A method and system for correcting keyboard input data

Abstract

This application relates to the field of keyboard input data processing technology, specifically, to a keyboard input data correction method and system. It includes: acquiring a release electrical signal and storing the initial release electrical signal as the initial signal recognition benchmark for the corresponding key; continuously monitoring the release electrical signal to obtain a release electrical signal sequence; comparing the release electrical signal sequence with the initial signal recognition benchmark to generate aging trajectory information; obtaining a dynamically adjusted signal recognition standard based on the aging trajectory information; when the release electrical signal of a key at a certain release exceeds the dynamically adjusted signal recognition standard corresponding to that key, identifying this release of the corresponding key as a pseudo-double press error; and correcting the pseudo-double press error. This addresses the problem that existing keyboard signal processing mechanisms are insufficient in handling random jitter signals generated after key release, leading to "pseudo-double press" errors and causing systemic risks.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of keyboard input data processing, in particular to a keyboard input data correction method and system. BACKGROUND

[0002] In high-precision input scenarios such as finance and medicine, long-term high-intensity use of keyboards can cause structures such as key springs and rubber bowls to gradually fatigue and age, and the keys no longer reset cleanly when released, but produce subtle and irregular mechanical vibrations and rebounds. This vibration appears as extremely short and random multiple on-off pulses on the electrical signal.

[0003] Most existing keyboards only set a fixed debounce timer for the "pressing moment" jitter, and lack effective filtering for low-frequency random jitter caused by aging in the "release phase". When the duration of a release jitter accidentally exceeds the debounce threshold, the system will misjudge an actual single release as an extremely fast double press, i.e., a "pseudo double press" error, for example, "123" should be input but is recorded as "1223" or "1233".

[0004] Such errors are random, low-frequency, and hidden, and operators often mistakenly believe that it is their own mistake, making it difficult to detect in a timely manner; traditional data correction mechanisms based on rules or fixed patterns also have difficulty identifying such errors caused by hardware aging and physical disturbance, which lack stable rules. As a result, in high-risk business systems such as finance and medicine, these minor errors can accumulate for a long time, and may only be exposed during subsequent reconciliation, audit, or business anomalies, resulting in high correction costs, and even serious and irreversible impacts on business decisions or patient safety.

[0005] The prior art needs to be improved in view of the above problems. SUMMARY

[0006] The present application discloses a keyboard input data correction method and system, aiming to solve the problem that the existing keyboard signal processing mechanism lacks the ability to process random jitter signals generated after the release of the keys, resulting in "pseudo double press" errors and causing systematic risks.

[0007] The technical solution of the present application is as follows:

[0008] In a first aspect, the present application discloses a keyboard input data correction method, specifically comprising:

[0009] Obtain the release electrical signal of each key in the keyboard in the initial state, and store the initial release electrical signal as the initial signal recognition reference of the corresponding key. The release electrical signal at least includes the duration of the transition process of the key from the electrical signal closed state to the stable open state and / or the voltage fluctuation range in the transition process;

[0010] In the process of keyboard use, the release electrical signal is continuously monitored, and the running state record corresponding to the key is updated to obtain the release electrical signal sequence of the key in the process of use;

[0011] The release electrical signal sequence is compared with the initial signal identification benchmark, and the gradual change trend of the release electrical signal of each key with time is judged as the aging track information of the corresponding key;

[0012] According to the aging track information, the signal identification standard of the key is adjusted to obtain the dynamically adjusted signal identification standard, and the signal identification standard includes a signal identification threshold for distinguishing normal release behavior from abnormal release behavior;

[0013] When the release electrical signal of the key at a certain release exceeds the dynamically adjusted signal identification standard corresponding to the key, the release of the corresponding key is identified as a pseudo double press error;

[0014] The pseudo double press error is corrected so that the corresponding key is only determined as a single valid press and / or only outputs a key code corresponding to the key to the host in this key operation.

[0015] Further, according to the aging track information, the signal identification standard of the key is adjusted, including:

[0016] Obtaining the waveform feature of the release electrical signal, the waveform feature including at least one of the local nonlinearity of the signal attenuation curve, the energy distribution of the high-frequency micro-oscillation, and the density and duration distribution of the micro-pulse cluster;

[0017] Obtaining the current environmental temperature;

[0018] According to the current environmental temperature and the long-term use of the corresponding key, the preset physical mechanism mapping table is dynamically adjusted to obtain the dynamically adjusted physical mechanism mapping table, and the physical mechanism mapping table includes feature fingerprint patterns of different physical defects on the waveform of the release electrical signal;

[0019] The waveform feature of the release electrical signal is compared with the dynamically adjusted physical mechanism mapping table to diagnose the specific physical mechanism causing the abnormal release electrical signal of the corresponding key, and the specific physical mechanism is used to distinguish the specific signal abnormality driven by the latent defect and the environmental situation from the conventional aging phenomenon based on the aging track information;

[0020] According to the specific physical mechanism, a correction strategy for the specific physical mechanism is activated, and the signal identification threshold in the signal identification standard of the key is adjusted.

[0021] Further, the waveform feature of the release electrical signal is obtained, including:

[0022] The monitoring window is started, and the start time of the monitoring window is determined based on the end time of the previous key release event and the preset minimum key interval time.

[0023] Within the monitoring window, the electrical signals of the corresponding key contacts and the mechanical vibration signals under the corresponding keycaps are collected simultaneously.

[0024] Time-domain alignment and cross-correlation analysis are performed on electrical signals and mechanical vibration signals to separate the waveform features of the electrical signals directly generated by the physical release of the current key, and to filter background noise and signal residues from the previous key release event, thereby extracting the waveform features of the release electrical signal based on the separated electrical signal waveform features.

[0025] Furthermore, time-domain alignment and cross-correlation analysis are performed on the electrical signals and mechanical vibration signals, including:

[0026] The mechanical vibration signal is preprocessed to filter out specific frequency components caused by keyboard structure resonance and transient high-amplitude pulses caused by instantaneous physical impact of adjacent keys, thus obtaining the preprocessed mechanical vibration signal.

[0027] Based on the preprocessed mechanical vibration signal, the calculation window of the cross-correlation function between the electrical signal and the preprocessed mechanical vibration signal is dynamically adjusted to determine the time-domain alignment reference between the electrical signal and the preprocessed mechanical vibration signal.

[0028] Based on the time-domain alignment reference, the electrical signal and the preprocessed mechanical vibration signal are time-domain aligned.

[0029] Based on the purity of the preprocessed mechanical vibration signal, the correlation weight between the electrical signal and the preprocessed mechanical vibration signal in the cross-correlation analysis is adaptively adjusted.

[0030] Based on the time-domain aligned electrical signal and the preprocessed mechanical vibration signal, and combined with the correlation weight, cross-correlation analysis is performed to separate the waveform characteristics of the electrical signal directly generated by the physical release of the current button, and to filter background noise and signal residue.

[0031] Furthermore, by comparing the release electrical signal sequence with the initial signal recognition benchmark, the gradual change trend of the release electrical signal of each key over time is determined, and this gradual change trend is used as the aging trajectory information of the corresponding key, including:

[0032] The release duration of each key release in the release electrical signal sequence is statistically analyzed to obtain the current release duration of the corresponding key.

[0033] The current release duration is updated based on a preset trend smoothing algorithm to obtain the smooth release duration of the corresponding key.

[0034] When the number of release duration samples used to calculate the smooth release duration reaches the preset minimum number of samples, the smooth release duration is compared with the release duration of the corresponding key in the initial signal recognition benchmark. The gradual change trend of the release electrical signal of each key with time is judged. If the difference between the smooth release duration and the release duration of the corresponding key in the initial signal recognition benchmark exceeds the preset time offset threshold multiple times in a preset time window, the continuous offset of the difference in the time window is determined as the aging trajectory information of the key.

[0035] Furthermore, the mechanical vibration signal is preprocessed, including:

[0036] Real-time monitoring of keyboard internal temperature data and cumulative keystroke intensity;

[0037] Based on the internal temperature data of the keyboard and the cumulative keystroke intensity, the preset resonant frequency range and resonant suppression filter parameters are dynamically adjusted to obtain the dynamically adjusted resonant frequency range and resonant suppression filter parameters.

[0038] Based on the dynamically adjusted resonant frequency range and resonant suppression filter parameters, specific frequency components caused by keyboard structure resonance are filtered out from the mechanical vibration signal.

[0039] Obtain the current pulse spectrum characteristics of the mechanical vibration signal;

[0040] Compare the similarity between the current pulse spectrum characteristics and the preset current key physical release characteristic spectrum, and compare the similarity between the current pulse spectrum characteristics and the preset adjacent key impact characteristic spectrum to obtain the similarity comparison results;

[0041] Based on the similarity comparison results, transient high-amplitude pulses generated by instantaneous physical impacts of adjacent buttons are distinguished and isolated, and the corresponding transient high-amplitude pulses are removed or marked from the mechanical vibration signal.

