Capacitive touch screen anti-interference method and system based on multi-mode dynamic atlas evolution and self-adaptive cooperative compensation

Abstract

The invention provides a capacitive touch screen anti-interference method and system based on multi-mode dynamic graph evolution and self-adaptive cooperative compensation. Belongs to the technical field of capacitive touch screen man-machine interaction anti-interference. The method comprises the following steps: carrying out environment, equipment and signal three-dimensional sensing area division on a working environment of the capacitive touch screen to generate multi-modal sensing area data; deploying multiple sensors according to the multi-modal sensing area data, and constructing a multi-modal sensing network; intelligent trigger learning condition judgment is carried out according to the multi-mode sensing network, and a learning process is started; in the learning process, wide-band multi-dimensional interference signal acquisition is carried out to obtain an original interference signal data set; through the multi-mode dynamic graph evolution and self-adaptive cooperative compensation technology, various interference sources can be accurately identified and dynamically adapted in a complex electromagnetic environment, the interference sensing capacity and compensation precision of the capacitive touch screen are remarkably improved, and the interference suppression rate is increased to 98% or above.

Description

Technical Field

[0001] This invention proposes a capacitive touchscreen anti-interference method and system based on multimodal dynamic graph evolution and adaptive collaborative compensation, belonging to the field of capacitive touchscreen human-computer interaction anti-interference technology. Background Technology

[0002] Capacitive touchscreens have become the mainstream human-computer interaction method due to their advantages such as no mechanical wear and intuitive operation. However, the capacitive sensing principle is extremely sensitive to electromagnetic interference. Currently, typical interference sources exhibit characteristics of "multi-source, complexity, and transient nature": power supply interference has been expanded to include high-frequency switching noise (2MHz-5MHz) from GaN fast charging and common-mode interference from vehicle high-voltage power distribution systems; display driver interference has expanded to include crosstalk of multi-channel driver signals in Mini / MicroLED screens; wireless communication interference covers terahertz signal leakage from 6G prototypes and burst pulses in automotive-grade V2X communication; and environmental coupling interference has been expanded to include strong electromagnetic radiation from medical radiofrequency ablation equipment and electromagnetic radiation from industrial laser equipment.

[0003] Existing anti-interference technologies suffer from three major shortcomings: First, hardware filtering schemes use fixed topologies, which cannot cope with wide-band interference; for example, the SAW filter's suppression effect drops sharply for interference above 2MHz. Second, software algorithms rely on static noise models, making them insufficiently adaptable to time-varying and non-stationary interference (such as transient interference from vehicle motor start-stop); Kalman filtering experiences error amplification of more than three times during sudden interference changes. Third, there is a "timing disconnect" between interference compensation and touch detection; traditional baseline tracking algorithms require more than 300ms to recover after sudden interference changes, far exceeding the real-time requirements (≤50ms) of automotive and medical scenarios. Therefore, it is urgent to break through the "passive filtering" paradigm and construct a full-link anti-interference technology system of "active perception - dynamic modeling - real-time compensation". Summary of the Invention

[0004] This invention provides a capacitive touchscreen anti-interference method and system based on multimodal dynamic spectrum evolution and adaptive collaborative compensation, to solve the problems mentioned in the background art above:

[0005] The present invention proposes an anti-interference method for capacitive touchscreens based on multimodal dynamic spectrum evolution and adaptive collaborative compensation, the method comprising:

[0006] S1. Divide the capacitive touchscreen's working environment into three-dimensional sensing areas (environment, equipment, and signals) to generate multimodal sensing area data; deploy multiple sensors based on the multimodal sensing area data to construct a multimodal sensing network;

[0007] S2. Based on the multimodal sensing network, intelligent trigger learning conditions are determined, and the learning process is started; during the learning process, wide-band multi-dimensional interference signals are collected to obtain the original interference signal dataset;

[0008] S3. Perform outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; based on the optimized interference signal data, construct a five-dimensional dynamic interference feature map set;

[0009] S4. During the touch detection phase, current environmental parameters, device status parameters, and touch area parameters are collected in real time. Intelligent spectrum matching is performed using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature spectrum. If no matching spectrum is found, a temporary spectrum is generated based on adjacent frequency band spectra and scene sub-spectrums.

[0010] S5. Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculation on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized to obtain optimized touch signal data.

[0011] S6. Based on optimized touch signal data, perform touch point coordinate fitting and error correction to generate high-precision touch point data; output touch point confidence and adjust the terminal system response strategy according to the confidence; add a system status monitoring mechanism to monitor performance indicators in real time.

[0012] The present invention proposes a system for implementing the capacitive touchscreen anti-interference method based on multimodal dynamic spectrum evolution and adaptive collaborative compensation as described above, the system comprising:

[0013] Network construction module: Divides the capacitive touchscreen's working environment into three-dimensional sensing areas (environment, devices, and signals) to generate multimodal sensing area data; deploys multiple sensors based on the multimodal sensing area data to construct a multimodal sensing network;

[0014] Condition judgment module: Based on the multimodal perception network, intelligent trigger learning conditions are determined to start the learning process; during the learning process, wide-band multi-dimensional interference signals are collected to obtain the original interference signal dataset;

[0015] Data reconstruction module: Performs outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; based on the optimized interference signal data, constructs a five-dimensional dynamic interference feature map set;

[0016] The map matching module: During the touch detection phase, it collects current environmental parameters, device status parameters, and touch area parameters in real time. It performs intelligent map matching using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature map. If there is no matching map, it generates a temporary map based on adjacent frequency band maps and scene sub-maps.

