. These oscillations, superimposed on the actual sensing signal, create strong interference, severely reducing the signal-to-noise ratio and real-time performance of the measurement signal.

[0003] Existing technologies lack a way to effectively suppress time-varying and unpredictable compound resonance (ringing) noise generated by the coupling between capacitive sensors and dynamically changing environments without sacrificing measurement real-time performance and bandwidth, thereby achieving high-precision and robust interactive measurements. One method for capacitive sensors, traditionally based on a fixed excitation signal, cannot actively adapt to different or time-varying resonance conditions, resulting in limited noise suppression effectiveness. Summary of the Invention

[0004] The purpose of this invention is to provide a system and method for capacitive sensors to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for a capacitive sensor includes the following steps:

[0007] S1: Apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment.

[0008] S2: Acquire the first output signal of the capacitive sensor in response to the detection excitation signal;

[0009] S3: Analyze the first output signal to simultaneously extract the damping coefficient eigenvalue and natural frequency deviation eigenvalue, which characterize the dynamic response of the composite resonant system; construct a binary eigenvector from the damping coefficient eigenvalue and natural frequency deviation eigenvalue, and use it as the input of the pre-established capacitive sensor control model; output a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge.

[0010] S4: According to the excitation shaping control parameters, control the excitation generation circuit to generate a shaped periodic excitation signal, and apply the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters.

[0011] S5: Acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process;

[0012] S6: Process the second output signal to obtain the final measurement value corresponding to the measured quantity.

[0013] As a further aspect of the present invention: S1 specifically includes:

[0014] A detection excitation signal is generated, which is an excitation sequence consisting of at least three consecutive voltage step pulses with a fixed time interval, wherein the polarity of adjacent step pulses alternately reverses and the amplitude increases.

[0015] The excitation sequence is applied to the excitation terminal of the capacitive sensor;

[0016] The acquisition of the first output signal is triggered synchronously with the application of the last step pulse in the excitation sequence.

[0017] As a further aspect of the present invention: the calculation process of the characteristic value of the damping coefficient is as follows:

[0018] From the first output signal, extract the free decaying oscillation waveform segment generated by the composite resonant system after excitation by the edge of a single pulse of the detection excitation signal;

[0019] Envelope detection is performed on the free decaying oscillation waveform segment to obtain the decaying envelope of its oscillation amplitude as a function of time;

[0020] Calculate the logarithmic decay rate of the attenuation envelope within a preset time window, and directly calculate and output the characteristic value of the damping coefficient based on the defined relationship between the logarithmic decay rate and the damping ratio.

[0021] As a further aspect of the present invention: the calculation process for the deviation of the natural frequency from the eigenvalue is as follows:

[0022] Zero-crossing detection is performed on the waveform segment in the first output signal corresponding to the steady-state forced oscillation of the composite resonant system, and the time interval sequence between consecutive zero-crossings is recorded;

[0023] Calculate the statistical variance of the time interval sequence and use the statistical variance as a primary deviation indicator to characterize the jitter of the oscillation period;

[0024] The primary deviation index is compared with a pre-stored baseline variance obtained under calibration conditions, and the ratio of the two is calculated and output as the natural frequency deviation characteristic value.

[0025] As a further aspect of the present invention: the construction process of the capacitive sensor control model is as follows:

[0026] The damping coefficient eigenvalues ​​and natural frequency deviation eigenvalues ​​are combined to construct a binary eigenvector;

[0027] The binary feature vector is input into the pre-established capacitive sensor control model, which is a multinomial regression model; the binary feature vector is used as input to output a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge.

[0028] The pre-training process of the multinomial regression model is as follows: multiple sets of known binary feature vectors are used as training samples, the ideal excitation shaping control parameters corresponding to each set of training samples are used as training objectives, and the training is carried out with the goal of minimizing the sum of squared errors between the predicted excitation shaping control parameters and the ideal excitation shaping control parameters, until the sum of squared errors converges.

[0029] Obtain the excitation shaping control parameters output by the multinomial regression model.

[0030] As a further aspect of the present invention: S4 specifically includes:

[0031] The parameter value used to specify the rise time in the excitation shaping control parameters is converted into the drive current value for the first controllable current source.

[0032] When a pulse rising edge needs to be generated, a first controllable current source is used to charge an integrating capacitor with a constant current, so that the voltage across the integrating capacitor rises linearly at a rate determined by the driving current value.

[0033] The parameter value used to specify the fall time in the excitation shaping control parameters is converted into the driving current value of the second controllable current source. When it is necessary to generate a pulse falling edge, the second controllable current source is switched to discharge the integrating capacitor with constant current, so that the voltage across the integrating capacitor decreases linearly at the corresponding rate.