[0042] Furthermore, the similarity between the current pulse spectrum characteristics and the preset current key physical release characteristic spectrum is compared, and the similarity between the current pulse spectrum characteristics and the preset adjacent key impact characteristic spectrum is also compared to obtain the similarity comparison results, including:

[0043] Real-time monitoring of microstructural changes in keyboard key materials and the impact angle and force distribution of adjacent keys;

[0044] Based on the microstructure change index of the keyboard key material, the reference range of the current physical release characteristic spectrum of the key is dynamically adjusted to obtain the dynamically adjusted current physical release characteristic spectrum of the key.

[0045] Based on the impact angle and force distribution, the distribution pattern of the impact feature spectrum of adjacent keys is updated to obtain the updated impact feature spectrum of adjacent keys.

[0046] The similarity of the current pulse spectrum feature is compared with the dynamically adjusted current physical release feature spectrum of the key. The similarity comparison takes into account the drift characteristics of the current physical release feature spectrum of the key at different life stages. The similarity of the current pulse spectrum feature is also compared with the updated impact feature spectrum of the adjacent key. The multimodal distribution characteristics of the impact feature spectrum of the adjacent key are taken into account. The similarity comparison results are obtained.

[0047] Furthermore, real-time monitoring of microstructural changes in keyboard key materials and the impact angle and force distribution of adjacent keys includes:

[0048] Real-time strain data is acquired by miniature strain sensors integrated at multiple discrete locations within the keyboard key material along the force-bearing area of ​​the key.

[0049] Based on strain data and a pre-set material fatigue characteristic curve, the microstructure change index at each discrete location is calculated.

[0050] Spatial interpolation is performed on the microstructure change index at each discrete location to generate a distribution map of the microstructure change of the key material;

[0051] Based on the microstructure change distribution map, the aging state of the overall key and local areas is evaluated to obtain the microstructure change index of the keyboard key material.

[0052] Real-time acquisition of piezoelectric signals, which are collected by miniature piezoelectric sensors deployed at multiple angles along the fingertip impact path under the keyboard keycaps;

[0053] Based on the amplitude and timing characteristics of the piezoelectric signal, the impact angle and force distribution of adjacent buttons are analyzed.

[0054] Furthermore, based on the microstructure change distribution map, the overall aging status and local areas of the button are evaluated, including:

[0055] Based on the distribution map of microstructure changes, identify regions with significant gradients where the gradient of microstructure changes is greater than a preset gradient threshold, and regions with concentrated local anomalies where the microstructure change index is greater than a preset index threshold, and obtain candidate evaluation regions.

[0056] Based on the spatial location, microstructure change gradient, and microstructure change index of the candidate evaluation region in the microstructure change distribution map, the microstructure change distribution map is dynamically divided to obtain multiple evaluation focus areas. Evaluation weights are then assigned to each evaluation focus area based on the microstructure change gradient and microstructure change index of each evaluation focus area.

[0057] Within multiple assessment areas of concern, and in combination with the corresponding assessment weights, the comprehensive aging index of each assessment area of ​​concern is calculated based on the microstructural change indicators at each discrete location within each assessment area of ​​concern.

[0058] The overall aging index of multiple assessment areas is used to evaluate the aging status of the buttons and the corresponding local areas of each assessment area.

[0059] Secondly, this application also discloses a keyboard input data correction system, comprising:

[0060] The acquisition module is used to acquire the release electrical signal of each key in the keyboard in the initial state, and store the initial release electrical signal as the initial signal identification reference of the corresponding key. The release electrical signal includes at least the duration of the transition process of the key from the electrical signal closed state to the stable open state and / or the voltage fluctuation range during the transition process.

[0061] The monitoring module is used to continuously monitor the released electrical signals during keyboard use and update the operating status record corresponding to the key to obtain the sequence of released electrical signals during key use.

[0062] The judgment module is used to compare the release electrical signal sequence with the initial signal recognition benchmark, judge the gradual change trend of the release electrical signal of each button over time, and use the gradual change trend as the aging trajectory information of the corresponding button.

[0063] The adjustment module is used to adjust the signal recognition standard of the button according to the aging trajectory information to obtain the dynamically adjusted signal recognition standard. The signal recognition standard includes the signal recognition threshold used to distinguish between normal release behavior and abnormal release behavior.

[0064] The identification module is used to identify the release of the corresponding button as a pseudo-double press error when the release electrical signal of the button exceeds the dynamically adjusted signal identification standard corresponding to the button.

[0065] The correction module is used to correct pseudo-double press errors, so that the corresponding key is only identified as a valid press in this key operation and / or only outputs the key code corresponding to the key once to the host.

[0066] Beneficial Effects: The keyboard input data correction method disclosed in this application obtains the release electrical signal of the key in its initial state as the initial signal recognition benchmark and continuously monitors the release electrical signal sequence during use. By comparing the release electrical signal sequence with the initial signal recognition benchmark, the gradual trend of the key release electrical signal changing over time is determined and used as the key's aging trajectory information. Based on the aging trajectory information, the key's signal recognition standard is dynamically adjusted, including a signal recognition threshold for distinguishing between normal and abnormal release behavior. When the release electrical signal of the key exceeds the dynamically adjusted signal recognition standard during a certain release, it is identified as a pseudo-double press error and corrected to ensure that the key operation is only determined as a valid press and / or only one key code is output to the host. This method effectively solves the problem in the prior art of difficulty in identifying and correcting pseudo-double press errors caused by key aging. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating a keyboard input data correction method provided in this application.

[0068] Figure 2 A flowchart of a keyboard input data correction system provided in this application.

[0069] In the diagram: 1. Acquisition module; 2. Monitoring module; 3. Judgment module; 4. Adjustment module; 5. Identification module; 6. Correction module. Detailed Implementation

[0070] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0071] Traditional keyboards inevitably experience aging of their key mechanisms under prolonged and intensive use, leading to minute, irregular contact vibrations upon key release. Current keyboard signal processing mechanisms typically rely on fixed debouncing timers to filter out instantaneous noise during key presses, but their processing capabilities are often insufficient for handling these random vibrations generated after key release. This can cause the system to incorrectly identify a single physical key release as two key presses, a so-called "pseudo-double press" error. Due to its randomness, low frequency, and operator cognitive biases, this type of error is often difficult to detect and correct in a timely manner, potentially accumulating in high-risk business systems and causing untraceable potential harm and systemic risks.

[0072] Reference Figure 1 In response, this application proposes a keyboard input data correction method, comprising:

[0073] S1000: Acquire the release electrical signal of each key in the keyboard in its initial state, and store the initial release electrical signal as the initial signal recognition reference for the corresponding key.

[0074] The release of the electrical signal includes at least the duration of the transition process from the button's closed state to a stable open state and / or the range of voltage fluctuations during that transition process.

[0075] S2000: During keyboard use, continuously monitor the released electrical signals and update the operating status record corresponding to the key to obtain the sequence of released electrical signals during key use;

[0076] S3000: Compares the release electrical signal sequence with the initial signal recognition benchmark, determines the gradual change trend of the release electrical signal of each key over time, and uses the gradual change trend as the aging trajectory information of the corresponding key.

[0077] S4000: Based on the aging trajectory information, the signal recognition standard of the button is adjusted to obtain a dynamically adjusted signal recognition standard. The signal recognition standard includes a signal recognition threshold used to distinguish between normal release behavior and abnormal release behavior.

[0078] S5000: When the release electrical signal of a button exceeds the dynamically adjusted signal recognition standard corresponding to the button during a certain release, the release of the corresponding button will be identified as a pseudo-double press error.

[0079] S6000: Corrects pseudo-double press errors so that the corresponding key is only identified as a valid press in this key operation and / or only outputs the key code corresponding to the key once to the host.

[0080] This application aims to effectively identify and correct "pseudo-double press" errors caused by key aging by dynamically adjusting the key signal recognition standard, thereby improving the accuracy of keyboard input data and system reliability. The method revolves around core concepts such as release electrical signal, initial signal recognition benchmark, aging trajectory information, signal recognition standard, and pseudo-double press error.

[0081] The release signal refers to the electrical signal generated during the transition of a button from a pressed state (electrical signal closed) to a fully released state (stable disconnection). This signal includes not only the instant the button contacts open, but also the electrical signal fluctuations caused by minor vibrations or rebounds that may occur during the mechanical reset process. Its characteristics may include, but are not limited to, the duration of the signal from closed to open, and the range of voltage or current fluctuations during this transition.

[0082] The initial signal recognition benchmark refers to the release electrical signal characteristics obtained for each key under ideal, wear-free conditions when the keyboard is manufactured or calibrated for the first time. This benchmark serves as a reference standard for subsequent judgment of whether the key is aging.

