Self-made glass sensor capacitance grid layout optimization method and device

Abstract

The invention relates to the technical field of touch display screens, in particular to a self-made glass sensor capacitance grid layout optimization method and device, and the method comprises the steps: carrying out the regional division of a capacitance grid layout, and constructing a perception feature database; basic capacitance response, coordinate positioning precision and pressure rate sensing threshold values of all the areas are determined based on the database, the influence of grid parameter changes on pressure sensing precision is analyzed, and parameter sensitivity is quantified; acquiring capacitance, coordinates and pressure rate data of each area in real time, and generating a grid parameter adjustment table when the capacitance, coordinates and pressure rate data exceed a threshold value; key parameters are extracted from the adjustment table, the perception importance degree is calculated, and a to-be-optimized parameter set is generated; and layering the region according to the parameter sensitivity, constructing a mapping relationship between a parameter adjustment amount and a performance improvement amount through a regression tree, generating a progressive adjustment rule, optimizing a parameter set, and driving layout parameterization reconstruction. According to the method, the multi-touch-point pressure sensing precision and the system robustness are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of touch display screen, and particularly relates to a self-made glass sensor capacitive grid layout optimization method and device. BACKGROUND

[0002] With the rapid development of flexible display and wearable electronic devices, the performance requirements of touch screen as the core component of human-computer interaction are increasingly stringent. The traditional capacitive touch screen mainly relies on accurate identification of touch point position, however, in advanced interaction scenarios (such as pressure drawing, game control, etc.), only point coordinate information cannot meet the demand, and fine perception of touch pressure and its change rate becomes crucial. Especially in the face of complex situations of multiple touch points and each point applying different pressure and change rate, how to accurately, quickly and independently perceive the pressure dynamics of each point becomes a key technical bottleneck to improve user experience.

[0003] The above content is only used to assist in understanding the technical solutions of the present application, and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0004] The main purpose of the present application is to provide a self-made glass sensor capacitive grid layout optimization method and device, aiming to solve the technical problem of how to accurately, quickly and independently perceive the pressure dynamics of each point in the face of complex situations of multiple touch points and each point applying different pressure and change rate.

[0005] To achieve the above purpose, the present application provides a self-made glass sensor capacitive grid layout optimization method, which comprises: The layout structure of the self-made glass sensor capacitive grid is regionally divided by a multi-dimensional perception analysis method to obtain a multi-perception region layout, and historical capacitance change data, historical touch point coordinate data and corresponding historical pressure change rate data of each perception region layout under multiple pressure loads are collected to construct a perception feature database; The basic capacitance response threshold, coordinate positioning accuracy threshold and pressure change rate perception threshold of each perception region layout are determined according to the perception feature database, and the influence of grid parameter change in each perception region layout on multi-touch point pressure perception accuracy is analyzed based on the perception feature database to obtain the grid parameter sensitivity of each perception region layout; Real-time capacitance change data, real-time touch point coordinate data and real-time pressure change rate data of each perception region layout are obtained, and a grid parameter adjustment table of each perception region layout is generated when the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold or pressure change rate perception threshold is exceeded; analyzing multiple correlations between the capacitance response, coordinate positioning, and pressure rate sensing of each of the sensing area layout based on the sensing feature database, and extracting grid parameters from the grid parameter adjustment table according to the multiple correlations to perform sensing importance calculation, to generate a set of grid parameters to be optimized for each of the sensing area layout; sensitivity grouping is performed on each of the sensing area layout according to the grid parameter sensitivity, to obtain a multi-level sensitivity layout area, so as to construct a mapping relationship between the grid parameter adjustment amount and the multi-touch point pressure sensing performance improvement amount of each level of the sensitivity layout area by using a regression tree, to generate a gradual layout adjustment rule to optimize the set of grid parameters to be optimized for each of the sensing area layout, and to control parameterization reconstruction of the layout of the self-made glass sensor capacitance grid based on the optimization result.

[0006] Optionally, the area division of the layout structure of the self-made glass sensor capacitance grid by using the multi-dimensional sensing analysis method to obtain a multi-sensing area layout comprises: obtaining the electrode density, unit capacitance base value, mutual coupling coefficient, and signal-to-noise ratio of each sub-area in the capacitance grid layout, to construct a layout complexity quantization matrix of each of the sub-areas; a sensing directed graph is established based on each of the layout complexity quantization matrix, and clustering operation is performed on the mean value and peak value of the capacitance change amplitude of each of the sub-areas under pressure load, to preliminarily divide each of the sub-areas, to obtain a capacitance response grouping result; corresponding touch point coordinate offset data of each of the sub-areas is obtained, a preset positioning tolerance threshold is compared with the coordinate offset data, and the capacitance response grouping result is refined and adjusted according to the comparison result, to obtain a multi-sensing area layout.