[0034] The voltage across the integrating capacitor is buffered and then output as the shaped periodic excitation signal.

[0035] As a further aspect of the present invention: the process of obtaining the driving current value is as follows:

[0036] Based on the specified rise time parameter value in the excitation shaping control parameters, a pre-stored rise time and drive current mapping table is queried to obtain the corresponding reference drive current value.

[0037] The reference drive current value is multiplied by a real-time correction factor to generate the drive current value;

[0038] The process for determining the real-time correction factor is as follows: Under the current ambient temperature, the integrating capacitor is charged using a first controllable current source with a reference current value, the actual charging time is measured, and the ratio of the rated charging time to the actual charging time is calculated as the real-time correction factor.

[0039] As a further aspect of the present invention: S5 specifically includes:

[0040] A sampling trigger pulse synchronized with the center point of the flat region of each pulse of the generated and shaped periodic excitation signal;

[0041] After the rising edge of the sampling trigger pulse is triggered, a programmable time delay is started to wait for the delay time specified by the excitation shaping control parameters in order to avoid the transient response phase that exists after the pulse edge switching;

[0042] After the delay time ends, the second output signal is synchronously sampled once to obtain a voltage sample value;

[0043] The system samples multiple consecutive pulse cycles of the formed periodic excitation signal and then performs an arithmetic average of the obtained voltage sample values ​​to obtain the second output signal.

[0044] As a further aspect of the present invention: S6 specifically includes:

[0045] The second output signal is differentially processed with a pre-calibrated baseline capacitance response value to obtain a baseline-corrected primary measurement signal.

[0046] Using the primary measurement signal as input, a nonlinear calibration curve storing the capacitance-physical quantity conversion relationship is queried to obtain preliminary physical quantity estimates.

[0047] Obtain the parameter values ​​used to specify the rise time and fall time of the pulse edge in the excitation shaping control parameters, and calculate the edge smoothness compensation coefficient based on these parameter values;

[0048] Multiply the preliminary physical quantity estimate by the edge smoothness compensation coefficient to calculate and output the final measurement value.

[0049] A system for a capacitive sensor includes:

[0050] The system excitation module is used to apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment.

[0051] The initial response acquisition module is used to acquire the first output signal of the capacitive sensor in response to the detection excitation signal;

[0052] The adaptive parameter decision module is used to analyze the first output signal to simultaneously extract the damping coefficient feature value and the natural frequency deviation feature value, which characterize the dynamic response of the composite resonant system. The damping coefficient feature value and the natural frequency deviation feature value are used to construct a binary feature vector, which is used as the input of the pre-established capacitive sensor control model. The module outputs a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge.

[0053] The waveform adjustable excitation generation module controls the excitation generation circuit to generate a shaped periodic excitation signal according to the excitation shaping control parameters, and applies the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters.

[0054] The response acquisition module is optimized to acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process;

[0055] The measurement value calculation module is used to process the second output signal to obtain the final measurement value corresponding to the measured quantity.

[0056] The beneficial effects of this invention are:

[0057] (1) When the sensor is coupled with an unknown or time-varying environment, the fixed excitation waveform of the traditional capacitive sensing method is prone to generating unpredictable composite resonances, producing strong "ringing noise" and severely degrading the signal quality. This invention first applies a probe signal to actively excite and quantify the key characteristics (damping and frequency stability) of the composite resonance system, and then dynamically generates an excitation signal with optimal edge smoothness to match it based on these characteristics, thereby minimizing excessive energy injection into the resonance system from the excitation source. This "sensing first and then adapting" adaptive mechanism can effectively suppress the amplitude of ringing noise, resulting in a significant improvement in the signal-to-noise ratio of the sensor output signal obtained in subsequent measurements. Especially in dynamic interactive scenarios such as contact force measurement and soft robot tactile sensing, more stable and accurate measurement results can be obtained.

[0058] (2) This invention compensates for the slow drift of analog circuit parameters (such as current sources and capacitors) through a real-time correction factor, enabling it to respond to changes in system resonance characteristics caused by environmental variations. The final signal processing stage introduces a smoothness compensation coefficient related to the excitation edge parameters, further correcting the systematic measurement deviations introduced by the excitation waveform changes themselves. This multi-level, closed-loop compensation and adaptation strategy enables the measurement system to automatically adjust to a better working state when facing environmental fluctuations, device aging, or changes in the measurement object, maintaining the consistency of measurement results, reducing frequent reliance on external calibration, and improving overall long-term reliability. Attached Figure Description

[0059] The invention will now be further described with reference to the accompanying drawings.