[0083] Aging trajectory information refers to the slow, gradual change trend of the release electrical signal of a button over time. It is used to reflect the material fatigue and wear degree of the button's mechanical structure (such as springs, rubber domes, etc.) and is a quantitative characterization of button aging.

[0084] Signal recognition standards refer to a set of rules or parameters used to distinguish between normal and abnormal button release behavior. These rules include at least a signal recognition threshold, which is used to define the normal fluctuation range of the released electrical signal.

[0085] A pseudo-double press error refers to a mechanical vibration caused by button aging, which causes the system to incorrectly identify a single physical button release as a double button press event.

[0086] In practice, the release electrical signal of each key on the keyboard in its initial state is first acquired, and the initial release electrical signal is stored as the initial signal identification reference for the corresponding key. The release electrical signal includes at least the duration of the transition process from the closed electrical signal state to the stable open state and / or the voltage fluctuation range during the transition process.

[0087] For example, after the keyboard is manufactured or before its first use, each key can be subjected to multiple standard press and release operations using automated testing equipment. During this process, a high-precision oscilloscope or data acquisition card is used to record the electrical signal waveform of each key at the moment of release. This waveform data, including the precise duration of the signal from closing to stable opening, and the range of any minute fluctuations in voltage or current that may occur during this period, is captured and stored, forming the initial signal identification benchmark for the key as a reference for its "health" status.

[0088] Secondly, during keyboard use, the released electrical signals are continuously monitored, and the operating status records corresponding to the keys are updated to obtain the sequence of released electrical signals during key use.

[0089] For example, the microcontroller inside the keyboard can integrate a low-power signal monitoring module. This module collects the release electrical signal in real time each time a key is released and records it chronologically to form a "release electrical signal sequence." This sequence includes both the latest release signal and historical data for subsequent trend analysis.

[0090] Next, the release electrical signal sequence is compared with the initial signal recognition benchmark to determine the gradual change trend of the release electrical signal of each button over time, and this gradual change trend is used as the aging trajectory information of the corresponding button.

[0091] For example, the system can analyze the release electrical signal sequence of each key according to a preset cycle (such as every certain number of key operations or every period of time). By comparing the latest data in the sequence with the initial signal recognition benchmark, it can observe whether there is a slow, cumulative shift in characteristics such as release duration and voltage fluctuation range, such as a gradual increase in release duration or an increase in voltage fluctuation range. This continuous shift is used as the aging trajectory information of the key.

[0092] Then, based on the aging trajectory information, the signal recognition standard of the button is adjusted to obtain the dynamically adjusted signal recognition standard, which includes a signal recognition threshold for distinguishing between normal release behavior and abnormal release behavior.

[0093] For example, once aging trajectory information of a button is detected, the system can dynamically adjust the signal recognition threshold of that button based on the degree of aging. If the aging trajectory shows a trend of increasing release duration, the maximum allowable duration threshold for judging a single release can be appropriately relaxed to adapt to the new normal after button aging. This adjustment process is gradual, ensuring that normal release and abnormal jitter can still be effectively distinguished even when button performance deteriorates.

[0094] Subsequently, when the release electrical signal of a button exceeds the dynamically adjusted signal recognition standard corresponding to that button during a certain release, the release of the corresponding button is identified as a pseudo-double press error.

[0095] For example, in the daily use of buttons, if the duration or fluctuation range of a release electrical signal exceeds the dynamically adjusted signal recognition threshold, the system will determine that the release is abnormal and mark it as a pseudo-double press error, indicating that the mechanical vibration of the button is sufficient to be mistakenly identified as two button operations under traditional recognition logic.

[0096] Finally, the pseudo-double press error is corrected so that the corresponding key is only identified as a valid press in this key operation and / or only the key code corresponding to the key is output to the host once.

[0097] For example, once a false double-press error is detected, the system immediately initiates a correction mechanism. Internally, the abnormal release event can be merged into part of the previous key press event, ensuring that only one key operation is recorded at the logical level. At the same time, the key codes output to the host are filtered, and only the key code corresponding to the key is output once, thereby avoiding duplicate input due to hardware aging.

[0098] The keyboard input data correction method proposed in this application can effectively solve the "pseudo-double press" problem caused by key aging during long-term use of traditional keyboards through the synergistic effect of the above steps. Its core lies in the introduction of a continuous monitoring and dynamic adjustment signal recognition standard for key "aging trajectory information".

[0099] In another embodiment of this application, S4000 is further proposed to include:

[0100] S4100: Acquire waveform characteristics of the released electrical signal, including at least one of the following: local nonlinearity of the signal attenuation curve, energy distribution of high-frequency micro-oscillation, and density and duration distribution of micro-pulse clusters;

[0101] S4200: Obtain the current ambient temperature;

[0102] S4300: Based on the current ambient temperature and the long-term usage of the corresponding buttons, the preset physical mechanism mapping table is dynamically adjusted to obtain the dynamically adjusted physical mechanism mapping table. The physical mechanism mapping table includes the characteristic fingerprint patterns of different physical defects on the waveform of the released electrical signal.

[0103] S4400: Compares the waveform characteristics of the released electrical signal with the dynamically adjusted physical mechanism mapping table to diagnose the specific physical mechanism that causes the abnormal release electrical signal of the corresponding button. The specific physical mechanism is used to distinguish between specific signal abnormalities driven by latent defects and environmental conditions and conventional aging phenomena based on aging trajectory information.

[0104] S4500: Based on a specific physical mechanism, activate a correction strategy for that specific physical mechanism and adjust the signal recognition threshold in the key signal recognition standard.

[0105] Specifically, acquiring the waveform characteristics of the release electrical signal refers to continuously acquiring the electrical signal during the button release process using a high sampling rate analog-to-digital converter (ADC), and performing signal processing such as Fourier transform and wavelet analysis on the acquired time-domain signal to extract its features in the frequency domain and time-frequency domain.

[0106] Among them, the local nonlinearity of the signal attenuation curve refers to the degree of nonlinearity of voltage or current change with time during the transition of an electrical signal from a closed state to an open state. For example, it can be quantified by calculating the rate of change of the signal attenuation slope or the residual of the fitted curve. The energy distribution of high-frequency micro-oscillation refers to the energy distribution of high-frequency noise or oscillation components superimposed on the main signal in different frequency ranges during the signal attenuation process. It is used to reflect phenomena such as poor contact of contacts and micro-vibration of materials. The density and duration distribution of micro-pulse clusters refers to the frequency and duration characteristics of short-term, high-amplitude pulse signals that may appear during the signal release process. It is used to characterize transient physical events such as contact bounce and material fatigue.

[0107] The current ambient temperature can be obtained through a temperature sensor integrated inside the keyboard, providing environmental parameters for the dynamic adjustment of the physical mechanism mapping table. In practical applications, the physical mechanism mapping table can be understood as a database storing the characteristic fingerprint patterns of the electrical signal waveforms generated by various known physical defects (such as contact oxidation, spring fatigue, microcracks, etc.) under different environmental conditions and lifespan stages. For example, the physical mechanism mapping table may include "typical waveform characteristic fingerprint patterns of contact oxidation at 25°C and under long-term use" and "typical waveform characteristic fingerprint patterns of spring fatigue in low-temperature environments." Dynamically adjusting the preset physical mechanism mapping table based on the current ambient temperature and the long-term usage of the corresponding keys means using real-time ambient temperature data and long-term usage data such as the cumulative number of key presses and total usage time to correct the parameters or adjust the weights of the characteristic fingerprint patterns in the mapping table, making it more accurately reflect the actual state of the current keys.

[0108] By comparing the waveform characteristics of the released electrical signal with a dynamically adjusted physical mechanism mapping table, the specific physical mechanism causing the abnormal release signal of the corresponding button can be diagnosed. This can be achieved through methods such as pattern recognition algorithms and machine learning classifiers. For example, the similarity or distance between the current waveform characteristics and the fingerprint patterns of each feature in the mapping table can be calculated to identify the most matching physical defect type. Specific physical mechanisms are used to distinguish between specific signal anomalies driven by both latent defects and environmental conditions and conventional aging phenomena based on aging trajectory information, thereby achieving more refined fault diagnosis. For example, a button may exhibit a conventional aging trend of gradually increasing release duration, but its waveform characteristics simultaneously show abnormal concentration of high-frequency micro-oscillation energy. In this case, it can be diagnosed as the specific physical mechanism of "slight contact oxidation," rather than simply "spring fatigue" or "structural loosening."

[0109] Based on a specific physical mechanism, a correction strategy targeting that mechanism is activated, and the signal recognition threshold in the key signal recognition standard is adjusted. The correction strategy can be a series of predefined parameter adjustment rules or algorithms. For example, when an anomaly caused by "contact oxidation" is diagnosed, a correction strategy specifically for oxidation can be activated. This strategy may include fine-tuning the recognition threshold for release duration and simultaneously adjusting the recognition threshold for voltage fluctuation range to more tolerantly handle specific waveform distortions caused by oxidation, while maintaining relatively strict recognition requirements for other types of anomalies. Through this technical solution, a more refined and adaptive adjustment of the keyboard key signal recognition standard can be achieved.

