Anti-interference filtering of touch feedback signal and instruction recognition method

Abstract

The application relates to the technical field of human-computer interaction devices, and discloses an anti-interference filtering and instruction recognition method for touch feedback signals, which comprises the following steps: acquiring a touch sensing array capacitance frame stream and establishing a space-time distribution matrix; determining an orthogonal axial capacitance evolution vector with a capacitance peak node as an origin, and extracting an asymmetric characteristic value representing anisotropy of deformation; calculating a first-order capacitance gradient between frames to generate a phase state conversion criterion; inputting the asymmetric characteristic value and the phase state conversion criterion into a touch recognition state machine; when a transition state of flexible fitting, rigid support and stable contact is met in turn and conforms to an asymmetric expansion law, the method determines that a real touch signal is obtained; the application converts a deformation process into a state machine logic sequence by using nonlinear impedance mutation characteristics of biological tissue pressing, removes symmetric expansion interference caused by water stains or mechanical vibration in principle, and enhances instruction recognition accuracy of the device under complex working conditions.

Description

Technical Field

[0001] This invention relates to an anti-interference filtering and command recognition method for touch feedback signals, belonging to the field of human-computer interaction device technology. Background Technology

[0002] Currently, capacitive sensing technology is the primary means of acquiring touch input. It converts physical touch actions by monitoring changes in capacitance on a sensor matrix. In ideal physical environments, recognition mechanisms based on spatial domain capacitance thresholds or area thresholds can provide stable interaction support. However, when interactive devices are in complex working conditions such as industrial control terminals or vehicle-mounted equipment, water splashes, electromagnetic interference, and high-frequency mechanical vibrations often generate parasitic interference signals with similar physical characteristics on the sensor matrix. Existing technologies typically assume that the interference signals are always weaker than the actual touch in terms of spatial morphology or peak intensity, and thus adopt a spatial domain filtering method with a set static capacitance threshold. Analysis shows that such schemes generally suffer from the design inertia of equating the touch deformation process with an isotropic symmetric diffusion model, focusing only on the scalar mean of the capacitance change rate in the contact area. When the capacitance distortion level generated by environmental parasitic noise coincides with the level of actual biological touch, the recognition mechanism based on static threshold comparison is prone to failure.

[0003] To address these challenges, some improvement approaches attempt to assist in discrimination by increasing sampling frequency or adding environmental sensors. However, such linear paths not only increase hardware deployment costs and exacerbate system processing latency, but also fail to penetrate the physical essence of the signal. The nonlinear impedance abrupt changes in biological tissues during compression, and the anisotropic viscoelastic relaxation characteristics generated by phalangeal support, are essential physical criteria for distinguishing between fluid disturbances and rigid vibrations, but are generally ignored in existing methods. Increasing sampling frequency or adding environmental sensors to assist in discrimination increases hardware deployment costs, exacerbates system processing latency, and has limitations at the algorithmic logic level. For example, the authorization announcement number CN11520250 Chinese invention patent 4B discloses a touch recognition method, device, and storage medium. It establishes a topology of capacitance-current change relationship through graph neural network and uses convolution operation to recognize touch patterns. The recognition mechanism based on graph topology belongs to data-driven feature clustering. The judgment logic depends on the preset edge relationship weight matrix. In dynamic evolution environments such as industrial and vehicle environments, the parasitic signals generated by the coupling of water flow patterns and mechanical vibration have spatiotemporal randomness. The pre-constructed topological substructure is difficult to cover all interference modes. The solution does not address the physical phase transformation logic of the change from flexible adhesion to rigid support when biological tissue is compressed. In the face of highly simulated environmental noise, it is difficult to avoid the risk of false positives and false negatives caused by relying on scalar feature judgment.