[0007] Optionally, the basic capacitance response threshold comprises a baseline capacitance value, a lower limit value of a dynamic range, and an upper limit value of saturation; the coordinate positioning accuracy threshold comprises a minimum resolvable distance and a linearity error limit; and the pressure change rate sensing threshold comprises a minimum perceptible rate gradient and a maximum following rate.

[0008] Optionally, the determination of the basic capacitance response threshold, the coordinate positioning accuracy threshold, and the pressure change rate sensing threshold of each of the sensing area layout according to the sensing feature database comprises: baseline capacitance sampling values of each of the sensing area layout in a non-touch state, capacitance saturation values and non-linear distortion points under different pressures are extracted from the sensing feature database, and a static response curve of capacitance-pressure is fitted by using a Gaussian process regression, to determine the baseline capacitance value, the lower limit value of the dynamic range, and the upper limit value of saturation of each of the sensing area layout; extracting multiple coordinate sampling data of each of the perception area patterns on standard test points from the perception feature database, calculating root mean square error and linear fitting residual of the data and theoretical coordinates, and determining minimum resolvable distance and linearity error limit of each of the perception area patterns through confidence interval estimation; extracting capacitance response delay data and rate tracking error of each of the perception area patterns under a preset pressure ramp signal from the perception feature database, smoothing measurement noise and identifying system response bandwidth by using Kalman filter, so as to determine minimum perceivable rate gradient and maximum following rate of each of the perception area patterns.

[0009] Optionally, the influence of grid parameter change on multi-touch point pressure perception accuracy in each of the perception area patterns is analyzed based on the perception feature database, so as to obtain grid parameter sensitivity of each of the perception area patterns, including: extracting historical touch point coordinate data, historical pressure change rate data and corresponding historical grid parameter change records of each of the perception area patterns from the perception feature database, so as to train a time series convolution network by using the training samples, and obtain a response model representing dynamic correlation between grid parameter change and multi-dimensional perception performance in each of the perception area patterns; based on the response model, coordinate positioning error change and pressure rate tracking error change caused by specific grid parameter change in each of the perception area patterns under a preset pressure mode sequence are counted, and the error change is subjected to Spearman rank correlation analysis according to pressure change rate interval, so as to obtain parameter-performance correlation strength index of each of the perception area patterns; a preset function weight of each of the perception area patterns in a global perception task is used to weight and fuse the corresponding parameter-performance correlation strength index, and the fusion result is divided into intervals according to a preset sensitivity classification threshold, so as to obtain a grid parameter sensitivity classification table to quantitatively determine the grid parameter sensitivity of each of the perception area patterns.

[0010] Optionally, the multiple correlations among capacitance response, coordinate positioning and pressure rate perception of each of the perception area patterns are analyzed based on the perception feature database, and grid parameters are extracted from the grid parameter adjustment table for perception importance calculation according to the multiple correlations, so as to generate a set of to-be-optimized grid parameters of each of the perception area patterns, including: extracting historical capacitance response sequence, historical coordinate positioning data and historical perception performance evaluation data of each of the perception area patterns under different pressure change rates from the perception feature database, and calculating perception behavior similarity under different pressure scenarios by using dynamic time warping algorithm, so as to perform pressure scenario clustering on each of the perception area patterns according to similarity calculation result, and obtain a typical pressure perception scenario set; The mean, variance, extreme value and covariance matrix of the capacitance response characteristic value, coordinate stability index and rate tracking accuracy in each of the sets of typical pressure sensing scenarios are calculated to construct a multi-modal sensing feature tensor, and a partial least squares regression method is used to quantify the complex correlation between the layout grid parameters and the multi-modal sensing performance indicators, and a multi-element coupling relationship matrix of the layout of each sensing area is obtained. Based on the multi-element coupling relationship matrix, candidate adjustment parameters are extracted from the grid parameter adjustment table of each sensing area layout, and the contribution of the extracted candidate grid parameters to the overall sensing performance improvement is quantified by a random forest algorithm, so as to quantify the sensing importance of the extracted candidate grid parameters based on the contribution, and select the candidate grid parameters whose importance calculation results exceed the importance threshold to form the set of to-be-optimized grid parameters of each sensing area layout.