[0110] In some preferred embodiments, a specific example is given below: Suppose that after prolonged use, the release duration of the "A" key gradually increases, which is identified as aging trajectory information. According to the solution of this application, the system further acquires the electrical signal waveform characteristics of the "A" key during release. For example, analysis reveals a significant increase in the local nonlinearity of its signal attenuation curve and a specific peak in the energy distribution of the high-frequency micro-oscillation; simultaneously, the current ambient temperature is 35°C. Based on this information, the system dynamically adjusts a preset physical mechanism mapping table according to the current ambient temperature and the long-term use of the "A" key. After comparing the waveform characteristics with the dynamically adjusted physical mechanism mapping table, the system diagnoses the specific physical mechanism causing this anomaly as "contact instability caused by micro-wear on the contact surface." For this specific physical mechanism, the system activates a preset "micro-wear correction strategy," which may include fine-tuning the recognition threshold of the release duration within a specific range and imposing stricter restrictions on the energy threshold of the high-frequency micro-oscillation to allow slight contact instability while still effectively recognizing genuine secondary presses. In this way, the system can more accurately adapt to the actual aging state and specific defect type of the key, avoiding misjudgment due to overly relaxed thresholds or missed judgment due to insufficient threshold adjustment, thereby ensuring the input accuracy of the "A" key in subsequent use.

[0111] In another embodiment of this application, S4100 is further proposed to include the following:

[0112] S4110: Start the monitoring window. The start time of the monitoring window is determined based on the end time of the previous key release event and the preset minimum key interval time.

[0113] S4120: Within the monitoring window, the electrical signal of the corresponding key contact and the mechanical vibration signal under the corresponding key cap are collected simultaneously.

[0114] S4130: Performs time-domain alignment and cross-correlation analysis on electrical signals and mechanical vibration signals to separate the waveform features of the electrical signals directly generated by the physical release of the current key, and filters background noise and signal residues from the previous key release event, thereby extracting the waveform features of the release electrical signal based on the separated electrical signal waveform features.

[0115] Specifically, the monitoring window refers to a preset time period used to accurately capture electrical and mechanical vibration signals during the key release process. The start time of the monitoring window is determined based on the end time of the previous key release event and the preset minimum key interval time to ensure that the current key release event is sufficiently separated from the previous event in time, avoiding monitoring starting too early or too late, thereby ensuring that the monitoring of each key release event is independent and improving the accuracy of data acquisition.

[0116] Within the monitoring window, both electrical signals from the corresponding key contacts and mechanical vibration signals from beneath the keycaps are simultaneously acquired. The electrical signals reflect changes in the electrical contact state of the key contacts, while the mechanical vibration signals provide direct evidence of the key's physical movement. By simultaneously acquiring these two signals, a correlation can be established between the release electrical signals and the key's physical release behavior, providing a more comprehensive data foundation for subsequent analysis.

[0117] In practical applications, time-domain alignment and cross-correlation analysis are performed on electrical signals and mechanical vibration signals. Time-domain alignment eliminates minor time deviations generated during acquisition or transmission, ensuring precise synchronization between the two signals on the time axis. Cross-correlation analysis identifies and quantifies the correlation between the two signals, separating the waveform characteristics of the electrical signal directly generated by the current physical release of the button, and filtering background noise and signal residue from the previous button release event. Based on the separated electrical signal waveform characteristics, further waveform features of the released electrical signal can be extracted, such as the local nonlinearity of the signal attenuation curve, the energy distribution of high-frequency micro-oscillations, and the density and duration distribution of micro-pulse clusters.

[0118] The above technical solution avoids signal overlap and interference between different key events by precisely controlling the start time of the monitoring window. On the other hand, it separates the electrical signal features directly related to the physical release of the current key from multiple dimensions through synchronous acquisition, time-domain alignment and cross-correlation analysis of electrical signals and mechanical vibration signals, effectively suppressing the influence of background noise and signal residue.

[0119] In another embodiment of this application, S4130 further includes:

[0120] S4131: Preprocess the mechanical vibration signal to filter out specific frequency components caused by keyboard structure resonance and transient high-amplitude pulses caused by instantaneous physical impact of adjacent keys, and obtain the preprocessed mechanical vibration signal.

[0121] S4132: Based on the preprocessed mechanical vibration signal, dynamically adjust the calculation window of the cross-correlation function between the electrical signal and the preprocessed mechanical vibration signal to determine the time-domain alignment reference between the electrical signal and the preprocessed mechanical vibration signal;

[0122] S4133: Perform time-domain alignment of electrical signals and preprocessed mechanical vibration signals according to the time-domain alignment reference;

[0123] S4134: Based on the purity of the preprocessed mechanical vibration signal, adaptively adjust the correlation weight between the electrical signal and the preprocessed mechanical vibration signal in the cross-correlation analysis;

[0124] S4135: Based on the time-domain aligned electrical signal and the preprocessed mechanical vibration signal, and combined with the correlation weight, cross-correlation analysis is performed to separate the waveform characteristics of the electrical signal directly generated by the physical release of the current button, and to filter background noise and signal residue.

[0125] Specifically, preprocessing the mechanical vibration signal refers to purifying the original mechanical vibration signal before performing time-domain alignment and cross-correlation analysis. This preprocessing is used to eliminate or significantly reduce interference components in the signal. For example, by applying specific digital filters, specific frequency components caused by the resonance effect generated by the overall keyboard structure during key operation are filtered out; and by using transient signal detection and suppression algorithms, pulse signals caused by the instantaneous, high-energy physical impact generated when adjacent keys are pressed are removed. The mechanical vibration signal obtained after preprocessing is the preprocessed mechanical vibration signal, which more accurately reflects the physical release behavior of the current key, providing a high-quality reference signal for subsequent time-domain alignment and cross-correlation analysis.

[0126] Specifically, the dynamic adjustment of the cross-correlation function calculation window between the electrical signal and the preprocessed mechanical vibration signal, based on the preprocessed mechanical vibration signal, to determine the time-domain alignment benchmark, means that the system intelligently adjusts the time window range used to calculate the cross-correlation function based on the characteristics of the preprocessed mechanical vibration signal (such as its starting point, peak value, or specific waveform features). This allows the cross-correlation calculation to focus on the most relevant time period, thereby more accurately obtaining the time-domain offset between the two signals, i.e., the time-domain alignment benchmark. This time-domain alignment benchmark is the basis for subsequent precise time-domain alignment of the two signals.

[0127] In practical applications, time-domain alignment of electrical signals and preprocessed mechanical vibration signals based on a time-domain alignment benchmark refers to synchronizing the two signals on the time axis after obtaining a precise time offset, using techniques such as time shifting or resampling, so that their physical event occurrence times precisely correspond. For example, an appropriate delay or advance can be applied to either the electrical signal or the preprocessed mechanical vibration signal to make their key feature points coincide in time.

[0128] Furthermore, based on the purity of the preprocessed mechanical vibration signal, the correlation weights between the electrical signal and the preprocessed mechanical vibration signal in the cross-correlation analysis are adaptively adjusted. This means that before performing the cross-correlation analysis, the system first evaluates the quality or reliability of the preprocessed mechanical vibration signal. If the purity is high (e.g., high signal-to-noise ratio, low residual interference), a higher correlation weight is assigned in the cross-correlation analysis; if the purity is relatively low, its correlation weight is reduced to minimize its adverse impact on the final separation result. Through this adaptive adjustment mechanism, the cross-correlation analysis can fully utilize the information from high-quality mechanical vibration signals while suppressing errors caused by low-quality signals.

[0129] Therefore, based on the time-domain aligned electrical signal and the preprocessed mechanical vibration signal, and combined with the correlation weights, cross-correlation analysis is performed to separate the electrical signal waveform characteristics directly generated by the current physical release of the button, and to filter background noise and signal residue. This means that, under the premise that the two signals have been accurately aligned in the time domain and the reference weight of the mechanical vibration signal has been adaptively set, a weighted cross-correlation operation is performed to extract the electrical signal components that are highly correlated with the current physical release behavior of the button from the mixed signal, and to effectively suppress or filter background noise, signal residue from the previous button event, and other irrelevant interference components, thereby obtaining a purer and more accurate electrical signal waveform characteristic that reflects the current physical release characteristics of the button.

[0130] By employing the above technical solutions, preprocessing mechanical vibration signals, dynamically adjusting the cross-correlation function calculation window, and adaptively adjusting the correlation weights, the accuracy of time-domain alignment between electrical signals and mechanical vibration signals can be significantly improved, and the robustness of cross-correlation analysis in complex noise environments can be enhanced.