[0004] Therefore, the technical problem to be solved by this invention is to start from the evolution law of nonlinear physical impedance during the human hand pressing process and construct a spatiotemporal state machine recognition mechanism that can accurately isolate the interference of isotropic fluid. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for anti-interference filtering and command recognition of touch feedback signals, comprising: Step S1: Obtain the capacitance frame stream of the touch sensing array within a continuous sampling period, and map the capacitance frame stream into a spatiotemporal distribution matrix characterizing the real-time changes in the capacitance of the touch array nodes. Step S2: Monitor the capacitance peak nodes in the spatiotemporal distribution matrix, take the capacitance peak nodes as the logical origin, calculate the capacitance evolution vector of the logical origin on the orthogonal axis, and determine the asymmetric characteristic value that characterizes the anisotropic deformation law of the touch object. Step S3: Calculate the first-order capacitance gradient of the spatiotemporal distribution matrix between adjacent sampling frames, capture the capacitance change rate trajectory that characterizes the physical contact state from flexible contact to rigid support, and generate phase transition criteria. Step S4: Input the asymmetric feature value and phase transition criterion into the preset touch recognition state machine. When the asymmetric feature value meets the preset asymmetric expansion envelope and the phase transition criterion triggers the touch recognition state machine to transition sequentially from the first transition state representing flexible bonding, the second transition state representing rigid support, to the third steady state representing stable contact, the capacitive frame stream is determined to be a real touch signal. Step S5: Generate interactive control commands based on the spatial centroid trajectory of the actual touch signal within the spatiotemporal distribution matrix.

[0006] Preferably, step S3 is further refined into the following sub-steps: S31, calculate the capacitance increment of the capacitance peak node in the current sampling frame relative to the previous sampling frame, and determine the first-order capacitance gradient; S32, compare the first-order capacitance gradient with a preset rate of change slope threshold, and when the first-order capacitance gradient shows a decreasing slope that meets the rate of change slope threshold, identify the physical contact as changing from flexible layer deformation to internal rigid structure support state; S33, mark the start time of the internal rigid structure support state as the logical trigger point of the phase transition criterion.

[0007] Preferably, step S4 specifically includes: S41, in response to the first-order capacitance gradient being greater than a preset initial surge calibration value, activating a first transition state; S42, while in the first transition state, in response to the first-order capacitance gradient in subsequent sampling frames exhibiting a decay exceeding a preset negative extreme value, activating a second transition state; S43, while in the second transition state, in response to the first-order capacitance gradient converging to the zero bias interval within a specified number of frames, activating a third steady state.

[0008] Preferably, the touch recognition state machine also has a reset path: during the first transition state or the second transition state, if the asymmetric feature value shows an equivalent change component in the orthogonal direction, the capacitive frame stream is determined to be an environmental fluid interference signal, the touch recognition state machine is forced to return to the idle state and the output interface of the interactive control command is blocked.

[0009] Preferably, between step S1 and step S2, the method further includes: calculating the cumulative energy distribution of the capacitor frame stream within a preset sampling window; and performing origin zero-position calibration on the physical coordinates of the capacitor peak node based on the geometric center displacement of the cumulative energy distribution.

[0010] Preferably, step S5 is further refined as follows: during the third steady state of the touch recognition state machine, the real-time physical coordinates of the capacitance peak node corresponding to the real touch signal in the spatiotemporal distribution matrix are continuously recorded; the displacement path composed of the real-time physical coordinates is mapped to the preset gesture command library, and when the overlap between the displacement path and the preset path is higher than 85%, the interactive control command is output.

[0011] Preferably, the method further includes: monitoring the noise floor variance of the touch sensing array during non-interactive periods and constructing a dynamic noise template accordingly; and adjusting the recognition sensitivity of the touch recognition state machine in real time based on the dynamic noise template to compensate for the initial surge calibration value in step S4.

[0012] Preferably, in step S2, when extracting asymmetric feature values, the method further includes delay detection of the signal phase of the orthogonal sampling channel, and filtering out isotropic fluid interference signals by identifying the difference between the physical response delay of biological tissue under pressure and the gravity response delay of water stain fluid sliding.

[0013] Preferably, the method is applied to vehicle-mounted touch terminals or industrial display terminals. By modeling the asymmetry of the capacitive frame stream in the data domain and using state machine logic to determine the signal interference of environmental noise on the actual touch behavior, it eliminates the interference of environmental noise on the signal of the actual touch behavior.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. In anti-interference filtering and command recognition, by monitoring the temporal evolution of the data stream of the capacitance sensing matrix, the rate evolution characteristics of the contact area as it expands over time are extracted. Utilizing the unique elastic deformation hysteresis effect when a biological body touches a rigid interface, the stepwise increasing gradient of the capacitance value of surrounding nodes is identified within a microsecond-level observation time window. This allows for the fundamental separation of real touch commands with specific physical deformation processes from environmental interference signals such as water droplet splashes or mechanical vibrations that present transient large-area jumps or disordered shaking at the original source, thereby improving the command recognition certainty of the interactive system under harsh working conditions.