[0011] Optionally, the sensitivity of each grid parameter is grouped according to the sensitivity of each sensing area layout to obtain a multi-level sensitivity layout area, and a regression tree is used to construct the mapping relationship between the grid parameter adjustment amount and the multi-touch point pressure sensing performance improvement amount of each level of sensitivity layout area, and a gradual layout adjustment rule is generated to optimize the set of to-be-optimized grid parameters of each sensing area layout, and after the parameterized reconstruction of the layout of the self-made glass sensor capacitor grid based on the optimization result, it includes: The performance index time series data of the self-made glass sensor after the optimization of the layout parameters under multiple pressure change rate test modes are obtained, and the performance index time series data is used as an input feature to train a generative adversarial network, so as to generate a high-fidelity virtual pressure touch signal in real time and simulate a sensing response through the trained network, and obtain virtual sensing response data; Based on the virtual sensing response data, the parameter tolerance boundary of each sensing area layout under extreme pressure change rate is obtained as a stability constraint condition, and a virtual layout simulation environment containing four groups of virtual grid parameters of electrode width, spacing, shape factor and interlayer coverage area is constructed by parameterized scanning; Taking the weighted sum of the comprehensive accuracy of multi-touch point pressure sensing and the minimum as the optimization objective, the comprehensive accuracy includes coordinate error, pressure value error and rate tracking error, and multi-objective Bayesian optimization is used to cooperatively optimize the virtual grid parameters of each sensing area layout in the virtual layout simulation environment, and a set of Pareto optimal virtual parameter combinations of each sensing area layout is obtained; Based on each set of Pareto optimal virtual parameter combinations, high-precision electromagnetic-structure coupling simulation is performed on the capacitor grid, and when the sensing performance index data obtained by simulation is stable in the preset performance target interval, the self-made glass sensor capacitor grid is controlled according to the optimized virtual parameter combination set to perform final parameterized reconstruction before manufacturing and output.

[0012] In addition, to achieve the above object, the application further provides a self-made glass sensor capacitor grid layout optimization device, which comprises: A data acquisition module is configured to divide the layout structure of the self-made glass sensor capacitor grid into regions by a multi-dimensional perception analysis method, obtain a multi-perception region layout, and collect historical capacitance change data, historical touch point coordinate data and corresponding historical pressure change rate data of each perception region layout under multiple pressure loads to construct a perception feature database. A parameter perception module is configured to determine a basic capacitance response threshold, a coordinate positioning accuracy threshold and a pressure change rate perception threshold of each perception region layout according to the perception feature database, and analyze the influence of grid parameter changes in each perception region layout on multi-touch point pressure perception accuracy based on the perception feature database to obtain a grid parameter sensitivity of each perception region layout. An adjustment generation module is configured to obtain real-time capacitance change data, real-time touch point coordinate data and real-time pressure change rate data of each perception region layout, and generate a grid parameter adjustment table of each perception region layout when the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold or pressure change rate perception threshold is exceeded. A correlation analysis module is configured to analyze the multiple correlations between the capacitance response, coordinate positioning and pressure rate perception of each perception region layout based on the perception feature database, and perform perception importance calculation on the extracted grid parameters from the grid parameter adjustment table according to the multiple correlations to generate a set of to-be-optimized grid parameters of each perception region layout. A parameter adjustment module is configured to group each perception region layout according to the grid parameter sensitivity to obtain a multi-level sensitivity layout region, construct a mapping relationship between the grid parameter adjustment amount and the multi-touch point pressure perception performance improvement amount of each level of sensitivity layout region by a regression tree, generate a gradual layout adjustment rule to optimize the set of to-be-optimized grid parameters of each perception region layout, and control the parameterization reconstruction of the layout of the self-made glass sensor capacitor grid based on the optimization result.

[0013] In addition, to achieve the above object, the application further provides a self-made glass sensor capacitor grid layout optimization device, which comprises:

[0014] In addition, to achieve the above objectives, the present invention also provides a storage medium storing a self-made glass sensor capacitor grid layout optimization program, wherein when the self-made glass sensor capacitor grid layout optimization program is executed by a processor, it implements the steps of the self-made glass sensor capacitor grid layout optimization method as described above.