[0017] Optimization processing module: Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculation on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized to obtain optimized touch signal data;

[0018] Status monitoring module: Based on optimized touch signal data, it performs touch point coordinate fitting and error correction to generate high-precision touch point data; outputs touch point confidence and adjusts the terminal system response strategy according to the confidence; adds a system status monitoring mechanism to monitor performance indicators in real time.

[0019] The beneficial effects of this invention are as follows: Through multimodal dynamic graph evolution and adaptive collaborative compensation technology, it can accurately identify and dynamically adapt to various interference sources in complex electromagnetic environments, significantly improving the interference perception capability and compensation accuracy of capacitive touchscreens, with an interference suppression rate exceeding 98%. Simultaneously, this technology reduces the compensation delay of traditional passive filtering schemes, compressing the computational delay to within 1.5ms, meeting the stringent real-time requirements of automotive and medical scenarios. Furthermore, the system enhances environmental adaptability through a dynamic graph evolution mechanism, maintaining a touch accuracy rate of ≥99.8% within a temperature range of -50℃ to 120℃ and a wide voltage input range of 8V to 36V. This method reduces the need for frequent hardware upgrades, adapting to multiple mainstream controllers through firmware updates, thus reducing upgrade costs for manufacturers. At the same time, it avoids touch drift and accidental touches caused by sudden interference changes, significantly improving the user experience. It meets the smooth interaction requirements of consumer electronics while also addressing the challenges of high-reliability scenarios such as industrial explosion-proof and medical privacy protection. Attached Figure Description

[0020] Figure 1 This is a diagram illustrating the steps of the method described in this invention;

[0021] Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0022] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0023] One embodiment of the present invention, such as Figure 1 As shown, a capacitive touchscreen anti-interference method based on multimodal dynamic spectrum evolution and adaptive collaborative compensation is described, the method comprising:

[0024] S1. Divide the working environment of the capacitive touch screen into three-dimensional sensing areas of environment, equipment and signals, and generate multimodal sensing area data; deploy multiple sensors based on the multimodal sensing area data, including motion sensing sensors, environmental sensing sensors and equipment status sensing sensors, and construct a multimodal sensing network.

[0025] S2. Based on the multimodal sensing network, determine the intelligent trigger learning conditions. When the basic trigger conditions and at least four multimodal verification conditions are met, or when an event trigger / user trigger command is received, start the learning process. In the learning process, collect wide-band (50kHz–5MHz) multi-dimensional (amplitude, phase, frequency response, time-domain waveform) interference signals to obtain the original interference signal dataset.

[0026] S3. Perform outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; based on the optimized interference signal data, construct a five-dimensional dynamic interference feature map set, which covers the 50kHz–5MHz frequency band, and the five dimensions include amplitude map, phase map, frequency response map, time-domain waveform map, and scene-specific sub-map.

[0027] S4. During the touch detection phase, current environmental parameters, device status parameters, and touch area parameters are collected in real time. Intelligent spectrum matching is performed using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature spectrum. If no matching spectrum is found, a temporary spectrum is generated based on adjacent frequency band spectra and scene sub-spectrums.

[0028] S5. Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculation on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized, including edge enhancement, noise suppression and threshold adjustment optimization, to obtain optimized touch signal data.

[0029] S6. Based on optimized touch signal data, perform touch point coordinate fitting and error correction processing to generate high-precision touch point data; output touch point confidence and adjust the terminal system response strategy according to the confidence; add a system status monitoring mechanism to monitor performance indicators in real time, including interference intensity level, touch accuracy, response delay and false touch rate; based on the performance indicator evaluation results, automatically trigger parameter optimization, low power consumption adjustment or fault diagnosis processing to generate self-optimized system operating status data, and realize the continuous and stable operation of capacitive touch screen in complex electromagnetic environment.

[0030] The working principle and effects of the above technical solution are as follows: The combination of multimodal perception and dynamic graph evolution significantly improves the accuracy of interference identification and wide-band interference detection capabilities, avoiding touch malfunctions caused by missed interference detection in complex electromagnetic environments. The four-dimensional collaborative compensation model significantly enhances interference suppression, reduces accidental touches and touch drift, and enhances the stability and accuracy of touch detection. The compensation calculation and graph update speed are greatly accelerated, avoiding response lag issues in real-time scenarios. Low power consumption adjustment effectively reduces energy consumption during non-working periods, meeting the battery life requirements of mobile terminals without affecting anti-interference performance. The core algorithm can be upgraded without hardware modifications, reducing manufacturer adaptation costs and enhancing the flexibility of multi-scenario customization. System status monitoring and fault diagnosis functions improve operational reliability, and encryption design strengthens information security, avoiding risks caused by equipment failure and data leakage.

[0031] In one embodiment of the present invention, S1 includes:

[0032] S11. Based on the spatial characteristics and interference propagation path of the capacitive touch screen's working environment, complete the three-dimensional perception area definition of the environment, equipment, and signals, clarify the perception focus and coverage of each area, and generate multimodal perception area data.

[0033] S12. Based on multimodal sensing area data, deploy motion sensing components, environmental sensing components and equipment status sensing components in the corresponding areas; the motion sensing components include a three-axis accelerometer and a six-axis gyroscope, the environmental sensing components include an infrared distance sensor, a temperature and humidity sensor and a miniature EMI sensor, and the equipment status sensing components involve a power management detection unit and a communication status detection unit, forming a multi-dimensional sensor layout.