[0060] Figure 1 This is a flowchart of the method of the present invention;

[0061] Figure 2 This is a system block diagram of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] Please see Figure 1 As shown, the present invention is a method for a capacitive sensor, comprising the following steps:

[0064] S1: Apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment.

[0065] S2: Acquire the first output signal of the capacitive sensor in response to the detection excitation signal;

[0066] S3: Analyze the first output signal to simultaneously extract the damping coefficient eigenvalue and natural frequency deviation eigenvalue, which characterize the dynamic response of the composite resonant system; construct a binary eigenvector from the damping coefficient eigenvalue and natural frequency deviation eigenvalue, and use it as the input of the pre-established capacitive sensor control model; output a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge.

[0067] S4: According to the excitation shaping control parameters, control the excitation generation circuit to generate a shaped periodic excitation signal, and apply the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters.

[0068] S5: Acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process;

[0069] S6: Process the second output signal to obtain the final measurement value corresponding to the measured quantity.

[0070] In steps S1 and S2, firstly, the probe excitation signal is generated. The probe excitation signal is not a single pulse, but an excitation sequence consisting of at least three consecutive voltage step pulses. Each pulse has a steep edge, with a typical width of 10 milliseconds. Adjacent pulses in the sequence have opposite polarities; that is, if the first pulse is a positive step, the second is a negative step, and the third returns to a positive step. Furthermore, the pulse amplitudes exhibit an increasing pattern; for example, the amplitude of the second pulse is set to 120% of the amplitude of the first pulse, and the amplitude of the third pulse is set to 120% of the amplitude of the second pulse. This amplitude increase aims to progressively enhance the energy injection into the composite resonant system based on the alternating polarity reversal, thereby making its oscillation response more significant and facilitating subsequent feature extraction. The specific parameters of this excitation sequence, including the number of pulses, time interval, fundamental amplitude, and increment ratio, can be precisely controlled by the microcontroller's timer module and generated by the corresponding analog voltage steps output by its built-in or external digital-to-analog converter.

[0071] Next, the excitation sequence is applied to the excitation terminal of the capacitive sensor. Physically, the output of the signal generation unit is directly connected to one electrode of the capacitive sensor via a low-impedance drive circuit. This drive circuit must have sufficient output current capability to ensure that the steep edge of the excitation signal is not affected by the sensor's equivalent input capacitance load. During application, the entire excitation sequence is continuously and uninterruptedly transmitted to the sensor as a complete test signal packet, thereby actively perturbing the composite resonant system formed by the coupling between the sensor's mechanical structure and its environment, forcing it to produce a mixed response of free-dampening oscillations and forced oscillations that incorporate the system's dynamic characteristics.

[0072] Finally, the acquisition of the first output signal is triggered synchronously. To ensure that the acquired signal can completely capture the final response state of the composite resonant system to the complete excitation sequence, the triggering time of the acquisition is strictly synchronized with the application of the last step pulse in the excitation sequence. Specifically, in the control logic, the microcontroller generates a hardware trigger signal, such as a rising edge, from its designated general-purpose input / output pin while sending the instruction for the last pulse. This trigger signal is directly connected to the external trigger input pin of the analog-to-digital converter to start the sampling process and begin recording the voltage change at the output of the capacitive sensor. From this trigger point, the acquisition process will continue for a preset duration, which at least covers the time required for the free oscillation of the composite resonant system to decay to a value less than 5% of the initial maximum oscillation amplitude, to ensure that the complete ringing decay process that can be used for quantitative analysis is recorded.

[0073] In S3, firstly, the damping coefficient characteristic value is extracted. From the stored first output signal data sequence, based on the known timing of the probe excitation pulse edge, a waveform data segment is identified and extracted after any (usually the last) steep pulse edge excitation, where the system enters the free-dampening oscillation phase. The length of this extracted segment must ensure that it includes the complete process of oscillation decaying to below a preset amplitude threshold (e.g., 5% of the initial amplitude). Envelope detection is performed on the extracted free-dampening oscillation waveform segment to obtain the decay trend line of the oscillation amplitude over time, rather than the instantaneous oscillation details. One implementation is to calculate the peak point sequence for each oscillation period of the waveform segment, and then connect these discrete peak points into a continuous, smooth decaying envelope using linear interpolation or curve fitting methods. Subsequently, the logarithmic decay rate of this decaying envelope within the time window is calculated. Specifically, two points separated by an integer number of oscillation periods are selected on the envelope, and their amplitudes are denoted as... and The number of complete oscillation cycles contained within the time interval between them is denoted as Logarithmic decay rate The following formula is used for calculation: ;

[0074] in, Represents the natural logarithm operation. It is the envelope amplitude at an earlier time. It is the envelope amplitude at a later point in time. (After obtaining...) After that, damping ratio Calculated based on its relationship with the definition of logarithmic decay rate: Finally, the calculated damping ratio The damping coefficient characteristic value is directly used as the output. .