[0015] This invention provides a method for optimizing the capacitive grid layout of a self-made glass sensor. The method, through multi-dimensional region partitioning and feature database construction, subdivides the capacitive grid layout into multiple sensing regions based on parameters such as electrode density and capacitance response characteristics, overcoming the limitations of traditional unified parameter design. Combining a response model trained on historical data with multiple correlation analysis, it can accurately capture the nonlinear mapping relationship between grid parameters and pressure sensing accuracy, solving the problems of signal crosstalk and nonlinear distortion in dynamic pressure sensing at multiple touch points, reducing pressure change rate sensing errors, and improving coordinate positioning accuracy to sub-millimeter level, meeting the needs of high-precision interactive scenarios. Through real-time threshold monitoring and parameter adjustment table generation mechanisms, the system can respond to dynamic changes in touch pressure in real time. When abnormal capacitance response, coordinate positioning deviation, or pressure rate sensing failure is detected, it can automatically trigger differentiated adjustments to the grid parameters, avoiding performance degradation of traditional static layouts under complex pressure loads. By combining regression tree-based progressive adjustment rules, hierarchical optimization can be implemented for layout regions with different sensitivities. For example, fine-tuning with small steps can be used for high-sensitivity regions, while batch parameter optimization can be performed for low-sensitivity regions. This shortens the system's response latency under multi-point pressure change scenarios and significantly improves real-time interaction. Traditional layout design relies on engineers' experience to adjust mesh parameters, resulting in long trial-and-error cycles and difficulty in covering the entire scenario. This solution quantifies the impact weight of mesh parameters on multi-dimensional perception performance through data-driven sensitivity analysis and importance calculation, avoiding blind adjustments. For example, Spearman's rank correlation analysis is used to determine the optimization priority of key parameters, and Pareto optimal parameter combinations are generated by combining Bayesian optimization and other algorithms. This shortens the layout design cycle and reduces the number of iterations of physical samples, thereby reducing R&D costs. In addition, the construction of a virtual simulation environment allows for early verification of parameter robustness during the design phase, avoiding performance risks in actual manufacturing. For the strong coupling characteristics between capacitive response, coordinate positioning, and pressure rate perception, the solution decomposes the complex multi-objective optimization problem into hierarchical processing tasks through multi-level sensitivity grouping and mapping relationship modeling. For example, a high-precision regression tree model is used to finely adjust the electrode spacing in highly sensitive areas, while simplified rules are used for batch optimization in low-sensitivity areas, reducing computational complexity while ensuring overall performance. This strategy effectively suppresses signal interference when multiple pressures are superimposed, improves the system's sensing stability under extreme conditions, and significantly enhances the reliability of the device in complex interaction scenarios. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of a self-made glass sensor capacitor grid layout optimization device for the hardware operating environment involved in the embodiments of the present invention; Figure 2 This is a flowchart illustrating an embodiment of the self-made glass sensor capacitor grid layout optimization method of the present invention; Figure 3 This is a structural block diagram of an embodiment of the self-made glass sensor capacitor grid layout optimization device of the present invention.

[0017] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a self-made glass sensor capacitor grid layout optimization device for the hardware operating environment involved in the embodiments of the present invention.

[0020] like Figure 1 As shown, the self-made glass sensor capacitance grid layout optimization device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen, and optionally, it may also include a standard wired interface or a wireless interface. In this invention, the wired interface of the user interface 1003 may be a USB interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a non-volatile memory (NVM), such as a disk storage device. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0021] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the self-made glass sensor capacitor grid layout optimization device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0022] likeFigure 1 As shown, the memory 1005, which serves as a computer storage medium, may include an operating device, a network communication module, a user interface module, and a self-made glass sensor capacitor grid layout optimization program.

[0023] exist Figure 1 In the self-made glass sensor capacitor grid layout optimization device shown, the network interface 1004 is mainly used to connect to the backend server and communicate with the backend server; the user interface 1003 is mainly used to connect to peripherals; the self-made glass sensor capacitor grid layout optimization device calls the self-made glass sensor capacitor grid layout optimization program stored in the memory 1005 through the processor 1001 and executes the self-made glass sensor capacitor grid layout optimization method provided in the embodiment of the present invention.

[0024] Based on the above hardware structure, an embodiment of the self-made glass sensor capacitor grid layout optimization method of the present invention is proposed.

[0025] Reference Figure 2 , Figure 2 This is a flowchart illustrating an embodiment of the self-made glass sensor capacitor grid layout optimization method of the present invention, which presents an embodiment of the self-made glass sensor capacitor grid layout optimization method of the present invention.