[0131] In some preferred embodiments, a specific example is given below. Assume that a miniature strain sensor acquires a mechanical vibration signal during a key release. First, the mechanical vibration signal is preprocessed. For example, a bandpass filter can be used to filter out specific frequency components caused by keyboard structural resonance (such as resonance peaks of 200Hz-300Hz), and a transient pulse detection algorithm (such as wavelet transform or adaptive thresholding) can be used to identify and remove transient high-amplitude pulses generated by accidental touches of adjacent keys, resulting in a purer mechanical vibration signal. Subsequently, based on the waveform characteristics of the preprocessed mechanical vibration signal (e.g., the starting point of its rising or falling edge), the system dynamically determines a cross-correlation function calculation window to calculate the cross-correlation function between the preprocessed mechanical vibration signal and the synchronously acquired electrical signal, thereby accurately determining the time-domain alignment reference. For example, if the rising edge of the preprocessed mechanical vibration signal starts at time t1, the calculation window can be set to [t1-Δt, t1+T], where Δt and T are preset time parameters. After determining the time-domain alignment reference, the electrical signal is time-shifted accordingly to ensure precise temporal alignment with the preprocessed mechanical vibration signal. During this process, the system can also assess the purity of the preprocessed mechanical vibration signal by calculating the signal-to-noise ratio or residual noise level: high purity results in a higher correlation weight (e.g., 0.8) in subsequent cross-correlation analysis; low purity results in a lower correlation weight (e.g., 0.5). Finally, combining the time-domain aligned electrical signal and the preprocessed mechanical vibration signal, and based on the adaptively adjusted correlation weights, cross-correlation analysis is performed to accurately separate the waveform features of the electrical signal directly generated by the physical release of the current button, effectively filtering background noise and signal residue, providing high-quality input data for subsequent signal recognition and correction.

[0132] In another embodiment of this application, S3000 is further proposed to include:

[0133] S3100: Statistically calculate the release duration of each key release in the release electrical signal sequence to obtain the current release duration of the corresponding key;

[0134] S3200: Updates the current release duration based on a preset trend smoothing algorithm to obtain the smooth release duration of the corresponding key.

[0135] S3300: When the number of release duration samples used to calculate the smooth release duration reaches the preset minimum number of samples, the smooth release duration is compared with the release duration of the corresponding key in the initial signal recognition reference. The gradual change trend of the release electrical signal of each key with time is judged. If the difference between the smooth release duration and the release duration of the corresponding key in the initial signal recognition reference exceeds the preset time offset threshold multiple times in a preset time window, the continuous offset of the difference in the time window is determined as the aging trajectory information of the key.

[0136] Specifically, the release duration of each key release in the electrical signal sequence is statistically analyzed. This refers to the system recording the time required for each key press to transition from a closed electrical signal state to a stable open state. This duration serves as a key indicator for evaluating the physical performance and aging of the key. The current release duration is updated based on a preset trend smoothing algorithm to obtain the smoothed release duration of the corresponding key. This involves using smoothing techniques such as moving averages, exponential smoothing, and Kalman filtering to reduce random noise and instantaneous fluctuations in the single release duration, thereby highlighting its long-term trend. The smoothed release duration more accurately reflects the aging state of the key.

[0137] In practical applications, the preset minimum sample size refers to the minimum number of key release duration samples that need to be accumulated before starting aging trend judgment. This ensures that the smooth release duration has sufficient statistical stability and avoids misjudgments due to insufficient samples. The preset time window refers to the time range considered when judging whether the difference exceeds the preset time offset threshold multiple times consecutively. For example, it can be set to the most recent key releases or key releases within a recent period. The preset time offset threshold is the maximum allowable deviation between the smooth release duration and the release duration in the initial signal recognition benchmark, used to define the boundary between "normal fluctuations" and "abnormal offsets". Therefore, determining the continuous offset of the difference within the preset time window as the key's aging trajectory information means that when the difference between the smooth release duration and the initial signal recognition benchmark exceeds the preset time offset threshold multiple times consecutively within the preset time window and shows a stable trend, the system recognizes this continuous offset pattern as the key's aging trajectory information, rather than random fluctuations.

[0138] In some preferred embodiments, a specific example is given below. Assume the initial release duration of a key is 15 milliseconds. During keyboard use, the system continuously monitors the key's release electrical signal and records 100 release duration data points in 100 consecutive key presses. To eliminate instantaneous fluctuations, the system uses a moving average algorithm with a window length of 50 key presses as a trend smoothing algorithm to calculate the smoothed release duration. When the cumulative sample size reaches a preset minimum sample size (e.g., 500 key presses), aging trend judgment begins. If the difference between the smoothed release duration of the key and the initial baseline of 15 milliseconds is found to exceed a preset time offset threshold of 1 millisecond for 10 consecutive times in the last 200 key presses (as a preset time window), and the smoothed release duration is consistently above 16.5 milliseconds, then the system determines this approximately 1.5 millisecond offset as the key's aging trajectory information. Based on this aging trajectory information, the system can dynamically adjust the signal recognition threshold in the signal recognition standard for this button. For example, it can appropriately relax the range of judgment on the release duration when distinguishing between normal release behavior and abnormal release behavior, so as to adapt to the slightly prolonged normal release process after aging and avoid misidentifying such normal release as a pseudo double press error.

[0139] In another embodiment of this application, S4131 further includes:

[0140] S4131-1: Real-time monitoring of keyboard internal temperature data and cumulative keystroke intensity;

[0141] S4131-2: Based on the internal temperature data of the keyboard and the cumulative key impact intensity, dynamically adjust the preset resonance frequency range and resonance suppression filter parameters to obtain the dynamically adjusted resonance frequency range and resonance suppression filter parameters.

[0142] S4131-3: Based on the dynamically adjusted resonant frequency range and resonant suppression filter parameters, filter out specific frequency components in mechanical vibration signals caused by keyboard structure resonance;

[0143] S4131-4: Obtain the current pulse spectrum characteristics of the mechanical vibration signal;

[0144] S4131-5: Compare the similarity between the current pulse spectrum characteristics and the preset current key physical release characteristic spectrum, and compare the similarity between the current pulse spectrum characteristics and the preset adjacent key impact characteristic spectrum to obtain the similarity comparison result;

[0145] S4131-6: Based on the similarity comparison results, distinguish and isolate transient high-amplitude pulses generated by instantaneous physical impacts of adjacent keys, and remove or mark the corresponding transient high-amplitude pulses from the mechanical vibration signal.

[0146] Specifically, real-time monitoring of the keyboard's internal temperature data and cumulative keystroke intensity refers to continuously acquiring real-time temperature information inside the keyboard through a temperature sensor integrated inside the keyboard, and accumulating and calculating the keystroke intensity by recording data such as the number of keystrokes and the force of each keystroke. These data collectively reflect the keyboard's current physical state and usage history.

[0147] The system dynamically adjusts the preset resonant frequency range and resonant suppression filter parameters based on the keyboard's internal temperature data and cumulative keystroke intensity. This means using monitored temperature and impact intensity data, combined with a pre-established physical model or empirical curve, to make real-time corrections to the frequency range and resonant suppression filter parameters used to filter out keyboard structural resonance. For example, when the temperature rises, the elastic modulus of the keyboard material changes, causing the resonant frequency to drift. The system adjusts the filter's center frequency, bandwidth, or Q value accordingly to keep the resonance suppression locked within the current resonant frequency range.

[0148] In practical applications, obtaining the current pulse spectrum characteristics of a mechanical vibration signal refers to performing Fourier transform or other spectral analysis on the acquired mechanical vibration signal to extract parameters such as energy distribution and peak frequency in the frequency domain, thus forming the "current pulse spectrum characteristics".

[0149] Furthermore, the similarity between the current pulse spectrum characteristics and the preset current button physical release characteristic spectrum is compared, and the similarity between the current pulse spectrum characteristics and the preset adjacent button impact characteristic spectrum is also compared. The resulting similarity comparison involves comparing the currently monitored mechanical vibration signal with two types of reference spectra: one is the current button physical release characteristic spectrum when the target button is released normally, and the other is the adjacent button impact characteristic spectrum generated through structural coupling when adjacent buttons are pressed or released. By calculating the similarity (e.g., using correlation coefficients, Euclidean distance, or discrimination scores based on machine learning classifiers), the degree of closeness between the current pulse and the aforementioned two types of reference spectra is quantified.

[0150] Therefore, based on the similarity comparison results, transient high-amplitude pulses generated by the instantaneous physical impact of adjacent keys are distinguished and isolated. The corresponding transient high-amplitude pulses are removed or marked from the mechanical vibration signal. This means that when the similarity between the current pulse and the impact characteristic spectrum of the adjacent key is significantly higher than the similarity with the physical release characteristic spectrum of the current key, the pulse is determined to be interference introduced by the instantaneous physical impact of the adjacent key, and is directly removed or marked as invalid data from the mechanical vibration signal to prevent it from entering the subsequent electrical signal analysis process.