[0075] Secondly, extract the characteristic values ​​of the natural frequency deviation. This requires analyzing the waveform segment in the first output signal corresponding to the steady-state forced oscillation of the composite resonant system. This waveform segment is usually located in the flat region of the last pulse in the excitation sequence, after the oscillation has reached steady state. Zero-crossing detection is performed on this steady-state waveform segment, recording the precise time points when the signal voltage crosses zero level from positive to negative or from negative to positive. The time interval between two consecutive zero-crossings is then calculated to obtain one oscillation period. The sequence of oscillations is used. To characterize the stability of the system frequency, the statistical variance of this oscillation period sequence is calculated. The variance is calculated as follows: first, calculate the average of all period values ​​of the sequence. Then calculate according to the following formula: ;in, The number of cycles, Indicates the first The variance calculated over each period This is a primary deviation indicator that characterizes the oscillation period jitter under the current environment. To obtain a dimensionless relative deviation measure with a clear physical meaning, this primary deviation index is used. With a pre-stored benchmark variance Compare the baseline variance. The oscillation period variance is calculated using the same method described above after applying the same detection excitation signal to the same capacitive sensor in a quiet, stable, and interference-free calibration environment. Ultimately, the natural frequency deviates from the eigenvalue. The ratio of the two is calculated as follows: ;

[0076] The ratio This intuitively reflects the degree to which the current measurement environment, relative to the ideal calibration environment, affects the inherent frequency stability of the system. The larger the value, the worse the frequency stability.

[0077] Finally, the damping coefficient characteristic value obtained above is used... Deviation from the eigenvalue of the natural frequency Construct an ordered binary array, i.e., a binary feature vector: ;

[0078] Binary feature vectors This serves as input to a pre-established capacitive sensor control model. The control model is essentially a feature mapping relationship established through training. Its pre-construction and training process includes: measuring and calculating multiple corresponding feature vector samples under laboratory conditions for various known composite resonant system scenarios with different damping and frequency stability; simultaneously, determining the optimal excitation pulse edge rise and fall times to effectively suppress ringing noise in each scenario through experimental optimization or simulation. Using these paired training data, a polynomial regression method is used for fitting and training. The general form of the polynomial regression model is that the output parameters are polynomial functions of the input feature values. For example, for the rise time... Its model can be expressed as including and The sum of polynomials of all orders and their interactions. Optimization algorithms such as least squares are used to adjust the coefficients of each term in the polynomials to improve the rise time of the model's predictions for all training samples. and descent time The sum of squared overall errors between the model and its actual optimal value is minimized. After training, the model internally solidifies the features derived from the feature space. To control parameter space The mapping relationship. In practical applications, the binary feature vectors calculated in real time will be used. Inputting this trained model, the model can output a set of corresponding specific excitation shaping control parameters, namely the target rise time of the excitation pulse, based on its internal polynomial calculation rules. and target descent time .

[0079] In S4, firstly, the parameter value used to specify the target rise time in the excitation shaping control parameters is converted into a specific drive command for the first controllable current source. This conversion is accomplished by querying a pre-stored rise time-drive current mapping table in non-volatile memory. This mapping table is pre-calibrated and established during the circuit production and debugging phase. Its construction method is as follows: for the fixed combination of the first controllable current source and the integrating capacitor in the circuit, an external precision instrument controls the first controllable current source to charge the integrating capacitor with a series of known different current values ​​using a constant current. Simultaneously, the actual time taken for the integrating capacitor voltage to rise from zero to the target high level (e.g., 3 volts) is accurately measured; this is the measured rise time. Multiple sets of corresponding data pairs of "drive current value" and "measured rise time" are recorded, thereby establishing and storing the lookup table. During actual operation, the microcontroller searches this table based on the target rise time parameter value, finding the drive current value corresponding to the closest measured rise time, and uses this as the reference drive current value. To compensate for the drift of circuit parameters such as the integrating capacitor capacitance and current source transconductance with temperature and time, a real-time correction factor is also introduced. The process for determining the real-time correction factor is as follows: Before each formal generation of the excitation signal, the microcontroller controls the first controllable current source to charge the same integrating capacitor with a known, fixed reference current value, and measures the actual charging time required for the integrating capacitor voltage to rise from zero to the target high level. The theoretically stored rated charging time (the time that should be achieved using this reference current value under calibration conditions) is divided by the measured actual charging time; the quotient is the real-time correction factor. Finally, the reference drive current value obtained from the query is multiplied by this real-time correction factor, and the product is used as the final actual drive current value set for the first controllable current source. This drive current value is typically sent in the form of a digital control word through the microcontroller's digital interface to the digital-to-analog converter or digital potentiometer controlling the first controllable current source, thereby generating the corresponding analog control voltage.