[0026] In one embodiment, the method for optimizing the capacitive grid layout of the self-made glass sensor includes the following steps: Step S100: The layout structure of the self-made glass sensor capacitor grid is divided into regions using a multi-dimensional sensing analysis method to obtain a multi-sensing region layout. Historical capacitance change data, historical touch point coordinate data, and corresponding historical pressure change rate data of each sensing region layout under various pressure loads are collected to construct a sensing feature database.

[0027] The self-made glass sensor capacitive grid can be a transparent capacitive sensing structure integrated into flexible displays or wearable devices. It consists of a grid array of cross electrodes and is used to sense touch position and pressure information, enabling the identification of touch point coordinates and the sensing of pressure intensity. In an exemplary embodiment, the self-made glass sensor capacitive grid can be formed by depositing conductive materials on a glass substrate using micro-nano fabrication processes to create an electrode network, utilizing capacitance changes to detect touch behavior. Furthermore, the self-made glass sensor capacitive grid can include, but is not limited to, one or more of ITO electrode grids, silver nanowire grids, and graphene electrode grids. The multi-sensing regions can be different functional areas divided from the capacitive grid layout based on parameters such as electrode density and response characteristics, which can be used to support differentiated parameter configurations to improve overall sensing performance. In this embodiment, the multi-sensing regions can be spatially divided using a multi-dimensional sensing analysis method to distinguish the sensing capabilities of different regions. For example, the multi-sensing regions can include, but are not limited to, one or more of high-density electrode areas, medium-response sensitive areas, and edge compensation areas. Historical capacitance change data, historical touch point coordinate data, and historical pressure change rate data can collectively form the data source of the sensing feature database. This database can be a collection of historical data for each sensing area under various pressure loads, which can be used to provide data support for response model training and correlation analysis.

[0028] In one specific embodiment, the sensing feature database can be constructed by applying a standardized pressure sequence to an experimental platform and simultaneously recording response data. Region division can be achieved by spatially segmenting the capacitor grid based on dimensions such as electrode density and capacitance response characteristics, thereby enabling differentiated design of functional areas. Furthermore, region division can be achieved by classifying the geometric and electrical characteristics of the layout using clustering algorithms or rule engines, allowing the system to configure corresponding parameters according to the sensing needs of different areas. Constructing the sensing feature database can be a process of collecting and integrating historical capacitance, coordinate, and pressure rate data under multiple operating conditions, providing a reliable data foundation for subsequent analysis. Furthermore, the sensing feature database can be constructed by repeatedly applying different pressure patterns under controlled conditions and simultaneously sampling sensor output signals, thereby achieving the technical effect of providing a reliable data source for model training and analysis.

[0029] Step S200: Determine the basic capacitance response threshold, coordinate positioning accuracy threshold, and pressure change rate sensing threshold for each sensing area based on the sensing feature database. Analyze the impact of grid parameter changes on the multi-touch point pressure sensing accuracy in each sensing area based on the sensing feature database to obtain the grid parameter sensitivity of each sensing area.

[0030] The basic capacitance response threshold can serve as a benchmark value for determining whether the capacitance signal is abnormal, and can be used to assess whether the current capacitance response deviates from the normal operating range. In this embodiment, the basic capacitance response threshold can be determined based on the capacitance change distribution under normal operating conditions in the sensing feature database. The coordinate positioning accuracy threshold can be the maximum allowable positioning deviation limit, and can be used to determine the reliability of the current touch point coordinates. For example, the coordinate positioning accuracy threshold can be set based on the statistical error distribution of historical touch point coordinate data. The pressure change rate sensing threshold can be the minimum identifiable rate standard for dynamic pressure changes, and can be used to determine whether the pressure sensing function is within the effective response range. In an exemplary embodiment, the pressure change rate sensing threshold can be obtained by analyzing historical pressure change rate data to determine the system response boundary. The grid parameter sensitivity can be a quantitative indicator reflecting the degree of influence of specific grid parameter changes on the multi-touch point pressure sensing accuracy, and can be used to guide the formulation of parameter adjustment priorities. Furthermore, the grid parameter sensitivity can be derived by analyzing the relationship between parameter changes and sensing performance in the sensing feature database to quantify the sensing accuracy fluctuations caused by different parameter adjustments.