[0034] S13. Connect all deployed sensing components, realize data communication between components through high-speed data transmission links, integrate sensing resources, and build a multimodal sensing network.

[0035] The working principle and effects of the above technical solution are as follows: The rational definition of the three-dimensional sensing area improves the targeting of sensing coverage and avoids missed interference detection due to ambiguous area division. The precise deployment of multi-dimensional sensors expands the sensing range and dimensions, enhances the ability to capture new interference sources and complex environmental parameters, and reduces sensing blind spots. High-speed data links enable efficient data exchange between components, integrating dispersed sensing resources. This ensures both the real-time nature of sensing data and the accuracy of data fusion, avoiding misjudgments of interference caused by data transmission delays or fragmentation. The construction of the overall sensing network allows the system to more comprehensively capture changes in the environment, equipment, and signals, providing reliable support for subsequent intelligent learning triggers and accurate interference identification, further enhancing the sensing adaptability in complex environments.

[0036] In one embodiment of the present invention, S2 includes:

[0037] S21. Set basic trigger rules, multimodal verification rules, event trigger rules, and user trigger rules; basic trigger rules are related to touch signal status, capacitance baseline fluctuation, and leakage current; multimodal verification rules involve motion status, environmental parameters, and device operating parameters; event trigger rules cover interference source switching, device mode changes, and touch error rate changes; user trigger rules include scenario-based calibration options and scan parameter adaptation logic.

[0038] S22. The multimodal perception network collects relevant data in real time and checks whether each condition is met by comparing it with the preset triggering rules. If the basic triggering rules are met and at least four multimodal verification rules meet the requirements, or if an event triggering instruction or a user triggering instruction is received, it is determined that the learning start-up conditions are met.

[0039] S23. After the triggering conditions are met, the learning process is started, the wideband signal acquisition module is activated synchronously, and the acquisition frequency range and data dimensions are determined.

[0040] S24. According to the set wide frequency range, cover five frequency bands: ultra-low frequency, low frequency, medium frequency, high frequency, and ultra-high frequency. Select characteristic frequency points in each frequency band, collect amplitude, phase, frequency response, and time-domain waveform data of the capacitor channel and mutual capacitance channel, and compile them to form the original interference signal dataset.

[0041] The working principle and effects of the above technical solution are as follows: Layered triggering rules precisely control the learning initiation timing, avoiding erroneous learning under interference or touch conditions, and reducing wasted computing power due to ineffective calculations. The combination of multimodal verification and active triggering ensures the purity of the learning scenario and enables timely learning initiation when interference sources switch or device modes change, enhancing responsiveness to environmental changes. Wideband coverage of five frequency bands and acquisition of multi-dimensional data improves the comprehensiveness of interference signal capture, making the original interference signal dataset more complete and laying a solid foundation for subsequent dynamic feature mapping. Scenario-based calibration and scanning parameter adaptation enhance the targeting of learning, avoiding feature omissions caused by general acquisition methods, and making subsequent anti-interference compensation more aligned with actual scenario needs. The overall process ensures both the reliability of data acquisition and enhances the flexibility of learning, providing strong support for improving the system's anti-interference performance.

[0042] In one embodiment of the present invention, S3 includes:

[0043] S31. The 3σ+isolated forest hybrid algorithm is used to identify and remove outliers in the original interference signal dataset, eliminating the influence of electrode noise and sudden environmental interference.

[0044] S32. For missing data points after outlier removal, Kriging interpolation combined with radial basis function is used to reconstruct the data, reduce data bias, and generate optimized interference signal data.

[0045] S33. Perform wavelet packet transform on the optimized interference signal data to extract time-frequency domain features, enrich the signal feature dimensions, and provide more comprehensive feature support for spectrum construction.

[0046] S34. Based on the optimized interference signal data and the enhanced feature information, construct the amplitude spectrum, phase spectrum, frequency response spectrum, and time-domain waveform spectrum, and record the basic characteristics of interference in different frequency bands;

[0047] S35. Combining the interference characteristics of different application scenarios, including consumer electronics, automotive, industrial, and medical applications, based on the basic feature map, construct scenario-specific sub-maps for each scenario to form a five-dimensional dynamic interference feature map set; embed incremental update identifiers and aging elimination judgment criteria into the map set to lay the foundation for subsequent dynamic evolution of the map.

[0048] The working principle and effects of the above technical solution are as follows: The hybrid algorithm accurately removes outliers, reducing errors caused by electrode noise and sudden interference, and preventing erroneous data from affecting subsequent spectrum construction. Data reconstruction fills in missing points, reducing data deviation and making the optimized data more reliable. Wavelet packet transform extracts more time-frequency domain features, enriching the signal dimensions and enhancing the spectrum's ability to capture interference features. The basic spectrum and scenario-specific sub-spectrums form a comprehensive five-dimensional spectrum set, improving the matching adaptability to different scenarios and avoiding the inadequacy of general-purpose spectra in dealing with specific interference. Embedded update and obsolescence markers pave the way for dynamic adjustments to the spectrum, reducing the resource consumption of invalid spectra, ensuring the timeliness of the spectrum, reducing computational power consumption, and making subsequent anti-interference compensation more closely aligned with actual interference conditions.