[0080] Secondly, when the rising edge of the pulse needs to be generated, the control circuit switches to the charging state. Specifically, an analog switch controlled by a microcontroller or dedicated logic circuit connects the output of the first controllable current source to one end of the integrating capacitor. The other end of the integrating capacitor is grounded or connected to a fixed reference potential. The first controllable current source begins constant current charging of the integrating capacitor according to the previously set actual drive current value. According to the basic voltage-current relationship of a capacitor, under constant current charging conditions, the voltage across the integrating capacitor will rise linearly at a constant rate, which is equal to the drive current value divided by the capacitance of the integrating capacitor. Therefore, by precisely controlling the magnitude of the drive current value, the slope of the voltage rise can be precisely controlled, thereby achieving precise control of the pulse rising edge time. When the voltage of the integrating capacitor reaches the preset target high-level voltage, a voltage comparator outputs a signal, triggering the control logic to stop the charging process, thus forming a flat top for the pulse.

[0081] Then, at the falling edge when a pulse needs to be generated, the control circuit switches to the discharge state. This process is similar to the rising edge but with the current direction reversed. First, the parameter value used to specify the target fall time in the excitation shaping control parameters is processed in a similar manner to the rising edge process. That is, by querying another pre-stored fall time and drive current mapping table (which is constructed in a similar way to the rise time table but obtained using a second controllable current source for discharge testing), and combining it with the same real-time correction factor (or a correction factor measured separately for the second current source based on the same principle), the actual drive current value set for the second controllable current source is calculated. Next, the control analog switch disconnects from the first controllable current source and connects the second controllable current source to the integrating capacitor. The second controllable current source is configured to sink current mode, which draws current from the integrating capacitor at a constant value of the set actual drive current, thereby causing the voltage across the integrating capacitor to decrease linearly at a constant rate, the rate of decrease being equal to the drive current value divided by the capacitance of the integrating capacitor. When the voltage of the integrating capacitor drops to a preset low level, another voltage comparator triggers the control logic to stop discharging, thus forming a flat bottom of the pulse and preparing for the rising edge of the next cycle.

[0082] Finally, the voltage signal across the integrating capacitor is processed to drive the sensor. Since the integrating capacitor has limited load-carrying capacity, its voltage cannot directly drive a sensor that may have a capacitive load. Therefore, a voltage buffer with high input impedance and low output impedance (such as a voltage follower circuit) is connected to the input of the integrating capacitor. This buffer draws almost no current from the preceding integrating capacitor, thus not affecting the linearity of the constant current charging and discharging process; simultaneously, its low output impedance provides sufficient current to ensure that the shaped voltage waveform can be applied quickly and accurately to the excitation terminal of the capacitive sensor without waveform distortion due to load effects. The signal obtained from the output of this buffer is the final required shaped periodic excitation signal, with pulse edge rise and fall times strictly controlled by the excitation shaping control parameters. This signal is continuously or on demand applied to the excitation terminal of the capacitive sensor for subsequent optimized measurements.

[0083] In S5, firstly, a sampling trigger pulse is generated, synchronized with the center point of each pulse flat region of the shaped periodic excitation signal. The waveform parameters of this excitation signal are known, and its pulse width, rise time, and fall time are determined by the excitation shaping control parameters. The microcontroller's timer module is configured to remain synchronized with the generation of this excitation signal. Within each pulse period of the excitation signal, the timer starts an internal countdown after detecting the end of the pulse rising edge (i.e., the start of the flat region). When the countdown reaches half the total duration of the pulse flat region, a high-level pulse is generated by a general-purpose input / output pin of the microcontroller as the sampling trigger pulse, thereby ensuring that each trigger point is located at the exact center of the pulse voltage stable region.