[0031] Determining thresholds can be an operation that sets critical judgment criteria for various performance indicators based on historical data distribution, enabling the system to identify anomalies. Furthermore, threshold determination can be achieved by using statistical methods such as percentiles or distribution fitting to determine a reasonable threshold range, thereby establishing an anomaly detection benchmark. Analyzing the impact of grid parameter changes on sensing accuracy can be a process of revealing the correlation between parameter adjustments and performance, the result of which is reflected in grid parameter sensitivity. This operation can be achieved through regression analysis or multivariate correlation modeling, enabling the system to identify key influencing factors.

[0032] Step S300: Obtain real-time capacitance change data, real-time touch point coordinate data, and real-time pressure change rate data for each sensing area map. When the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold, or pressure change rate sensing threshold is exceeded, generate a grid parameter adjustment table for each sensing area map.

[0033] The real-time capacitance change data can be the dynamic signal value output by the capacitance grid under the current touch state, which can be used to reflect the presence and intensity of the touch event. In this embodiment, the real-time capacitance change data can be obtained by sampling the capacitance value between electrodes in real time through the sensor circuit. The real-time touch point coordinate data can be the position information of the current touch point on the screen plane, which can be used to provide spatial input for human-computer interaction. For example, the real-time touch point coordinate data can be obtained by solving the peak position of the electrode signal using a capacitance differential algorithm. The real-time pressure change rate data can be the amount of change in touch pressure per unit time, which can be used to characterize the dynamic characteristics of user operation. In a specific embodiment, the real-time pressure change rate data is obtained by differentiating the continuous pressure sampling values. The grid parameter adjustment table can be a mapping table that records the grid parameters to be adjusted and their target values, which can be used to drive subsequent optimization execution. Furthermore, the grid parameter adjustment table can automatically generate adjustment instruction entries when the real-time data exceeds the corresponding threshold.

[0034] Generating a grid parameter adjustment table can be an action that creates a list of parameter correction tasks when real-time monitoring data exceeds preset judgment criteria, enabling the system to respond promptly to performance deviations. Furthermore, generating the grid parameter adjustment table can be achieved by setting a dynamic weighting mechanism to prioritize marking parameters in areas with significant deviations, thereby initiating an adaptive optimization process.

[0035] Step S400 involves analyzing the multiple correlations between capacitance response, coordinate positioning, and pressure rate sensing in each sensing region based on the sensing feature database. Grid parameters are then extracted from the grid parameter adjustment table based on these multiple correlations to calculate sensing importance, thereby generating a set of grid parameters to be optimized for each sensing region. The multiple correlations can represent the nonlinear coupling relationship between capacitance response, coordinate positioning, and pressure rate sensing, revealing the interdependence mechanisms among multidimensional sensing indicators. In this embodiment, multiple correlations can be obtained through statistical analysis of the dependency structure between variables in the sensing feature database. Sensing importance can be a comprehensive score measuring the contribution of a grid parameter to overall sensing performance, used to determine the priority of parameter optimization. For example, sensing importance can be obtained by weighted calculation combining multiple correlation and sensitivity analysis results. The set of grid parameters to be optimized can be a set of key grid parameters to be adjusted, used to narrow the optimization search space and improve computational efficiency. In an exemplary embodiment, the set of grid parameters to be optimized can be formed by extracting high-impact parameters from the grid parameter adjustment table after sorting by sensing importance.

[0036] Perceived importance calculation can be an operation that integrates the results of multiple correlation and sensitivity analysis to quantify the influence of parameters, allowing the optimization process to focus on parameters with the greatest potential for improvement. Furthermore, perceived importance calculation can be achieved through comprehensive evaluation using the analytic hierarchy process (AHP) or machine learning scoring models, thereby accurately selecting parameters that need to be prioritized for optimization.

[0037] Step S500: Based on the sensitivity of each grid parameter, the layout of each sensing region is grouped by sensitivity to obtain multi-level sensitivity layout regions. A regression tree is used to construct the mapping relationship between the grid parameter adjustment amount and the improvement amount of multi-touch point pressure sensing performance of each level of sensitivity layout region. A progressive layout adjustment rule is generated to optimize the set of grid parameters to be optimized in each sensing region layout. Based on the optimization results, the layout of the self-made glass sensor capacitor grid is parametrically reconstructed.