[0049] In one embodiment of the present invention, S32 includes:

[0050] A full-domain traversal is performed on the original interference signal data after outlier removal to locate the frequency band positions and coordinate information of missing data, and integrate them to form the distribution information of missing data points;

[0051] Based on the distribution information of missing data points, the amplitude, phase, frequency response and time-domain waveform features of adjacent valid data points are extracted. The preliminary data of the missing points are calculated by the Kriging interpolation algorithm to complete the preliminary filling of the missing data and generate preliminary reconstructed data.

[0052] By comparing the feature correlation between the preliminary reconstructed data and the surrounding valid data, the bias data is adjusted by radial basis function to optimize the continuity and consistency of the data and generate refined reconstructed data;

[0053] The deviation between the finely reconstructed data and the original valid data is calculated. The reconstructed data with deviations within the allowable range is retained, the out-of-tolerance data is removed, and the interpolation and correction process is re-executed to finally generate optimized interference signal data.

[0054] The working principle and effects of the above technical solution are as follows: Full-domain traversal accurately locates the frequency bands and coordinates of missing data, avoiding reconstruction deviations caused by ambiguous point positioning. Multi-dimensional features of surrounding valid data are extracted, and initial filling is completed through interpolation, providing reasonable support for the missing data area. Radial basis functions optimize data correlation, reducing data fragmentation after reconstruction and enhancing overall data consistency. Deviation verification filters qualified data, removes out-of-tolerance portions, and re-executes the processing flow, reducing the risk of data distortion. The entire process efficiently fills data gaps while strictly controlling reconstruction accuracy, making the final optimized interference signal data more reliable and preventing data incompleteness or inaccuracy from affecting the quality of subsequent interference feature map construction, laying a solid foundation for subsequent accurate anti-interference compensation.

[0055] In one embodiment of the present invention, step S4 includes:

[0056] S41. During the touch detection phase, the temperature and humidity of the current environment, electromagnetic radiation intensity, power supply voltage and communication status of the device, as well as the real-time operating parameters of the touch area are collected synchronously through a multimodal sensing network. The real-time operating parameters include location information, etc.

[0057] S42. Input the real-time running parameters into the weighted cosine similarity and scene priority algorithm, call the five-dimensional dynamic interference feature map set for matching operation, and quickly filter the interference feature map with the highest degree of fit with the current state.

[0058] S43. Check the matching operation results. If there is an optimal interference feature map with a matching degree that meets the standard, directly determine the map as the basis for compensation; if no matching map is found, start the temporary map generation process.

[0059] S44. Retrieve adjacent frequency band spectra and corresponding scene sub-spectrums from the spectra set, and perform data fusion and feature completion through a lightweight generative adversarial network model to generate a temporary spectra.

[0060] The working principle and effects of the above technical solution are as follows: Real-time acquisition of multi-dimensional operating parameters provides comprehensive data support for spectrum matching, improving the accuracy of the matching results. The combination of weighted cosine similarity and scene priority algorithms accelerates the matching speed, reduces invalid calculations, and avoids compensation lag caused by the long processing time of traditional matching. The optimal spectrum with satisfactory fit is prioritized to ensure targeted compensation; when no matching spectrum is available, a temporary spectrum is quickly generated to avoid anti-interference interruptions due to missing spectrums. The temporary spectrum is completed by fusing adjacent frequency bands with scene sub-spectrums, which not only conforms to the current interference characteristics but also maintains the continuity of compensation, enhancing the system's adaptability to complex and changing interference environments and laying a solid foundation for subsequent accurate compensation.

[0061] In one embodiment of the present invention, S42 includes:

[0062] The collected real-time operating parameters, such as temperature and humidity, electromagnetic radiation intensity, power supply voltage, communication status, and touch area location, are classified and integrated, and the data is sorted and arranged in a preset format to generate a normalized parameter set.

[0063] Based on the degree of influence of different parameters on interference feature matching, corresponding weight coefficients are assigned to each parameter in the normalized parameter set to form a weight configuration table, thereby strengthening the influence of key parameters on the matching results; the data reading function of the intelligent storage module is activated to retrieve the stored five-dimensional dynamic interference feature map set, which includes a basic map and various scene-specific sub-maps, providing data support for the matching operation;

[0064] Substitute the normalized parameter set and weight configuration table into the weighted cosine similarity algorithm, and compare them one by one with the map data in the five-dimensional dynamic interference feature map set. Calculate the degree of fit between each map and the current state to generate a preliminary matching result set.

[0065] Based on the current application scenario of the device, the priority of the corresponding scenario sub-map is set, and the preliminary matching result set is sorted by the fit value and scenario priority to generate a sorted matching list.

[0066] Extract the graph with the highest matching score from the sorted matching list, and use it as the interference feature graph with the highest matching score to the current state to complete the filtering process.

[0067] The working principle and effects of the above technical solution are as follows: After parameter classification and integration, the parameters are sorted according to a preset format, and then weights are allocated according to their degree of influence to strengthen the role of key parameters, improve matching targeting, and avoid the deviation in matching degree caused by interference from irrelevant parameters. A five-dimensional dynamic atlas set containing the basic atlas and sub-atlases of each scene is retrieved to provide comprehensive data support for matching operations and reduce matching errors caused by incomplete data. The weighted cosine similarity algorithm is used to map the atlas data one by one to accurately calculate the matching degree of each atlas and generate reliable preliminary matching results. The sub-atlases are prioritized according to the current application scenario, and the preliminary results are comprehensively sorted to avoid the best atlas being ignored due to general sorting. Finally, the atlas with the highest matching degree is extracted to provide accurate basis for subsequent anti-interference compensation, enhance the effectiveness of compensation, make touch detection more stable in complex environments, and reduce touch abnormalities caused by inaccurate matching.