[0015] 2. By utilizing the anisotropic vector characteristics of the capacitor matrix nodes during the expansion process, the capacitance difference in the orthogonal directions around the logic origin is calculated and the principal force axis is established. The asymmetric viscoelastic relaxation phase difference generated by the internal skeletal constraints of biological tissue is extracted, enabling the system to identify specific temporal misalignment patterns determined by the force direction of the finger bones. This eliminates fluid spreading interference that exhibits isotropic symmetrical diffusion characteristics driven by gravity or wind. Without changing the physical structure of the sensor, the device's immunity to directional environmental noise is enhanced by reconstructing the timing logic of the array signal.

[0016] 3. A time-series state machine verification mechanism based on nonlinear biomechanical impedance evolution is introduced. By calculating the capacitance change rate between adjacent sampling frames in real time, the mechanism captures the first-order gradient of capacitance abruptly decays as the human finger presses, from flexible contact to rigid support. The complex physical phase transition process is translated into a unidirectional transition sequence of the underlying state machine. The instruction is only released when the data flow fully conforms to the trigger path of the inherent biomechanical impedance characteristics. This resolves the conflict between the homology of high-simulation environmental noise and real biological touch in terms of absolute energy scalar characteristics, and avoids the risk of false positives and false negatives caused by relying on static threshold judgment. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the feature modeling and instruction recognition of the touch feedback signal in this invention. Figure 2 This is a physical phase sequence transition diagram of the touch recognition state machine of the present invention.

[0018] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0020] A method for anti-interference filtering and command recognition of touch feedback signals, comprising: Step S1: Obtain the capacitance frame stream of the touch sensing array within a continuous sampling period, and map the capacitance frame stream into a spatiotemporal distribution matrix characterizing the real-time changes in the capacitance of the touch array nodes. Step S2: Monitor the capacitance peak nodes in the spatiotemporal distribution matrix, take the capacitance peak nodes as the logical origin, calculate the capacitance evolution vector of the logical origin on the orthogonal axis, and determine the asymmetric characteristic value that characterizes the anisotropic deformation law of the touch object. Step S3: Calculate the first-order capacitance gradient of the spatiotemporal distribution matrix between adjacent sampling frames, capture the capacitance change rate trajectory that characterizes the physical contact state from flexible contact to rigid support, and generate phase transition criteria. Step S4: Input the asymmetric feature value and phase transition criterion into the preset touch recognition state machine. When the asymmetric feature value meets the preset asymmetric expansion envelope and the phase transition criterion triggers the touch recognition state machine to transition sequentially from the first transition state representing flexible bonding, the second transition state representing rigid support, to the third steady state representing stable contact, the capacitive frame stream is determined to be a real touch signal. Step S5: Generate interactive control commands based on the spatial centroid trajectory of the actual touch signal within the spatiotemporal distribution matrix.

[0021] Preferably, step S3 is further refined into the following sub-steps: S31, calculate the capacitance increment of the capacitance peak node in the current sampling frame relative to the previous sampling frame, and determine the first-order capacitance gradient; S32, compare the first-order capacitance gradient with a preset rate of change slope threshold, and when the first-order capacitance gradient shows a decreasing slope that meets the rate of change slope threshold, identify the physical contact as changing from flexible layer deformation to internal rigid structure support state; S33, mark the start time of the internal rigid structure support state as the logical trigger point of the phase transition criterion.

[0022] Preferably, step S4 specifically includes: S41, in response to the first-order capacitance gradient being greater than a preset initial surge calibration value, activating a first transition state; S42, while in the first transition state, in response to the first-order capacitance gradient in subsequent sampling frames exhibiting a decay exceeding a preset negative extreme value, activating a second transition state; S43, while in the second transition state, in response to the first-order capacitance gradient converging to the zero bias interval within a specified number of frames, activating a third steady state.

[0023] Preferably, the touch recognition state machine also has a reset path: during the first transition state or the second transition state, if the asymmetric feature value shows an equivalent change component in the orthogonal direction, the capacitive frame stream is determined to be an environmental fluid interference signal, the touch recognition state machine is forced to return to the idle state and the output interface of the interactive control command is blocked.