[0084] Secondly, after the rising edge of the sampling trigger pulse, sampling is not performed immediately; instead, a programmable time delay is initiated. The length of this delay is determined by the excitation shaping control parameters, and its purpose is to avoid transient voltage fluctuations or ringing transients that may exist inside the sensor and its interface circuits due to distributed parameters after the pulse edge switching. Specifically, this delay time is set to the greater of the rise time and fall time of the current excitation pulse, multiplied by a preset safety factor (e.g., 1.5). After the trigger edge arrives, the microcontroller starts another internal timer or programmable delay counter, counting to the delay time value determined by the above calculation.

[0085] Then, after the preset delay time ends, the microcontroller immediately sends a start conversion command to the analog-to-digital converter through its control bus, or the delay end signal is directly used as an external trigger signal for the analog-to-digital converter to synchronously sample and convert the instantaneous voltage value of the second output signal output by the capacitive sensor, thereby obtaining an accurate voltage sample value corresponding to the stable state of the excitation signal.

[0086] Finally, to improve the measurement signal-to-noise ratio and accuracy, the triggering, delaying, and sampling process is repeated for multiple consecutive pulse cycles (e.g., 16 cycles) of the formed periodic excitation signal. Multiple voltage sample values ​​(e.g., 16 values) are temporarily stored in memory. After a predetermined number of samples are completed, the microcontroller reads these data and performs an arithmetic average calculation. Specifically, all sample values ​​are summed to obtain a total, which is then divided by the total number of sample values. The quotient is the second output signal value representing the current stable response of the sensor, used for subsequent processing. This averaging process effectively suppresses the influence of random noise on the single sampling result.

[0087] In S6, firstly, baseline correction is performed to eliminate the inherent sensor offset. With the sensor unaffected by the measured quantity (i.e., zero-load state), a stable output value is pre-measured using the complete procedure described herein and stored in non-volatile memory as the pre-calibrated baseline capacitance response value. When processing the real-time acquired second output signal, the microcontroller reads this baseline value from memory and subtracts it from the value of the second output signal; the difference is the baseline-corrected primary measurement signal. This operation removes the inherent DC offset of the sensor and circuitry.

[0088] Secondly, the primary measurement signal is converted into a preliminary estimate of the physical quantity. Since there is usually a non-linear relationship between the output electrical signal of the capacitance sensor and the measured physical quantity, direct linear conversion will introduce errors. Therefore, before the equipment leaves the factory, a set of discrete "capacitance signal-physical quantity" corresponding data points are established through precise calibration experiments, and a continuous calibration curve characterizing this non-linear relationship is generated using methods such as polynomial fitting. This curve is pre-stored in the form of coefficients of a mathematical formula or a lookup table. The microcontroller takes the value of the primary measurement signal as input, searches within this pre-stored curve relationship, or substitutes it into the formula for calculation, thereby obtaining a preliminary estimate of the physical quantity proportional to the measured quantity.

[0089] Then, the edge smoothness compensation coefficient is calculated. Since this invention uses variable pulse edge rise and fall times, different edge velocities can cause slight differences due to non-ideal factors such as charge injection effects within the sensor, thus systematically affecting the absolute accuracy of the measurement results. To compensate for this effect, an edge smoothness compensation coefficient is introduced. The calculation process for the edge smoothness compensation coefficient is as follows: The target rise time parameter value (denoted as Tr) and the target fall time parameter value (denoted as Tf) are read from the excitation shaping control parameters. Simultaneously, a predefined reference edge time parameter (denoted as Tref, set as the minimum edge time achievable by the method) is used. The edge smoothness compensation coefficient (denoted as Kc) is obtained through the calculation rules described below: the edge smoothness compensation coefficient Kc equals the number one, plus the product of a first weighting coefficient and (Tr minus Tref), plus the product of a second weighting coefficient and (Tf minus Tref). The first and second weighting coefficients are a set of constants determined by linear regression analysis after measuring different combinations of Tr and Tf under standard experimental conditions and comparing the results with the benchmark true values. Their values ​​have been pre-calibrated and stored.

[0090] Finally, the final measurement value is calculated and output. The preliminary physical quantity estimate is multiplied by the calculated edge smoothness compensation coefficient Kc. The product is the final measurement value after nonlinear calibration and edge effect compensation. This value can then be output via a digital interface or used for subsequent display and control.

[0091] Please see Figure 2 As shown, a system for a capacitive sensor includes:

[0092] The system excitation module is used to apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment.

[0093] The initial response acquisition module is used to acquire the first output signal of the capacitive sensor in response to the detection excitation signal;

[0094] The adaptive parameter decision module is used to analyze the first output signal to simultaneously extract the damping coefficient feature value and the natural frequency deviation feature value, which characterize the dynamic response of the composite resonant system. The damping coefficient feature value and the natural frequency deviation feature value are used to construct a binary feature vector, which is used as the input of the pre-established capacitive sensor control model. The module outputs a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge.