[0068] In one embodiment of the present invention, step S5 includes:

[0069] S51. Based on the current interference intensity, temperature and humidity conditions, frequency characteristics and time domain characteristics, adjust the dynamic weights corresponding to amplitude, phase, frequency and time domain, as well as the temperature and humidity compensation coefficient, frequency adaptation coefficient, voltage compensation coefficient and time domain characteristic coefficient.

[0070] S52. Substitute the feature data of the optimal interference feature map or temporary map, along with the determined compensation parameters, into the amplitude, phase, frequency, and time-domain four-dimensional collaborative compensation model to calculate the real-time capacitance data, offset the interference effect, and generate the compensated capacitance data.

[0071] S53. The dynamic radius Laplacian operator is used to adjust the enhancement radius according to the interference intensity of the edge region, thereby improving the signal strength of the edge region in the compensated capacitance data and improving the edge touch detection effect.

[0072] S54. Enable adaptive bilateral filtering, switch the filter core according to the type of residual noise, and retain the detailed information of the touch signal while effectively suppressing noise;

[0073] S55. Referencing the dynamic range map in the dynamic interference feature map set and the current real-time interference intensity, adjust the touch detection threshold to ensure that the threshold is adapted to the current interference environment and generate optimized touch signal data.

[0074] The working principle and effects of the above technical solution are as follows: Dynamic weights and multi-dimensional compensation coefficients are adjusted as needed, making the compensation more suitable for the current interference, temperature, humidity, and frequency characteristics, improving the specificity of the compensation, and avoiding incomplete interference cancellation caused by general parameters. The four-dimensional collaborative compensation model accurately responds to multi-source interference, significantly improving the interference suppression effect and reducing interference residues in the touch signal. Edge enhancement flexibly adjusts the radius according to the interference intensity, improving the problem of weak signals in edge areas and reducing edge mis-touch and touch drift. Adaptive bilateral filtering switches the filter core according to the noise type, which can effectively suppress residual noise while completely preserving touch details and enhancing signal purity; the dynamic threshold is adjusted in real time according to the interference intensity, avoiding the situation where fixed thresholds are not well adapted to the environment, making touch signal recognition more accurate. The entire process makes the final generated optimized touch signal data more reliable, providing strong support for subsequent high-precision touch recognition, and making touch operation more stable and smooth in complex environments.

[0075] In one embodiment of the present invention, S53 includes:

[0076] The compensated capacitance data is scanned across the entire domain to identify the coordinate range corresponding to the edge of the touch screen, and the capacitance signal data of the edge region is separated to generate a subset of edge region data.

[0077] The signal fluctuation characteristics of the edge region data subset are extracted, and combined with the electromagnetic radiation intensity data collected by the multimodal sensing network, the current interference intensity of the edge region is analyzed to generate a quantitative value of the edge interference intensity.

[0078] Set the range of Laplacian operator enhancement radius corresponding to different interference intensity quantization values. The higher the interference intensity, the larger the enhancement radius, forming an adaptation table of interference intensity and enhancement radius. Compare the edge interference intensity quantization value with the adaptation table to determine the target enhancement radius corresponding to the current interference intensity, ensuring that the enhancement strength matches the interference level.

[0079] The dynamic radius Laplacian operator is loaded, and the target enhancement radius and edge region data subset are input to perform signal enhancement operations, thereby improving the amplitude and discernibility of the edge region signal. The enhanced edge region data is then fused with the compensated capacitance data of the non-edge region to maintain the continuity and integrity of the data, generating edge-enhanced capacitance data.

[0080] The working principle and effects of the above technical solution are as follows: Full-area scanning accurately separates the capacitance signals in edge regions, avoiding enhancement deviations caused by the confusion between edge and non-edge data. By combining electromagnetic radiation intensity analysis with the degree of edge interference, a quantized value is generated and matched with a corresponding enhancement radius, ensuring the enhancement intensity closely matches the actual interference situation. This improves the targeting of edge signal enhancement and avoids insufficient enhancement or signal distortion caused by a fixed radius. The dynamic radius Laplacian operator effectively improves the signal amplitude and recognition in edge regions, reducing misjudgments and lack of response caused by weak edge touch signals. The enhanced data is fused with non-edge region data, maintaining overall continuity and integrity, avoiding poor signal connection between the edge and center regions. The entire process not only enhances the edge touch detection effect under complex interference but also maintains the overall stability of capacitance data, making touchscreen edge operations more sensitive and accurate, and significantly improving the user's touch experience in edge regions.

[0081] In one embodiment of the present invention, S55 includes:

[0082] The signal fluctuation parameters and frequency band characteristic data of the dynamic range spectrum are retrieved from the dynamic interference feature spectrum set, and the quantitative values ​​of the current real-time interference intensity are collected simultaneously to form a threshold adjustment reference dataset.

[0083] By comparing the signal variation range of the dynamic range spectrum in the reference dataset with the real-time interference intensity values, we can analyze the fluctuation trend of the signal dynamic range under different interference intensities and clarify the correlation between the two.