[0024] Preferably, between step S1 and step S2, the method further includes: calculating the cumulative energy distribution of the capacitor frame stream within a preset sampling window; and performing origin zero-position calibration on the physical coordinates of the capacitor peak node based on the geometric center displacement of the cumulative energy distribution.

[0025] Preferably, step S5 is further refined as follows: during the third steady state of the touch recognition state machine, the real-time physical coordinates of the capacitance peak node corresponding to the real touch signal in the spatiotemporal distribution matrix are continuously recorded; the displacement path composed of the real-time physical coordinates is mapped to the preset gesture command library, and when the overlap between the displacement path and the preset path is higher than 85%, the interactive control command is output.

[0026] Preferably, the method further includes: monitoring the noise floor variance of the touch sensing array during non-interactive periods and constructing a dynamic noise template accordingly; and adjusting the recognition sensitivity of the touch recognition state machine in real time based on the dynamic noise template to compensate for the initial surge calibration value in step S4.

[0027] Preferably, in step S2, when extracting asymmetric feature values, the method further includes delay detection of the signal phase of the orthogonal sampling channel, and filtering out isotropic fluid interference signals by identifying the difference between the physical response delay of biological tissue under pressure and the gravity response delay of water stain fluid sliding.

[0028] Preferably, the method is applied to vehicle-mounted touch terminals or industrial display terminals. By modeling the asymmetry of the capacitive frame stream in the data domain and using state machine logic to determine the signal interference of environmental noise on the actual touch behavior, it eliminates the interference of environmental noise on the signal of the actual touch behavior.

[0029] Example 1: The method of the present invention is applicable to in-vehicle touch control environments with liquid adhesion and mechanical vibration. When the in-vehicle touch terminal is covered by windshield spray liquid under the condition of mechanical vibration caused by engine operation, the touch control system acquires the capacitance frame stream of the touch sensing array in a continuous sampling period and maps it into a spatiotemporal distribution matrix characterizing the real-time changes in capacitance of the touch array nodes. When the capacitance peak node in the spatiotemporal distribution matrix exceeds a preset noise energy threshold, the system uses the capacitance peak node as the logical origin, calculates the capacitance evolution vector of the logical origin on the orthogonal axis, determines the asymmetric feature value characterizing the anisotropic deformation law of the touch object, extracts the preset sensing node grid data adjacent to the capacitance peak node in physical space, and uses the spatial difference operator to calculate the sum of the absolute capacitance changes of the nodes on both sides of the horizontal wiring channel and the vertical wiring channel respectively. The cumulative difference values ​​of capacitance on the two sets of orthogonal independent axes are directly defined as the horizontal evolution component and the vertical evolution component. Thus, the scalar capacitance stream of the bottom array nodes is mapped into a two-dimensional evolution vector that can directly perform algebraic operations.

[0030] The system synchronously opens an observation time window, calculates the first-order capacitance gradient of the spatiotemporal distribution matrix between adjacent sampling frames, and determines the capacitance peak node by calculating the capacitance increment of the current sampling frame relative to the previous sampling frame. The asymmetric eigenvalues ​​and the phase transition criterion generated by the first-order capacitance gradient are input into a preset touch recognition state machine. The touch recognition state machine performs state transitions based on the dynamic changes of the first-order capacitance gradient. When the first-order capacitance gradient exceeds a preset initial surge calibration value... At that time, among them To calibrate the capacitance change rate as a measure of the initial strength of physical contact, the system enters a first transition state, corresponding to the flexible bonding process of the touch object's surface. In subsequent sampling frames, if the descent slope of the first-order capacitance gradient is found to meet a preset threshold for the rate of change slope, the system enters a second transition state, corresponding to the support process of the rigid structure inside the touch object. When the first-order capacitance gradient converges to the zero-bias interval within a specified sampling period, the system enters a third steady state, corresponding to the stable contact process.