[0038] The multi-level sensitivity map region can be a grouping of the sensing region into different optimization levels based on the sensitivity of grid parameters, which can be used to achieve differentiated allocation of optimization strategies. In this embodiment, the multi-level sensitivity map region can be obtained by clustering or hierarchical classification based on sensitivity values. For example, the multi-level sensitivity map region can include, but is not limited to, one or more of high-sensitivity regions, medium-sensitivity regions, and low-sensitivity regions. The progressive map adjustment rule can be control logic describing the mapping relationship between the amount of grid parameter adjustment and the improvement of sensing performance, which can be used to guide the parametric reconstruction process. Further, the progressive map adjustment rule can be generated by using a regression tree model to learn the relationship between parameter adjustment path and performance gain. Parametric reconstruction can be a process of reconfiguring the physical or logical parameters of the capacitor grid according to the optimization results, which can be used to complete the iterative upgrade of map performance. In a specific embodiment, parametric reconstruction can be implemented by dynamically rewriting the driving parameters or triggering the hardware fine-tuning mechanism through software definition. Sensitivity grouping can be an operation of classifying the sensing region into different levels according to the sensitivity, so that optimization resources can be allocated as needed. Furthermore, sensitivity grouping can be achieved automatically by using K-means clustering or threshold splitting, thereby supporting the formulation of hierarchical optimization strategies.

[0039] Generating incremental layout adjustment rules can be achieved by using regression trees to model the relationship between parameter adjustments and performance improvements, forming an executable optimization path. Furthermore, generating incremental layout adjustment rules can be accomplished by constructing a multi-layered decision tree structure to implement a combined strategy of small-step fine-tuning and batch optimization, thereby achieving efficient and stable parameter optimization.

[0040] Taking pressure-based interaction on a touchscreen display as an example, the self-made glass sensor capacitive grid layout optimization method in this embodiment allows users to apply different pressures with two fingers to draw strokes in a touchscreen drawing application. The system uses a self-made glass sensor to capture the pressure magnitude and rate of change of two touch points in real time. When a sudden increase in pressure at a certain point causes abnormal capacitive response, the system immediately calls the sensing feature database to compare historical patterns and determines whether the electrode driving parameters of the corresponding area need to be adjusted. Subsequently, based on multiple correlation analysis, the coupling effect between the coordinate positioning of the area and pressure sensing is confirmed, generating a set of parameters to be optimized, and fine-tuning is performed on high-sensitivity areas according to regression tree rules. The entire process is completed in milliseconds, ensuring that the stroke thickness and transparency change naturally with pressure, avoiding misjudgment or delay.

[0041] This embodiment provides a method for optimizing the capacitive grid layout of a self-made glass sensor. It uses a multi-dimensional sensing analysis method to divide the capacitive grid layout structure into regions and construct a sensing feature database. Based on historical data, it determines various sensing response thresholds and quantifies the sensitivity of grid parameters. During real-time operation, it combines real-time monitoring data with threshold comparisons to trigger a parameter adjustment mechanism. It utilizes multi-correlation modeling and sensing importance calculation to screen key optimization parameters. Then, it groups regions according to sensitivity levels and establishes a mapping rule between parameter adjustment amounts and performance improvements. Ultimately, it achieves progressive optimization of the parameter set to be optimized and parameterized reconstruction of the layout. Through the synergistic effect of these steps, it can maintain high-precision coordinate positioning and stable pressure sensing response in multi-point dynamic touch and complex pressure change scenarios, while reducing unnecessary global parameter adjustments, lowering computational load and response latency, and improving the sensor's adaptability and reliability in diverse interactive environments.

[0042] In one embodiment, the layout structure of the self-made glass sensor capacitor grid is divided into regions using a multi-dimensional sensing analysis method to obtain a multi-sensing region layout, including: The electrode density, cell capacitance base value, mutual coupling coefficient and signal-to-noise ratio of each sub-region in the capacitor grid layout are obtained to construct the layout complexity quantization matrix of each sub-region. Electrode density can be the density of the distribution of capacitor grid electrodes per unit area, and can be used to influence regional sensitivity and spatial resolution. In an exemplary embodiment, electrode density can be obtained by analyzing the layout geometry data and calculating the length or number of electrodes per square millimeter. Furthermore, electrode density can be one of the fundamental parameters for constructing a layout complexity quantification matrix. Exemplarily, electrode density can include, but is not limited to, one or more of high-density, medium-density, and low-density areas. The unit capacitance base value can be the initial capacitance value of a single grid cell in a non-touch state, and can be used to reflect the regional electrical stability and dynamic response range. In this embodiment, the unit capacitance base value can be obtained through circuit measurement or simulation to obtain the static capacitance value when undisturbed. Furthermore, the unit capacitance base value can participate in the construction of the layout complexity quantification matrix. The mutual coupling coefficient can be an indicator of the signal interference intensity caused by electromagnetic field overlap between adjacent electrodes, and can be used to characterize signal isolation capability and multi-point identification anti-interference capability. In a specific embodiment, the mutual coupling coefficient can be calculated by electromagnetic simulation or measured differential signal crosstalk amplitude. Furthermore, the mutual coupling coefficient can be used to assess the signal interference risk between sub-regions and incorporated into the complexity modeling process. Signal-to-noise ratio (SNR) can be the ratio of the effective capacitance change signal to the background noise power, and can be used to measure sensing reliability and minimum detectable stress level. For example, SNR can be calculated by comparing the peak signal value with the noise fluctuation value under standard load conditions. Furthermore, SNR can serve as a stability dimension parameter in a layout complexity quantification matrix.