[0084] Based on the correlation patterns and the preset threshold range, and combined with the signal fluctuation characteristics of the current interference environment, an initial threshold range that can cover the interference fluctuation range is defined, providing a basis for precise adjustment;

[0085] Based on the differences in real-time interference intensity levels, the threshold size is refined and adjusted within the initial threshold range; the threshold is appropriately increased when the interference intensity increases and reasonably decreased when the interference intensity decreases, to determine the target threshold that is suitable for the current environment.

[0086] The target threshold is applied to the capacitance data after edge enhancement and noise suppression to detect the effective recognition rate and false touch rate of the touch signal, and to determine whether the threshold can accurately distinguish between the touch signal and the interference signal.

[0087] If the threshold adaptation effect meets the requirements, the target threshold is directly fused with the processed capacitance data to generate optimized touch signal data; if the adaptation effect is not good, the threshold is readjusted until the adaptation requirements are met.

[0088] The working principle and effects of the above technical solution are as follows: It integrates dynamic range spectrum data with real-time interference intensity quantification values, clarifying the correlation between the two to provide a solid basis for threshold adjustment and avoid adaptation deviations caused by blind setting. The initial threshold range covers the interference fluctuation range, and is then refined according to the interference intensity level to improve the fit between the threshold and the current environment, reducing the possibility of missed touch signals due to excessively high thresholds or misjudged interference due to excessively low thresholds. The adaptation effect detection and fine-tuning process ensures that the target threshold can accurately distinguish between touch signals and interference signals, enhancing the reliability of touch recognition. The entire process can flexibly respond to changes in interference intensity while maintaining the stability of touch signal recognition, avoiding the problem of fixed thresholds being difficult to adapt to complex interference environments. This makes the optimized touch signal data more accurate, providing strong support for subsequent high-precision touch operations.

[0089] In one embodiment of the present invention, step S6 includes:

[0090] S61. Input the optimized touch signal data into the lightweight neural network to accurately fit the touch point coordinates, improve the coordinate positioning accuracy, and generate preliminary touch point data.

[0091] S62. Combining historical touch data and scene-based sub-maps, the initial touch point data is corrected a second time to generate high-precision touch point data, including misidentified points and multi-point mis-touches.

[0092] S63. Calculate and output the confidence level of each touch point. The terminal system executes different response strategies based on the confidence level differences to ensure the accuracy and smoothness of touch interaction. Activate the newly added system status monitoring mechanism to monitor key performance indicators in real time. The key performance indicators include interference intensity level, touch accuracy, response latency, and false touch rate.

[0093] S64. Regularly conduct a comprehensive evaluation of the monitored performance indicators to determine whether each indicator meets the preset standards;

[0094] S65. Based on the performance evaluation results, if the indicators do not meet the standards, parameter optimization is automatically triggered; if it is a non-touch period, low power consumption adjustment is performed; if an abnormal situation is detected, fault diagnosis is initiated, and self-optimized system operating status data is generated.

[0095] The working principle and effects of the above technical solution are as follows: A lightweight neural network enables more accurate touch point coordinate fitting, reducing positioning deviations. By combining historical touch data with scene sub-maps, it corrects issues such as misidentified points and multi-touch errors, generating more reliable high-precision touch point data. Confidence output allows the terminal to execute response strategies based on differences, ensuring both the accuracy of touch interaction and the smoothness of operation. System status monitoring captures key performance indicators in real time, and after periodic comprehensive evaluation, automatically triggers parameter optimization, low-power adjustments during non-touch periods, or abnormal fault diagnosis, preventing performance degradation, excessive energy consumption, or escalation of faults. This entire process significantly improves the accuracy of touch operation and user experience, enhances system adaptability, and allows capacitive touchscreens to operate stably and continuously in complex environments, reducing subsequent maintenance costs.

[0096] One embodiment of the present invention, such as Figure 2 As shown, a system for implementing the capacitive touchscreen anti-interference method based on multimodal dynamic graph evolution and adaptive collaborative compensation as described above is provided, the system comprising:

[0097] Network construction module: Divides the capacitive touchscreen's working environment into three-dimensional sensing areas (environment, devices, and signals) to generate multimodal sensing area data; deploys multiple sensors based on the multimodal sensing area data, including motion sensors, environmental sensors, and device status sensors, to construct a multimodal sensing network;

[0098] Condition Judgment Module: Based on the multimodal sensing network, intelligent trigger learning conditions are determined. When the basic trigger conditions and at least four multimodal verification conditions are met, or when an event trigger / user trigger command is received, the learning process is started. During the learning process, wide-band (50kHz–5MHz) multi-dimensional (amplitude, phase, frequency response, time-domain waveform) interference signals are collected to obtain the original interference signal dataset.

[0099] Data reconstruction module: Performs outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; Based on the optimized interference signal data, constructs a five-dimensional dynamic interference feature map set, which covers the 50kHz–5MHz frequency band. The five dimensions include amplitude map, phase map, frequency response map, time-domain waveform map, and scene-specific sub-map.

[0100] The map matching module: During the touch detection phase, it collects current environmental parameters, device status parameters, and touch area parameters in real time. It performs intelligent map matching using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature map. If there is no matching map, it generates a temporary map based on adjacent frequency band maps and scene sub-maps.