[0031] Based on the above physical phase state recognition, the system compares the asymmetric feature values ​​with the preset asymmetric expansion envelope. If the asymmetric feature values ​​match the asymmetric expansion envelope, and the touch recognition state machine sequentially experiences the first transition state, the second transition state, and the third steady state, then the capacitive frame stream is determined to be a real touch signal. At this time, the system records the spatial centroid trajectory of the capacitive peak node within the spatiotemporal distribution matrix and matches it with the preset gesture command library, outputting the corresponding interactive control command. Here, the isotropic characteristics exhibited by fluid interference do not refer to the absolutely symmetrical spread of macroscopic fluid under gravity or environmental wind pressure, but rather to the contact of a trace amount of fluid with the surface within a limited microsecond-level extremely short observation time window. The transient initial wetting process of the surface is completely dominated by surface tension. The microscopic capacitance increment gradient generated in each orthogonal direction exhibits similar orders of magnitude. By capturing the equivalent characteristics of the transient rate of change within this microscopic time window, the system blocks the response logic in advance in the data domain, thereby removing the subsequent macroscopic asymmetric tailing deformation interference caused by gravity. For environmental fluid interference lacking internal displacement constraints, since its capacitance evolution vector exhibits equivalent distribution characteristics in orthogonal directions, the asymmetric eigenvalues ​​deviate from the asymmetric expansion envelope. The touch recognition state machine is forced to return to the idle state and stop command output. Thus, by utilizing the nonlinear change characteristics of biomechanical impedance, isotropic environmental noise components are removed at the physical level.

[0032] Example 2: In this example, the verification process uses a physical test platform including a 16x10 touch node array and a 12-bit signal sampling depth. The data source is the original capacitance sequence collected by the physical test platform under simulated vehicle electromagnetic interference environment, with a sampling period of... The determination depends on a trade-off between the completeness of touch feature capture and the energy efficiency of the sensing circuit, among which, The frame scan interval of the touch array, based on the physical characteristics of biological pressing actions, is more than 20ms. To capture the transient phase transition from epidermal deformation to finger bone support, the sampling period is... The time was set to 5ms, and an ambient noise with a signal-to-noise ratio of 15dB was superimposed in the data link to construct a composite working condition that includes 5.0mL water splash and 30Hz mechanical simple harmonic vibration.

[0033] The system maps the real-time collected touch sensing array signals into a spatiotemporal distribution matrix characterizing the evolution of node capacitance. Monitoring is then performed on the sample group of this invention and a control group that uses only a fixed capacitance threshold for determination. During touch interaction, the peak number of nodes in the original input capacitance increases from 42 units to 515 units. Key intermediate feature calculation results show that the asymmetric feature value induced by finger pressure... Reaching 1.88, it exhibits anisotropic distribution, while the asymmetric eigenvalues ​​generated by simulated water droplet interference... The fluctuation range is between 1.02 and 1.14, where, This is a dimensionless scaling factor used to quantify the deformation symmetry of the touched object, while the first-order capacitive gradient exceeds the initial surge calibration value of 90 units per frame at the moment of triggering. ,in, The capacitance change rate calibrated value, which characterizes the initial strength of physical contact, conforms to the judgment model of the transition from flexible contact to rigid support. The sample group of this invention maintains 99.2% command recognition accuracy in a noisy environment, while the control group, due to the inability to filter the symmetrical capacitance jump caused by water stains, has a command false trigger rate of 38.6%.

[0034] The water stain volume gradient control experiment observed the system response by quantitatively adding 1.0 mL, 3.0 mL, 5.0 mL, and 10.0 mL of liquid to the sensing surface. The experimental data showed that when the liquid volume was between 1.0 mL and 5.0 mL, the system maintained a recognition accuracy of over 98.5% by utilizing the synergistic effect of asymmetric eigenvalues ​​and phase transition criteria, exhibiting shielding characteristics against fluid spreading interference. When the liquid volume reached 10.0 mL, exceeding the operating range, the peak capacitance gradient saturated and decayed due to the liquid forming a large-area common-mode load across multiple sensing units, causing the recognition accuracy to drop to 75.3%. This performance inflection point indicates the physical boundary defined by the sampling window and the feature envelope. By removing a portion of the missing second transition state judgment in the touch recognition state machine, the false trigger rate of the system increased by 21.8% under a simple mechanical vibration environment, confirming the contribution of the nonlinear evolution logic of biomechanical impedance in stripping away environmental parasitic signals.