[0095] The waveform adjustable excitation generation module controls the excitation generation circuit to generate a shaped periodic excitation signal according to the excitation shaping control parameters, and applies the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters.

[0096] The response acquisition module is optimized to acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process;

[0097] The measurement value calculation module is used to process the second output signal to obtain the final measurement value corresponding to the measured quantity.

[0098] The working principle of this invention is as follows: First, a detection excitation signal consisting of at least three steep step pulses with alternating polarity reversals and increasing amplitudes is applied to the capacitive sensor to actively excite the composite resonant system formed by the coupling between the sensor and the measurement environment, and its initial response signal is acquired simultaneously. Next, the response signal is analyzed, and the damping coefficient characteristic value and natural frequency deviation characteristic value, which characterize the dynamic properties of the system, are extracted and constructed into a binary feature vector. This feature vector is input into a control model pre-established through polynomial regression training, and the model outputs a set of adaptive excitation shaping control parameters. Then, based on these control parameters, the current of the constant current source charging and discharging the integrating capacitor is precisely controlled by querying a pre-stored mapping table and combining it with a real-time correction factor, thereby generating a shaped periodic excitation signal with strictly controlled pulse edge rise and fall times, which is then applied to the sensor. Subsequently, at the center point of the stable region of each pulse of the shaped excitation signal, after a programmable delay to avoid transients, the sensor output signal is synchronously sampled, and the sampled values ​​of multiple cycles are averaged to obtain a stable second output signal. Finally, baseline correction and nonlinear calibration are performed on the signal to obtain preliminary estimates of physical quantities. A smoothness compensation coefficient calculated based on the current excitation edge time is then introduced for correction, ultimately outputting high-precision measurement results. This method, through a closed loop of "excitation-analysis-adaptive adjustment-optimized measurement," effectively suppresses composite resonance effects and compensates for systematic errors under different measurement environments.

[0099] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for a capacitive sensor, characterized in that, Includes the following steps: S1: Apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment. S2: Acquire the first output signal of the capacitive sensor in response to the detection excitation signal; S3: Analyze the first output signal to simultaneously extract the damping coefficient eigenvalue and natural frequency deviation eigenvalue, which characterize the dynamic response of the composite resonant system; construct a binary eigenvector from the damping coefficient eigenvalue and natural frequency deviation eigenvalue, and use it as the input of the pre-established capacitive sensor control model; output a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge. S4: According to the excitation shaping control parameters, control the excitation generation circuit to generate a shaped periodic excitation signal, and apply the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters. S5: Acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process; S6: Process the second output signal to obtain the final measurement value corresponding to the measured quantity, specifically including: The second output signal is differentially processed with a pre-calibrated baseline capacitance response value to obtain a baseline-corrected primary measurement signal. Using the primary measurement signal as input, a nonlinear calibration curve storing the capacitance-physical quantity conversion relationship is queried to obtain preliminary physical quantity estimates. Obtain the parameter values ​​used to specify the rise time and fall time of the pulse edge in the excitation shaping control parameters, and calculate the edge smoothness compensation coefficient based on these parameter values; Multiply the preliminary physical quantity estimate by the edge smoothness compensation coefficient to calculate and output the final measurement value.

2. The method for a capacitive sensor according to claim 1, characterized in that, S1 specifically includes: A detection excitation signal is generated, which is an excitation sequence consisting of at least three consecutive voltage step pulses with a fixed time interval, wherein the polarity of adjacent step pulses alternately reverses and the amplitude increases. The excitation sequence is applied to the excitation terminal of the capacitive sensor; The acquisition of the first output signal is triggered synchronously with the application of the last step pulse in the excitation sequence.

3. The method for a capacitive sensor according to claim 1, characterized in that, The calculation process for the characteristic value of the damping coefficient is as follows: From the first output signal, extract the free decaying oscillation waveform segment generated by the composite resonant system after excitation by the edge of a single pulse of the detection excitation signal; Envelope detection is performed on the free decaying oscillation waveform segment to obtain the decaying envelope of its oscillation amplitude as a function of time; Calculate the logarithmic decay rate of the attenuation envelope within a preset time window, and directly calculate and output the characteristic value of the damping coefficient based on the defined relationship between the logarithmic decay rate and the damping ratio.