[0151] The above technical solution enables adaptive optimization of mechanical vibration signal preprocessing: On the one hand, by dynamically adjusting the resonant frequency range and resonant suppression filter parameters using internal keyboard temperature data and cumulative key impact intensity, the problem of incomplete suppression by fixed parameter filters when the resonant frequency drifts with material properties and usage conditions is solved, thus improving the accuracy of resonant noise suppression; on the other hand, by introducing a similarity comparison mechanism between the current pulse spectrum characteristics and the current key physical release characteristic spectrum and the impact characteristic spectrum of adjacent keys, the physical release signal of the target key and the instantaneous physical impact interference of adjacent keys can be accurately distinguished, avoiding misjudgment and signal residue. This ensures the accuracy and reliability of subsequent cross-correlation analysis of electrical signals and mechanical vibration signals, providing cleaner and more reliable input data for the identification and correction of pseudo-double-press errors.

[0152] In some preferred embodiments, it is assumed that after prolonged use, the internal temperature of a keyboard increases due to environmental changes or long-term operation, and the cumulative impact intensity of a certain key has reached a high level. The system monitors the keyboard's internal temperature in real time, which is 35°C, and the cumulative impact intensity of the target key is 100,000 times. Based on preset temperature-resonance frequency mapping tables and impact intensity-resonance damping coefficient mapping tables, the system dynamically adjusts the resonance frequency range from the initial 200-250Hz to 210-260Hz, and adjusts the Q value of the resonance suppression filter from 5 to 4.5. After the mechanical vibration signal is acquired, the updated filter parameters are applied to accurately filter out specific frequency components caused by keyboard structural resonance. At this time, when a high-amplitude pulse appears in the mechanical vibration signal, the system obtains the current pulse spectrum characteristics of the pulse, for example, its spectrum has a significant energy peak in the 300-400Hz range, while the energy is lower in the 500-600Hz range. The system compares the current pulse spectrum characteristics with preset current key physical release characteristic spectrums (e.g., main energy concentrated in 500-600Hz) and preset adjacent key impact characteristic spectrums (e.g., main energy concentrated in 300-400Hz). If the comparison shows that the similarity between the pulse and the adjacent key impact characteristic spectrum is 0.85, while the similarity with the current key physical release characteristic spectrum is only 0.2, the system determines that the high-amplitude pulse is interference caused by the instantaneous physical impact of the adjacent key and removes or marks it from the mechanical vibration signal. Through the above processing, even under keyboard aging or environmental changes, the preprocessing effect of the mechanical vibration signal can still be maintained at a relatively good level, providing high-quality input for subsequent electrical signal waveform feature extraction.

[0153] In another embodiment of this application, S4131-5 further includes:

[0154] S4131-51: Real-time monitoring of microstructural changes in keyboard key materials and the impact angle and force distribution of adjacent keys;

[0155] S4131-52: Based on the microstructure change index of the keyboard key material, dynamically adjust the reference range of the current key physical release characteristic spectrum to obtain the dynamically adjusted current key physical release characteristic spectrum.

[0156] S4131-53: Based on the impact angle and force distribution, update the distribution pattern of the impact feature spectrum of adjacent keys to obtain the updated impact feature spectrum of adjacent keys.

[0157] S4131-54: Compare the current pulse spectrum characteristics with the dynamically adjusted current key physical release characteristic spectrum. When comparing the similarity, the drift characteristics of the current key physical release characteristic spectrum at different life stages are considered. Compare the current pulse spectrum characteristics with the updated adjacent key impact characteristic spectrum. When comparing the similarity, the multimodal distribution characteristics of the adjacent key impact characteristic spectrum are considered. The similarity comparison results are obtained.

[0158] Specifically, real-time monitoring of the microstructure changes in keyboard key materials refers to comprehensive microstructure change indicators used to characterize the aging degree of keyboard key materials. These indicators acquire information about physical structural changes that occur in the key materials during long-term use, such as material fatigue, wear, and changes in elastic modulus. These changes directly affect the physical vibration characteristics of the keys. The impact angle and force distribution between adjacent keys refer to the incidental impact generated by a user's finger or pressing tool on adjacent keys when a user presses a key. This impact varies depending on factors such as the user's pressing habits, key position, and pressing speed.

[0159] The system dynamically adjusts the reference range of the physical release characteristic spectrum of the keyboard keys based on the microstructural changes in the key material. This means that as keys age, the vibration spectrum generated during physical release will exhibit subtle changes such as resonant frequency shifts and alterations in damping characteristics. By monitoring these microstructural changes, the system compensates for parameters such as the peak frequency range and bandwidth of the current physical release characteristic spectrum. The current physical release characteristic spectrum used for similarity comparison reflects the key's current physical state in real time, rather than a static preset value. For example, when increased material fatigue is detected, the peak frequency range can be appropriately shifted downwards or widened.

[0160] In practical applications, updating the distribution pattern of the impact feature spectrum of adjacent keys based on the impact angle and force distribution means that when a user presses a key, the accompanying impact on adjacent keys may present in various modes, such as a light touch, a heavy blow, or an oblique impact. By monitoring the impact angle and force distribution in real time, the system constructs or updates the adjacent key impact feature spectrum library to include multiple impact modes, so as to more comprehensively and accurately match actual adjacent key impact events when comparing similarity. For example, the spectrum mode can be divided into three levels—low, medium, and high—based on the impact force, and the distribution of spectral energy in different frequency bands can be adjusted according to the impact angle.

[0161] Furthermore, when comparing the current pulse spectrum characteristics with the dynamically adjusted current button physical release characteristic spectrum, the algorithm considers the drift characteristics of the current button physical release characteristic spectrum at different stages of its lifespan. This means that the comparison algorithm will combine the cumulative usage time or number of key presses to weight or offset the reference value of the current button physical release characteristic spectrum, in order to adapt to the gradual changes in the physical characteristics of the button throughout its lifespan, from new to aged. For example, the button's spectrum may be close to ideal in the initial stage, but drift towards lower frequencies later due to material fatigue. The algorithm will predict and compensate for this drift, avoiding misjudging normal aging as abnormal.

[0162] Furthermore, when comparing the current pulse spectrum features with the updated adjacent key impact feature spectra, the multimodal distribution characteristics of the adjacent key impact feature spectra are considered. This means that the comparison algorithm does not match only a single mode, but calculates the similarity between the current pulse spectrum features and multiple modes in the adjacent key impact feature spectrum library, and selects the best matching mode to distinguish adjacent key interference caused by different impact methods or forces. For example, when the adjacent key impact feature spectrum contains both "light touch mode" and "heavy impact mode", the algorithm will compare it with both modes simultaneously to determine which type of impact the current pulse is closer to.

[0163] In some preferred embodiments, it is assumed that after prolonged use, the keycap material of a keyboard key undergoes slight plastic deformation and a decrease in elastic modulus due to repeated pressing, causing a slight shift in the peak frequency of its vibration spectrum during physical release to a lower frequency. Traditional preset spectrum comparison methods may fail to detect this shift, thus misjudging normal release signals as abnormal. According to the solution of this application, the system monitors the microstructural changes of the keyboard key material in real time using microsensors integrated within the key material, such as detecting material strain or fatigue levels. When the detected material fatigue reaches a threshold, the system dynamically adjusts the reference range of the current physical release characteristic spectrum of the key, for example, by slightly adjusting the preset peak frequency range downwards by 5Hz to match the current aging state of the key.

[0164] For example, when a user rapidly and repeatedly presses the "A" key, their finger may slightly brush against the adjacent "S" key, generating a weak transient impact. If the user habitually presses at a specific angle and with a certain force, the spectral characteristics of this impact will exhibit a stable pattern. This application's solution uses miniature piezoelectric sensors deployed under the keycaps along the fingertip impact path at multiple angles to collect impact signals in real time, thereby monitoring the impact angle and force distribution of adjacent keys. Based on this data, the system updates the distribution pattern of the impact characteristic spectrum of adjacent keys, identifying patterns such as "slight oblique impact mode" and "vertical light touch mode." During similarity comparison, the system compares the current pulse spectral characteristics with the aforementioned multimodal adjacent key impact characteristic spectra one by one, accurately determining whether the current pulse originates from the normal release of the "A" key or from slight brushing interference from the "S" key, even if the two have similarities in some frequency bands. Through this dynamic adjustment and multimodal consideration, the system can more accurately filter out transient high-amplitude pulses caused by adjacent key impacts, ensuring the purity of the key signals.

[0165] In another embodiment of this application, it is further proposed that S4131-51 can be implemented in the following manner.

[0166] S4131-511: Real-time acquisition of strain data, which is acquired by miniature strain sensors integrated at multiple discrete locations along the force-bearing area of ​​the keyboard key material.

[0167] S4131-512: Based on strain data and a preset material fatigue characteristic curve, calculate the microstructure change index at each discrete location.