[0101] The optimization processing module: Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculations on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized, including edge enhancement, noise suppression and threshold adjustment optimization, to obtain optimized touch signal data;

[0102] Status monitoring module: Based on optimized touch signal data, it performs touch point coordinate fitting and error correction to generate high-precision touch point data; outputs touch point confidence and adjusts the terminal system response strategy according to the confidence; adds a system status monitoring mechanism to monitor performance indicators in real time, including interference intensity level, touch accuracy, response latency, and false touch rate; based on the performance indicator evaluation results, it automatically triggers parameter optimization, low power consumption adjustment, or fault diagnosis processing to generate self-optimized system operating status data, enabling the capacitive touch screen to operate continuously and stably in complex electromagnetic environments.

[0103] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A capacitive touchscreen anti-interference method based on multimodal dynamic graph evolution and adaptive collaborative compensation, characterized in that, The method includes: S1. Divide the capacitive touchscreen's working environment into three-dimensional sensing areas (environment, equipment, and signals) to generate multimodal sensing area data; deploy multiple sensors based on the multimodal sensing area data to construct a multimodal sensing network; S2. Based on the multimodal sensing network, intelligent trigger learning conditions are determined, and the learning process is started; during the learning process, wide-band multi-dimensional interference signals are collected to obtain the original interference signal dataset; S3. Perform outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; based on the optimized interference signal data, construct a five-dimensional dynamic interference feature map set; S4. During the touch detection phase, current environmental parameters, device status parameters, and touch area parameters are collected in real time. Intelligent spectrum matching is performed using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature spectrum. If no matching spectrum is found, a temporary spectrum is generated based on adjacent frequency band spectra and scene sub-spectrums. S5. Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculation on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized to obtain optimized touch signal data. S6. Based on optimized touch signal data, perform touch point coordinate fitting and error correction to generate high-precision touch point data; output touch point confidence and adjust the terminal system response strategy according to the confidence; add a system status monitoring mechanism to monitor performance indicators in real time.

2. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, S1 includes: S11. Based on the spatial characteristics and interference propagation path of the capacitive touch screen's working environment, complete the three-dimensional perception area definition of the environment, equipment, and signals, clarify the perception focus and coverage of each area, and generate multimodal perception area data. S12. Based on multimodal sensing area data, deploy motion sensing components, environmental sensing components, and device status sensing components in the corresponding areas to form a multi-dimensional sensor layout; S13. Connect all deployed sensing components, realize data communication between components through high-speed data transmission links, integrate sensing resources, and build a multimodal sensing network.

3. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, The S2 includes: S21. Set basic trigger rules, multimodal verification rules, event trigger rules, and user trigger rules; S22. The multimodal perception network collects relevant data in real time and checks whether each condition is met by comparing it with the preset triggering rules. If the basic triggering rules are met and at least four multimodal verification rules meet the requirements, or if an event triggering instruction or a user triggering instruction is received, it is determined that the learning start-up conditions are met. S23. After the triggering conditions are met, the learning process is started, the wideband signal acquisition module is activated synchronously, and the acquisition frequency range and data dimensions are determined. S24. According to the set wide frequency range, cover five frequency bands: ultra-low frequency, low frequency, medium frequency, high frequency, and ultra-high frequency. Select characteristic frequency points in each frequency band, collect amplitude, phase, frequency response, and time-domain waveform data of the capacitor channel and mutual capacitance channel, and compile them to form the original interference signal dataset.

4. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, The S3 includes: S31. The 3σ+isolated forest hybrid algorithm is used to identify and remove outliers in the original interference signal dataset, eliminating the influence of electrode noise and sudden environmental interference. S32. For the missing data points after removing outliers, Kriging interpolation combined with radial basis functions is used to reconstruct the data and generate optimized interference signal data. S33. Perform wavelet packet transform on the optimized interference signal data to extract time-frequency domain features and enrich the signal feature dimensions; S34. Based on the optimized interference signal data and the enhanced feature information, construct the amplitude spectrum, phase spectrum, frequency response spectrum, and time-domain waveform spectrum, and record the basic characteristics of interference in different frequency bands; S35. Based on the basic feature map, construct scenario-specific sub-maps for each application scenario, combining the interference characteristics of different application scenarios, to form a five-dimensional dynamic interference feature map set; embed incremental update identifiers and aging and elimination judgment criteria into the map set.

5. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, The S4 includes: S41. During the touch detection phase, the temperature and humidity of the current environment, the intensity of electromagnetic radiation, the power supply voltage and communication status of the device, and the real-time operating parameters of the touch area are collected synchronously through a multimodal sensing network. S42. Input the real-time running parameters into the weighted cosine similarity and scene priority algorithm, call the five-dimensional dynamic interference feature map set for matching operation, and filter the interference feature map with the highest degree of fit with the current state. S43. Check the matching operation results. If there is an optimal interference feature map with a matching degree that meets the standard, directly determine the map as the basis for compensation; if no matching map is found, start the temporary map generation process. S44. Retrieve adjacent frequency band spectra and corresponding scene sub-spectrums from the spectra set, and perform data fusion and feature completion through a lightweight generative adversarial network model to generate a temporary spectra.

6. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 5, characterized in that, S42 includes: The collected real-time operating parameters are classified and integrated, and the data is sorted and arranged according to a preset format to generate a normalized parameter set. Based on the degree of influence of different parameters on interference feature matching, corresponding weight coefficients are assigned to each parameter in the normalized parameter set to form a weight configuration table. The data reading function of the intelligent storage module is then activated to retrieve the stored five-dimensional dynamic interference feature map set. Substitute the normalized parameter set and weight configuration table into the weighted cosine similarity algorithm, and compare them one by one with the map data in the five-dimensional dynamic interference feature map set. Calculate the degree of fit between each map and the current state to generate a preliminary matching result set. Based on the current application scenario of the device, the priority of the corresponding scenario sub-map is set, and the preliminary matching result set is sorted by the fit value and scenario priority to generate a sorted matching list. Extract the graph with the highest matching score from the sorted matching list, and use it as the interference feature graph with the highest matching score to the current state to complete the filtering process.

7. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, The S5 includes: S51. Based on the current interference intensity, temperature and humidity conditions, frequency characteristics and time domain characteristics, adjust the dynamic weights corresponding to amplitude, phase, frequency and time domain, as well as the temperature and humidity compensation coefficient, frequency adaptation coefficient, voltage compensation coefficient and time domain characteristic coefficient. S52. Substitute the feature data of the optimal interference feature map or temporary map, along with the determined compensation parameters, into the amplitude, phase, frequency, and time-domain four-dimensional collaborative compensation model to calculate the real-time capacitance data and generate the compensated capacitance data. S53. Employ the dynamic radius Laplacian operator to adjust the enhancement radius according to the interference intensity in the edge region; S54. Enable adaptive bilateral filtering, switch the filter core according to the type of residual noise, and retain the detailed information of the touch signal while effectively suppressing noise; S55. Referencing the dynamic range map in the dynamic interference feature map set and the current real-time interference intensity, adjust the touch detection threshold to generate optimized touch signal data.

8. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 7, characterized in that, The S55 includes: The signal fluctuation parameters and frequency band characteristic data of the dynamic range spectrum are retrieved from the dynamic interference feature spectrum set, and the quantitative values ​​of the current real-time interference intensity are collected simultaneously to form a threshold adjustment reference dataset. By comparing the signal variation range of the dynamic range spectrum in the reference dataset with the real-time interference intensity values, we can analyze the fluctuation trend of the signal dynamic range under different interference intensities and clarify the correlation between the two. Based on the correlation patterns and the preset threshold range, and combined with the signal fluctuation characteristics of the current interference environment, an initial threshold range that can cover the interference fluctuation range is defined, providing a basis for precise adjustment; Based on the differences in real-time interference intensity levels, the threshold size is refined and adjusted within the initial threshold range; the threshold is appropriately increased when the interference intensity increases and reasonably decreased when the interference intensity decreases, to determine the target threshold that is suitable for the current environment. The target threshold is applied to the capacitance data after edge enhancement and noise suppression to detect the effective recognition rate and false touch rate of the touch signal, and to determine whether the threshold can accurately distinguish between the touch signal and the interference signal. If the threshold adaptation effect meets the requirements, the target threshold is directly fused with the processed capacitance data to generate optimized touch signal data; if the adaptation effect is not good, the threshold is readjusted until the adaptation requirements are met.

9. The anti-interference method for capacitive touchscreens based on multimodal dynamic graph evolution and adaptive collaborative compensation according to claim 1, characterized in that, The S6 includes: S61. Input the optimized touch signal data into a lightweight neural network to accurately fit the touch point coordinates and generate preliminary touch point data. S62. Combining historical touch data with scene-based sub-maps, the initial touch point data is corrected a second time to generate high-precision touch point data. S63. Calculate and output the confidence level of each touch point. The terminal system executes different response strategies based on the difference in confidence level, activates the newly added system status monitoring mechanism, and monitors key performance indicators in real time. S64. Regularly conduct a comprehensive evaluation of the monitored performance indicators to determine whether each indicator meets the preset standards; S65. Based on the performance evaluation results, if the indicators do not meet the standards, parameter optimization is automatically triggered; if it is a non-touch period, low power consumption adjustment is performed; if an abnormal situation is detected, fault diagnosis is initiated, and self-optimized system operating status data is generated.

10. A system for implementing the anti-interference method for a capacitive touchscreen based on multimodal dynamic graph evolution and adaptive collaborative compensation as described in claim 1, characterized in that, The system includes: Network construction module: Divides the capacitive touchscreen's working environment into three-dimensional sensing areas (environment, devices, and signals) to generate multimodal sensing area data; deploys multiple sensors based on the multimodal sensing area data to construct a multimodal sensing network; Condition judgment module: Based on the multimodal perception network, intelligent trigger learning conditions are determined to start the learning process; during the learning process, wide-band multi-dimensional interference signals are collected to obtain the original interference signal dataset; Data reconstruction module: Performs outlier removal and data reconstruction on the original interference signal dataset to generate optimized interference signal data; based on the optimized interference signal data, constructs a five-dimensional dynamic interference feature map set; The map matching module: During the touch detection phase, it collects current environmental parameters, device status parameters, and touch area parameters in real time. It performs intelligent map matching using weighted cosine similarity and scene priority algorithms to locate the optimal interference feature map. If there is no matching map, it generates a temporary map based on adjacent frequency band maps and scene sub-maps. Optimization processing module: Based on the matched optimal interference feature map or temporary map, a four-dimensional collaborative compensation model of amplitude, phase, frequency and time domain is used to perform multi-dimensional collaborative compensation calculation on the real-time capacitance data with dynamic weight adjustment to generate compensated capacitance data; the compensated capacitance data is then optimized to obtain optimized touch signal data; Status monitoring module: Based on optimized touch signal data, it performs touch point coordinate fitting and error correction to generate high-precision touch point data; outputs touch point confidence and adjusts the terminal system response strategy according to the confidence; adds a system status monitoring mechanism to monitor performance indicators in real time.

Images