[0035] Example 3: In a calibration scenario where factory parameters of a human-computer interaction device are determined, the system quantitatively analyzes the nonlinear impedance characteristics generated by the pressure of biological tissue, establishes a judgment criterion for the touch recognition state machine, acquires capacitance data when the object touches the touch sensing array, and extracts the capacitance peak nodes in the spatial coordinate system. Axial direction and Evolution vector components in the axial direction and Calculate the ratio of the two to determine the asymmetric eigenvalues. That is, satisfying ,in, This represents the lateral evolution component of the capacitance peak node within the current observation time window. As a longitudinal evolution component, it was determined by statistically analyzing 500 sets of biological touch sample data. The numerical distribution region between 1.6 and 2.1 represents an asymmetric expansion envelope. When extracting asymmetric eigenvalues, the touch terminal microcontroller extracts the eigenvalues ​​and eigenvectors of the capacitance distribution covariance matrix of the adjacent sensing node array centered on the capacitance peak node. The microcontroller establishes the direction of the eigenvector corresponding to the maximum eigenvalue as the principal axis of force and the direction orthogonal to the principal axis of force as the secondary axis. It calculates the first capacitance evolution gradient along the principal axis of force from the logic origin and the second capacitance evolution gradient along the secondary axis, respectively. The ratio of the first capacitance evolution gradient to the second capacitance evolution gradient is used to generate asymmetric eigenvalues, thus freeing the extraction reference from the constraints of the physical orthogonal wiring channels of the sensing array. The asymmetric eigenvalues ​​are then extracted and aligned with the correct direction. When detecting phase delay of the sampling channel signal, the touch control chip triggers a local high-frequency interrupt scanning mode to pause the global sensing array traversal scanning when the capacitance increment of the peak node reaches the initial surge calibration value. It alternately transmits square wave excitation signals to the horizontal driving channel and the vertical receiving channel where the peak node is located. The built-in hardware timer records the first time stamp when the capacitance response of the horizontal driving channel reaches the preset offset threshold and the second time stamp when the capacitance response of the vertical receiving channel reaches the preset offset threshold. The absolute time difference between the first and second time stamps is calculated to obtain the physical response delay parameter. When the physical response delay parameter is greater than the underlying circuit's bias delay, the physical response delay of biological tissue under pressure is identified.

[0036] The system synchronously determines the initial surge calibration value. The system utilizes a sensing circuit to capture the first-order capacitance gradient at the moment of finger skin contact, monitors the peak change rate of the capacitance growth curve within the first 10ms, and collects the transient current variation generated by the flexible skin contacting the sensing surface, converting it into capacitance change. The statistical average of the capacitance change is then used to... The threshold is set within a range of 85 to 105 capacitance units per frame. The determination of the slope threshold depends on the zero-point detection of the second derivative of the capacitance gradient. When the second derivative drops from a positive value to below zero, the deformation of the touched object changes from surface adhesion to internal skeletal support. The slope of the decreasing capacitance gradient corresponding to this inflection point serves as the criterion for the driving state machine to switch from the first transition state to the second transition state. To eliminate the interference of the operator's subjective pressing and deceleration behavior on the slope evolution characteristics, the system introduces a basic [mechanism / mechanism] before performing zero-point mapping. Impedance window self-consistency verification: Only when the absolute magnitude of the first-order capacitance gradient has accumulated to the system's pre-calibrated initial skin deformation energy threshold, and the drop in the second derivative exhibits extremely short-term nonlinear abrupt changes consistent with biorheology, is it determined to be a necessary physical phase transition caused by the rigid support of the skeleton, rather than a simple artificial force reduction delay. When the touch recognition state machine switches from the first transition state to the second transition state, the duration of the first-order capacitance gradient maintaining a negative extreme value decay spans a preset number of consecutive sampling frames. The number of consecutive sampling frames is determined by dividing the time constant of the first-order capacitance gradient recorded by the bionic silicone head impacting the touch sensing array within a preset standard range (within which Shore hardness is within a preset standard range) from its peak to the zero bias interval by the system sampling period. If the duration is greater than the basic frame number, the second transition state is activated. If the duration is less than or equal to the basic frame number, the input signal is determined to be a transient rigid impact, forcing the touch recognition state machine to return to the idle state. The system maps the real-time signal features in the spatiotemporal distribution matrix to the touch recognition state machine. When the real-time collected first-order capacitance gradient is detected to be greater than 95 units per frame and the duration is greater than 3 sampling periods, the initial surge condition is determined to be met, and the logic pointer enters the first transition state. In subsequent frame stream processing, if the rate of decrease of the first-order capacitance gradient reaches the slope threshold of 15 units per frame squared, the logic pointer enters the second transition state. This corresponds to the physical process of the finger transitioning from soft contact with the epidermis to rigid support of the phalanx. The above physical characteristics correspond to deterministic parameter indicators and state switching criteria, providing a reproducible execution path for instruction recognition.