[0168] S4131-513: Spatial interpolation is performed on the microstructure change index at each discrete location to generate a distribution map of the microstructure change of the key material;

[0169] S4131-514: Based on the microstructure change distribution diagram, assess the overall and local aging status of the keypad and obtain the microstructure change index of the keyboard keypad material.

[0170] S4131-515: Real-time acquisition of piezoelectric signals, which are acquired by miniature piezoelectric sensors deployed at multiple angles along the fingertip impact path under the keyboard keycaps.

[0171] S4131-516: Based on the amplitude and timing characteristics of the piezoelectric signal, analyze the impact angle and force distribution of adjacent buttons.

[0172] Strain data refers to the deformation information of a material under external load, which is collected in real time by miniature strain sensors integrated at multiple discrete locations within the keyboard key material along the force-bearing area of ​​the key. These sensors can capture the minute deformation experienced by the key during each key press and release. The material fatigue characteristic curve is a preset model describing the performance degradation law of the material under cyclic loading. Combined with the real-time collected strain data, the microstructural change index at each discrete location can be accurately calculated, such as the cumulative damage degree of the material or the local change in the elastic modulus.

[0173] Furthermore, by performing spatial interpolation on the microstructural change indicators at these discrete locations, a continuous distribution map of the key material's microstructure changes can be generated, visually demonstrating the fatigue and aging status of the key material in different areas. Based on this distribution map, the overall aging status of the key and specific local areas can be assessed, thereby obtaining more comprehensive and refined microstructural change indicators of the keyboard key material.

[0174] Furthermore, piezoelectric signals refer to the electrical charge signals generated by piezoelectric materials when subjected to mechanical stress. These signals are acquired in real time by miniature piezoelectric sensors deployed under the keycaps of a keyboard along the fingertip impact path at multiple angles. These sensors can capture the mechanical information generated when a fingertip impacts a key. By analyzing the amplitude (representing the impact force) and timing characteristics (representing the time and duration of the impact) of the acquired piezoelectric signals, the specific angle and force distribution of adjacent keys when impacted can be accurately determined, providing crucial physical context information for subsequent signal processing.

[0175] The above technical solution enables real-time and precise monitoring of microstructural changes in keyboard key materials and the distribution of impact angles and forces between adjacent keys. This refined monitoring method allows for more accurate dynamic adjustment of the reference range of the current key's physical release characteristic spectrum when comparing current pulse spectrum characteristics, and updates to the distribution pattern of the impact characteristic spectrum of adjacent keys.

[0176] In another embodiment of this application, steps S4131-514 are further described as follows:

[0177] S4131-5141: Based on the microstructure change distribution map, identify the gradient-significant regions where the microstructure change gradient is greater than the preset gradient threshold, and the local anomaly concentration regions where the microstructure change index is greater than the preset index threshold, to obtain candidate evaluation regions.

[0178] S4131-5142: Based on the spatial location, microstructure change gradient, and microstructure change index of the candidate evaluation region in the microstructure change distribution map, the microstructure change distribution map is dynamically divided to obtain multiple evaluation focus areas, and evaluation weights are assigned to each evaluation focus area according to the microstructure change gradient and microstructure change index of each evaluation focus area.

[0179] S4131-5143: Within multiple assessment areas of concern, and in combination with the corresponding assessment weights, calculate the comprehensive aging index for each assessment area of ​​concern based on the microstructural change indicators at each discrete location within each assessment area of ​​concern.

[0180] S4131-5144: Based on the comprehensive aging index of multiple assessment focus areas, assess the overall aging status of the buttons and the corresponding local areas of each assessment focus area.

[0181] Specifically, identifying regions with significant gradients where the microstructure change gradient exceeds a preset gradient threshold involves processing the microstructure change distribution map using spatial differentiation or edge detection algorithms to find areas with high microstructure change rates. These regions typically represent boundaries or transition zones of material fatigue or accelerated wear. Locally concentrated abnormal regions can be understood as areas where microstructure change indicators (e.g., crack density, percentage decrease in material hardness) are significantly higher than the average level within a local area, potentially indicating potential failure points or severe defects. The preset gradient threshold and preset indicator threshold are set based on material properties, key design life, and empirical data to distinguish between normal aging and accelerated aging or abnormal degradation. Through these steps, a series of candidate evaluation regions requiring focused attention can be obtained.

[0182] The dynamic division of the microstructure change distribution map to obtain multiple evaluation focus areas involves dividing the entire button material area into several sub-regions with different aging characteristics based on the characteristics of the candidate evaluation areas, such as their size, shape, and distance from the button's force center. An evaluation weight is assigned to each evaluation focus area to reflect its importance to the overall button function and lifespan. For example, the area below the button contact may be given a higher weight because its microstructure changes have a more direct impact on the electrical signal. The evaluation weights can be dynamically adjusted based on the area's physical location, historical usage data, and the severity of its microstructure changes.

[0183] In practical applications, a comprehensive aging index is calculated for each of the multiple assessment areas, taking into account their respective assessment weights. The aim is to provide a quantitative and comprehensive indicator of aging levels. This comprehensive aging index can be the result of a weighted average of microstructural change indicators at all discrete locations within the area, maximum value extraction, or calculation based on a specific aging model. This approach avoids the bias of judging aging status solely based on a single indicator or local observation.

[0184] The aforementioned technical solution enables a more refined and accurate assessment of the aging status of keyboard key materials. This regional, weighted assessment method allows the system to identify potential weak points or areas of accelerated aging within the keys, providing a more precise basis for subsequent adjustments to signal recognition standards. This not only improves the accuracy of identifying false double-press errors but also extends the keyboard's lifespan and enhances the user experience.

[0185] Reference Figure 2 In response, this application proposes a keyboard input data correction system, comprising:

[0186] The acquisition module 1 is used to acquire the release electrical signal of each key in the keyboard in the initial state, and store the initial release electrical signal as the initial signal identification reference of the corresponding key. The release electrical signal includes at least the duration of the transition process of the key from the electrical signal closed state to the stable open state and / or the voltage fluctuation range during the transition process.

[0187] Monitoring module 2 is used to continuously monitor the released electrical signals during keyboard use and update the running status record corresponding to the key to obtain the sequence of released electrical signals during key use.

[0188] The judgment module 3 is used to compare the release electrical signal sequence with the initial signal recognition benchmark, judge the gradual change trend of the release electrical signal of each button over time, and use the gradual change trend as the aging trajectory information of the corresponding button.

[0189] Adjustment module 4 is used to adjust the signal recognition standard of the button according to the aging trajectory information to obtain the dynamically adjusted signal recognition standard. The signal recognition standard includes a signal recognition threshold for distinguishing between normal release behavior and abnormal release behavior.

[0190] The identification module 5 is used to identify the release of the corresponding button as a pseudo-double press error when the release electrical signal of the button exceeds the dynamically adjusted signal identification standard corresponding to the button.

[0191] The correction module 6 is used to correct pseudo-double press errors so that the corresponding key is only identified as a valid press in this key operation and / or only outputs the key code corresponding to the key once to the host.

[0192] The keyboard input data correction system proposed in this application aims to dynamically monitor the aging status of the keys through the coordinated work of its various internal functional modules, and adjust the signal recognition standard accordingly, thereby effectively identifying and correcting the "pseudo-double press" error caused by key aging, and significantly improving the accuracy of keyboard input data and system reliability.

[0193] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method of correcting keyboard input data, characterized by, The method comprises: acquiring the release electrical signal of each key in the keyboard in the initial state, and storing the initial release electrical signal as the initial signal recognition reference of the corresponding key, wherein the release electrical signal at least includes the duration of the transition process of the key from the electrical signal closed state to the stable open state and / or the voltage fluctuation range in the transition process; during the use of the keyboard, continuously monitoring the release electrical signal and updating the running state record corresponding to the key to obtain the release electrical signal sequence of the key during the use; comparing the release electrical signal sequence with the initial signal recognition reference to judge the gradual change trend of the release electrical signal of each key over time as the aging track information of the corresponding key; adjusting the signal recognition standard of the key according to the aging track information to obtain the dynamically adjusted signal recognition standard, wherein the signal recognition standard includes the signal recognition threshold for distinguishing normal release behavior from abnormal release behavior; when the release electrical signal of the key at a certain release exceeds the dynamically adjusted signal recognition standard corresponding to the key, the release of the corresponding key is recognized as a pseudo double press error; correcting the pseudo double press error so that the corresponding key is only determined as a valid press once and / or only outputs a key code corresponding to the key to the host once in this key operation; adjusting the signal recognition standard of the key according to the aging track information, which comprises: acquiring the waveform characteristics of the release electrical signal, wherein the waveform characteristics include at least one of the local nonlinearity of the signal decay curve, the energy distribution of high-frequency micro-oscillation, and the density and duration distribution of micro-pulse clusters; acquiring the current environmental temperature; dynamically adjusting the preset physical mechanism mapping table according to the current environmental temperature and the long-term use of the corresponding key to obtain the dynamically adjusted physical mechanism mapping table, wherein the physical mechanism mapping table includes the feature fingerprint pattern of different physical defects on the waveform of the release electrical signal; comparing the waveform characteristics of the release electrical signal with the dynamically adjusted physical mechanism mapping table to diagnose the specific physical mechanism causing the abnormal release electrical signal of the corresponding key, wherein the specific physical mechanism is used to distinguish the specific signal abnormality driven by the latent defect and the environmental context from the conventional aging phenomenon based on the aging track information; activating the correction strategy for the specific physical mechanism according to the specific physical mechanism and adjusting the signal recognition threshold in the signal recognition standard of the key.