4. The method for a capacitive sensor according to claim 1, characterized in that, The calculation process for the deviation of the natural frequency from the eigenvalue is as follows: Zero-crossing detection is performed on the waveform segment in the first output signal corresponding to the steady-state forced oscillation of the composite resonant system, and the time interval sequence between consecutive zero-crossings is recorded; Calculate the statistical variance of the time interval sequence and use the statistical variance as a primary deviation indicator to characterize the jitter of the oscillation period; The primary deviation index is compared with a pre-stored baseline variance obtained under calibration conditions, and the ratio of the two is calculated and output as the natural frequency deviation characteristic value.

5. A method for a capacitive sensor according to claim 1, characterized in that, The process of constructing the capacitive sensor control model is as follows: The damping coefficient eigenvalues ​​and natural frequency deviation eigenvalues ​​are combined to construct a binary eigenvector; The binary feature vector is input into the pre-established capacitive sensor control model, which is a multinomial regression model; the binary feature vector is used as input to output a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge. The pre-training process of the multinomial regression model is as follows: multiple sets of known binary feature vectors are used as training samples, the ideal excitation shaping control parameters corresponding to each set of training samples are used as training objectives, and the training is carried out with the goal of minimizing the sum of squared errors between the predicted excitation shaping control parameters and the ideal excitation shaping control parameters, until the sum of squared errors converges. Obtain the excitation shaping control parameters output by the multinomial regression model.

6. A method for a capacitive sensor according to claim 1, characterized in that, S4 specifically includes: The parameter value used to specify the rise time in the excitation shaping control parameters is converted into the drive current value for the first controllable current source. When a pulse rising edge needs to be generated, a first controllable current source is used to charge an integrating capacitor with a constant current, so that the voltage across the integrating capacitor rises linearly at a rate determined by the driving current value. The parameter value used to specify the fall time in the excitation shaping control parameters is converted into the driving current value of the second controllable current source. When it is necessary to generate a pulse falling edge, the second controllable current source is switched to discharge the integrating capacitor with constant current, so that the voltage across the integrating capacitor decreases linearly at the corresponding rate. The voltage across the integrating capacitor is buffered and then output as the shaped periodic excitation signal.

7. A method for a capacitive sensor according to claim 6, characterized in that, The process for obtaining the driving current value is as follows: Based on the specified rise time parameter value in the excitation shaping control parameters, a pre-stored rise time and drive current mapping table is queried to obtain the corresponding reference drive current value. The reference drive current value is multiplied by a real-time correction factor to generate the drive current value; The process for determining the real-time correction factor is as follows: Under the current ambient temperature, the integrating capacitor is charged using a first controllable current source with a reference current value, the actual charging time is measured, and the ratio of the rated charging time to the actual charging time is calculated as the real-time correction factor.

8. A method for a capacitive sensor according to claim 1, characterized in that, S5 specifically includes: A sampling trigger pulse synchronized with the center point of the flat region of each pulse of the generated and shaped periodic excitation signal; After the rising edge of the sampling trigger pulse is triggered, a programmable time delay is started to wait for the delay time specified by the excitation shaping control parameters in order to avoid the transient response phase that exists after the pulse edge switching; After the delay time ends, the second output signal is synchronously sampled once to obtain a voltage sample value; The system samples multiple consecutive pulse cycles of the formed periodic excitation signal and then performs an arithmetic average of the obtained voltage sample values ​​to obtain the second output signal.

9. A system for a capacitive sensor, characterized in that, A method for performing a capacitive sensor according to any one of claims 1-8 includes: The system excitation module is used to apply a probe excitation signal to the excitation terminal of the capacitive sensor. The probe excitation signal contains a pulse with a steep edge to excite the composite resonant system formed by the coupling of the mechanical structure of the capacitive sensor with the current measurement environment. The initial response acquisition module is used to acquire the first output signal of the capacitive sensor in response to the detection excitation signal; The adaptive parameter decision module is used to analyze the first output signal to simultaneously extract the damping coefficient feature value and the natural frequency deviation feature value, which characterize the dynamic response of the composite resonant system. The damping coefficient feature value and the natural frequency deviation feature value are used to construct a binary feature vector, which is used as the input of the pre-established capacitive sensor control model. The module outputs a set of excitation shaping control parameters for controlling the waveform of the excitation pulse edge. The waveform adjustable excitation generation module controls the excitation generation circuit to generate a shaped periodic excitation signal according to the excitation shaping control parameters, and applies the periodic excitation signal to the excitation terminal of the capacitive sensor. The rise time and fall time of the pulse edge of the shaped periodic excitation signal are specified by the excitation shaping control parameters. The response acquisition module is optimized to acquire the second output signal of the capacitive sensor in response to the periodic excitation signal of the forming process; The measurement value calculation module is used to process the second output signal to obtain the final measurement value corresponding to the measured quantity.

Images