[0037] Example 4: In a newly installed industrial control terminal deployment scenario, to compensate for the differences in reference capacitance at each node caused by uneven mechanical stress distribution of the mounting bracket and parasitic inductance introduced by the wiring environment, the system continuously collects 100 frames of capacitance frame streams under zero-contact conditions and calculates the electrostatic field bias value of each sensing node. The resulting mean matrix is ​​stored as a reference distribution map. The system monitors the capacitance fluctuation of the touch sensing array during no-load operation and calculates the standard deviation of the node capacitance change. Determine the noise energy threshold, specifically the standard deviation. The value is set as three times the initial threshold for triggering the command. Random electrical noise caused by frequency converter interference is filtered out at the signal level. This pre-calibration process provides a differentially processed input source for subsequent extraction of the anisotropic deformation law of the touch object, making the asymmetric feature values... The calculation is no longer affected by the background noise of a specific installation environment.

[0038] When the touch terminal enters continuous operation and faces the drift of the cover plate dielectric constant due to ambient temperature cycling, the system uses the idle state monitoring gap in the touch recognition state machine to adjust the initial surge calibration value. The environmental compensation factor is determined by monitoring the background change slope of the non-touch area in the spatiotemporal distribution matrix, along with the change rate slope threshold. ,Will With the factory specifications The original values ​​are multiplied to obtain the real-time updated action trigger threshold, so as to maintain the sensitivity of the first-order capacitance gradient judgment during the flexible contact stage of the finger skin. When the time axis shift of the zero point of the second derivative of the node capacitance is detected, the system compensates for the slope threshold of the rate of change according to the shift, so as to match the change of material relaxation time constant caused by the increase of cover temperature. Through this dynamic baseline tracking and parameter compensation mechanism, the touch system maintains real-time calibration of the command recognition criteria and enters a stable interactive state.

[0039] Example 5: In a production line operation where the factory preset values ​​of the touch control chip are determined, the system controls bio-touch simulation heads with different geometric radii to determine the numerical boundary of the asymmetric expansion envelope. The system adjusts the simulation head to contact the touch sensing array at an angle of 10 to 45 degrees through a drive mechanism and collects the capacitance evolution vector in the spatiotemporal distribution matrix. as well as Calculate the ratio of the two to determine the asymmetric eigenvalues. That is, satisfying ,in, These are asymmetric eigenvalues. This represents the lateral evolution component of the capacitance peak node within the observation time window. For the longitudinal evolution component, the numerical interval corresponding to the 95% confidence level in the statistical distribution is selected as the fixed asymmetric expansion envelope, and the induced current intensity of the simulation head at the instant of downward contact is detected simultaneously. And according to the relation The transient energy integral of the capacitive frame stream is converted into the capacitance change rate, and the minimum value of the capacitance change rate in five consecutive calibration tests is selected as the initial surge calibration value. The trigger limit is used to determine the activation criterion for the first transition state.

[0040] When the touch system completes signal recognition and enters the command matching stage of spatial center-of-gravity trajectory, the system processes the displacement path of the capacitance peak node through coordinate normalization. It maps the acquired real-time trajectory coordinate point sequence to a standard 1x1 unit coordinate system to eliminate feature offsets caused by differences in the physical size of the touch surface. The system then calculates the Euclidean distance between the normalized real-time trajectory and the standard sequence in the preset gesture command library at the corresponding sampling node. and using the formula Determine the degree of overlap, where, The percentage of overlap. This represents the spatial distance between the corresponding sampling nodes. The total number of coordinate points, when the calculated... When the value is higher than 85% and the touch recognition state machine is stably in the third steady state, the system outputs the corresponding interactive control command. Through normalization processing and distance vector operation, the command recognition process has a definite judgment logic when facing trajectory fluctuations caused by mechanical vibration. When performing the overlap degree extraction calculation, in order to prevent the calculation result from being a negative number that does not conform to the physical meaning due to excessive spatial distance deviation, the system is configured with a truncation judgment mechanism in the underlying numerical processing logic. When the Euclidean distance value of a single sampling node is greater than the unit coordinate system limit of 1 due to abnormal drift, the calculation module forcibly truncates the distance deviation of that point and limits it to 1. At the same time, the lower limit of the overlap degree result is preset to zero to ensure that the multidimensional trajectory comparison formula always converges monotonically within a certain percentage scale space.