2. The keyboard input data correction method of claim 1, wherein acquiring the waveform characteristics of the release electrical signal, which comprises: starting a monitoring window, wherein the starting time point of the monitoring window is determined according to the end time of the previous key release event and the preset minimum key interval time; synchronously collecting the electrical signal of the corresponding key contact and the mechanical vibration signal below the corresponding key cap in the monitoring window; The electrical signal and the mechanical vibration signal are subjected to time domain alignment and cross-correlation analysis to separate the electrical signal waveform features directly generated by the current key physical release and filter background noise and signal residues of the previous key release event, so as to extract the waveform features of the release electrical signal based on the separated electrical signal waveform features.

3. The keyboard input data correction method of claim 2, wherein, The time domain alignment and cross-correlation analysis of the electrical signal and the mechanical vibration signal comprises: The mechanical vibration signal is preprocessed to filter out specific frequency components caused by keyboard structure resonance and transient high-amplitude pulses caused by instantaneous physical impact of adjacent keys, to obtain a preprocessed mechanical vibration signal; According to the preprocessed mechanical vibration signal, the cross-correlation function calculation window of the electrical signal and the preprocessed mechanical vibration signal is dynamically adjusted to determine the time domain alignment reference of the electrical signal and the preprocessed mechanical vibration signal; According to the time domain alignment reference, the electrical signal and the preprocessed mechanical vibration signal are subjected to time domain alignment; According to the purity of the preprocessed mechanical vibration signal, the correlation weight of the electrical signal and the preprocessed mechanical vibration signal in cross-correlation analysis is adaptively adjusted; According to the time domain aligned electrical signal and the preprocessed mechanical vibration signal, and in combination with the correlation weight, cross-correlation analysis is performed to separate the electrical signal waveform features directly generated by the current key physical release and filter background noise and signal residues.

4. The keyboard input data correction method of claim 1, wherein The release electrical signal sequence and the initial signal recognition reference are compared to determine the gradual change trend of the release electrical signal of each key over time, and the gradual change trend is taken as the aging trajectory information of the corresponding key, comprising: The release duration of each key release in the release electrical signal sequence is counted to obtain the current release duration of the corresponding key; The current release duration is updated based on a preset trend smoothing algorithm to obtain a smoothed release duration of the corresponding key; When the number of release duration samples used to calculate the smoothed release duration reaches a preset minimum sample number, the smoothed release duration is compared with the release duration of the corresponding key in the initial signal recognition reference to determine the gradual change trend of the release electrical signal of each key over time. If the difference between the smoothed release duration and the release duration of the corresponding key in the initial signal recognition reference exceeds a preset time offset threshold continuously for multiple times within a preset time window, the continuous offset of the difference within the time window is determined as the aging trajectory information of the key.

5. The keyboard input data correction method of claim 3, wherein, The preprocessing of the mechanical vibration signal comprises: Real-time monitoring of keyboard internal temperature data and key cumulative hitting intensity; According to the keyboard internal temperature data and the key cumulative hitting intensity, the preset resonance frequency range and resonance suppression filter parameters are dynamically adjusted to obtain dynamically adjusted resonance frequency range and resonance suppression filter parameters; According to the dynamically adjusted resonance frequency range and the resonance suppression filter parameters, specific frequency components in the mechanical vibration signal caused by keyboard structure resonance are filtered out; Obtain the current pulse spectrum feature of the mechanical vibration signal; Compare the similarity of the current pulse spectrum feature with the preset current key physical release feature spectrum and the similarity of the current pulse spectrum feature with the preset adjacent key impact feature spectrum to obtain a similarity comparison result; According to the similarity comparison result, distinguish and isolate the transient high-amplitude pulse generated by the instantaneous physical impact of the adjacent key, and remove or mark the corresponding transient high-amplitude pulse from the mechanical vibration signal.

6. The keyboard input data correction method of claim 5, wherein, Compare the similarity of the current pulse spectrum feature with the preset current key physical release feature spectrum and the similarity of the current pulse spectrum feature with the preset adjacent key impact feature spectrum to obtain a similarity comparison result, including: Real-time monitoring of the microstructure change index of the keyboard key material and the impact angle and force distribution of the adjacent key; According to the microstructure change index of the keyboard key material, dynamically adjust the reference range of the current key physical release feature spectrum to obtain a dynamically adjusted current key physical release feature spectrum; According to the impact angle and force distribution, update the distribution mode of the adjacent key impact feature spectrum to obtain an updated adjacent key impact feature spectrum; Compare the similarity of the current pulse spectrum feature with the dynamically adjusted current key physical release feature spectrum, taking into account the drift characteristics of the current key physical release feature spectrum at different service life stages during similarity comparison, and compare the similarity of the current pulse spectrum feature with the updated adjacent key impact feature spectrum, taking into account the multi-modal distribution characteristics of the adjacent key impact feature spectrum during similarity comparison, to obtain a similarity comparison result.

7. The keyboard input data correction method of claim 6, wherein, Real-time monitoring of the microstructure change index of the keyboard key material and the impact angle and force distribution of the adjacent key, including: Real-time acquisition of strain data, which is collected by micro-strain sensors integrated at multiple discrete positions along the key stress area in the keyboard key material; According to the strain data, combined with the preset material fatigue characteristic curve, calculate the microstructure change index of each discrete position; Spatially interpolate the microstructure change index of each discrete position to generate a microstructure change distribution map of the key material; According to the microstructure change distribution map, evaluate the aging state of the key as a whole and in local areas to obtain the microstructure change index of the keyboard key material; Real-time acquisition of piezoelectric signals, which are collected by micro-piezoelectric sensors arranged in multiple angular directions along the fingertip impact path under the keyboard keycap; According to the amplitude and timing characteristics of the piezoelectric signal, analyze the impact angle and force distribution of the adjacent key.

8. The keyboard input data correction method of claim 7, wherein, According to the microstructure change distribution map, evaluate the aging state of the key as a whole and in local areas, including: According to the microstructure change distribution map, a gradient significant area where the microstructure change gradient is greater than a preset gradient threshold is identified, and a local anomaly concentrated area where the microstructure change index is greater than a preset index threshold is identified, to obtain a candidate evaluation area; According to the spatial position of the candidate evaluation area in the microstructure change distribution map, the microstructure change gradient, and the microstructure change index, the microstructure change distribution map is dynamically divided to obtain a plurality of evaluation focus areas, and each evaluation focus area is assigned an evaluation weight according to the microstructure change gradient and the microstructure change index of the evaluation focus area; In the plurality of evaluation focus areas, according to the microstructure change index of each discrete position in each evaluation focus area, a comprehensive aging index of each evaluation focus area is calculated in combination with the corresponding evaluation weight; According to the comprehensive aging indexes of the plurality of evaluation focus areas, the aging states of the entire keyboard and the local areas corresponding to the evaluation focus areas are evaluated.

9. A keyboard input data correction system characterized by, In combination with the keyboard input data correction method of claim 1, the system comprises: An acquisition module is configured to acquire a release electrical signal of each key in the keyboard in an initial state, and store the initial release electrical signal as an initial signal identification reference of the corresponding key, wherein the release electrical signal at least includes a duration of a transition process of the key from an electrical signal closed state to a stable open state and / or a voltage fluctuation range in the transition process; A monitoring module is configured to continuously monitor the release electrical signal during use of the keyboard, and update a running state record corresponding to the key to obtain a release electrical signal sequence of the key during use; A judgment module is configured to compare the release electrical signal sequence with the initial signal identification reference, judge a gradual change trend of the release electrical signal of each key over time as aging trajectory information of the corresponding key; An adjustment module is configured to adjust a signal identification standard of the key according to the aging trajectory information to obtain a dynamically adjusted signal identification standard, wherein the signal identification standard includes a signal identification threshold for distinguishing between normal release behavior and abnormal release behavior; An identification module is configured to identify a pseudo double press error when a release electrical signal of the key at a certain release time exceeds the dynamically adjusted signal identification standard corresponding to the key; A correction module is configured to correct the pseudo double press error so that the corresponding key is determined to be pressed only once and / or outputs only one key code corresponding to the key in the key operation.

Images