[0041] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for anti-interference filtering and command recognition of touch feedback signals, characterized in that, include: Step S1: Obtain the capacitance frame stream of the touch sensing array within a continuous sampling period, and map the capacitance frame stream into a spatiotemporal distribution matrix characterizing the real-time changes in the capacitance of the touch array nodes. Step S2: Monitor the capacitance peak nodes in the spatiotemporal distribution matrix, take the capacitance peak nodes as the logical origin, calculate the capacitance evolution vector of the logical origin on the orthogonal axis, and determine the asymmetric characteristic value that characterizes the anisotropic deformation law of the touch object. Step S3: Calculate the first-order capacitance gradient of the spatiotemporal distribution matrix between adjacent sampling frames, capture the capacitance change rate trajectory that characterizes the physical contact state from flexible contact to rigid support, and generate phase transition criteria. Step S4: Input the asymmetric feature value and phase transition criterion into the preset touch recognition state machine. When the asymmetric feature value meets the preset asymmetric expansion envelope and the phase transition criterion triggers the touch recognition state machine to transition sequentially from the first transition state representing flexible bonding, the second transition state representing rigid support, to the third steady state representing stable contact, the capacitive frame stream is determined to be a real touch signal. Step S5: Generate interactive control commands based on the spatial centroid trajectory of the actual touch signal within the spatiotemporal distribution matrix.

2. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, Step S3 is further refined into the following sub-steps: S31, calculate the capacitance increment of the capacitance peak node in the current sampling frame relative to the previous sampling frame, and determine the first-order capacitance gradient; S32, compare the first-order capacitance gradient with the preset rate of change slope threshold, and when the first-order capacitance gradient shows a decreasing slope that meets the rate of change slope threshold, identify the physical contact as changing from flexible layer deformation to internal rigid structure support state; S33, mark the start time of the internal rigid structure support state as the logical trigger point of the phase transition criterion.

3. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, Step S4 specifically includes: S41, in response to the first-order capacitance gradient being greater than a preset initial surge calibration value, activating the first transition state; S42, while in the first transition state, in response to the first-order capacitance gradient in subsequent sampling frames exhibiting a decay exceeding a preset negative extreme value, activating the second transition state; S43, while in the second transition state, in response to the first-order capacitance gradient converging to the zero bias interval within a specified number of frames, activating the third steady state.

4. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, The touch recognition state machine also has a reset path: during the first transition state or the second transition state, if the asymmetric feature value shows an equivalent change component in the orthogonal direction, the capacitive frame stream is determined to be an environmental fluid interference signal, the touch recognition state machine is forced to return to the idle state and the output interface of the interactive control command is blocked.

5. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, Between steps S1 and S2, the method further includes: calculating the cumulative energy distribution of the capacitor frame stream within a preset sampling window; and performing origin zero-position calibration on the physical coordinates of the capacitor peak node based on the geometric center displacement of the cumulative energy distribution.

6. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, Step S5 is further refined as follows: During the third steady state of the touch recognition state machine, the real-time physical coordinates of the capacitance peak node corresponding to the real touch signal in the spatiotemporal distribution matrix are continuously recorded; the displacement path composed of the real-time physical coordinates is mapped to the preset gesture command library, and when the overlap between the displacement path and the preset path is higher than 85%, the interactive control command is output.

7. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, Also includes: Monitor the noise floor variance of the touch sensing array during non-interactive periods and construct a dynamic noise template accordingly; The initial surge calibration value in step S4 is used to adjust the recognition sensitivity of the touch recognition state machine based on the real-time compensation of the dynamic noise template.

8. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, In step S2, when extracting asymmetric feature values, the delay detection of the signal phase of the orthogonal sampling channel is also included. By identifying the difference between the physical response delay of biological tissue under pressure and the gravity response delay of water stain fluid sliding, isotropic fluid interference signals are filtered out.

9. The method for anti-interference filtering and command recognition of touch feedback signals according to claim 1, characterized in that, The method is applied to vehicle-mounted touch terminals or industrial display terminals. By modeling the asymmetry of the capacitive frame stream in the data domain and using state machine logic to identify it, the signal interference of environmental noise on the actual touch behavior is eliminated.

Images