[0043] Constructing a layout complexity quantification matrix can be achieved by integrating four indicators—electrode density, unit capacitance base value, mutual coupling coefficient, and signal-to-noise ratio—into a structured data matrix. Furthermore, this matrix can be constructed by normalizing the aforementioned parameters of each sub-region and combining them into row vectors, then organizing them into a numerical matrix covering the entire layout. This allows for a digital representation of the comprehensive electrical characteristics of each sub-region.

[0044] A perceptual directed graph is established based on the complex quantification matrix of each region, and clustering is performed on the mean and peak values ​​of capacitance change amplitude of each sub-region under pressure load to preliminarily partition each sub-region and obtain the capacitance response grouping results. The quantification matrix of map complexity can be a multidimensional numerical matrix constructed using the four parameters mentioned above as dimensions. It can be used to quantify the differences in technical characteristics across different regions and support systematic zoning decisions. In this embodiment, the quantification matrix of map complexity serves as the input data structure for generating a directed perceptual graph. Establishing the directed perceptual graph can be a process of constructing weighted directed edges by calculating similarity or influence relationships based on the quantification matrix of map complexity, with sub-regions as nodes. Furthermore, establishing the directed perceptual graph can be achieved by using graph neural networks or correlation propagation algorithms to generate connection structures reflecting signal flow, thereby enhancing the understanding of electrical coupling relationships and signal propagation paths in region partitioning. The average capacitance change amplitude can be the average change in capacitance response of a sub-region during multiple pressure loading processes, and can be used to reflect the overall response sensitivity of the region. In an exemplary embodiment, the average capacitance change amplitude can be obtained by taking the arithmetic mean of continuously sampled data. Furthermore, the average capacitance change amplitude can participate in clustering operations to achieve preliminary zoning. The peak capacitance change amplitude can be the maximum instantaneous change in capacitance response of a sub-region under pressure, and can be used to characterize the dynamic carrying capacity and nonlinear boundaries of the region. For example, the peak value of the capacitance change amplitude can be obtained by extracting the maximum value from the dynamic signal sequence. Furthermore, the peak value of the capacitance change amplitude can be used together with the mean as a clustering feature variable.

[0045] Clustering can be an operation that classifies sub-regions based on the mean and peak values ​​of capacitance change amplitude. Furthermore, clustering can be implemented by introducing a weighted distance metric to highlight the influence of key features, thereby improving group consistency and automatically merging regions with similar response characteristics. The capacitance response grouping results can be a preliminary functional region division formed after clustering based on capacitance change amplitude characteristics, which can be used to form a set of regions with similar dynamic response characteristics. In this embodiment, the capacitance response grouping results serve as the input basis for subsequent coordinate offset correction. For example, the capacitance response grouping results may include, but are not limited to, one or more of the following: high response group, standard response group, hysteresis response group, etc.

[0046] The touch point coordinate offset data corresponding to each sub-region is obtained, and the coordinate offset data is compared with the preset positioning tolerance threshold. The capacitive response grouping results are then refined and adjusted based on the comparison results to obtain the multi-sensing area map.

[0047] The touch point coordinate offset data can be the spatial deviation between the actual touch position and the system-recognized position, which can be used to evaluate positioning accuracy. In one specific embodiment, the touch point coordinate offset data can be obtained by recording the actual touch point with a high-precision external positioning device and comparing it with the system output. Furthermore, the touch point coordinate offset data can be used to verify and correct the capacitive response grouping results. The preset positioning tolerance threshold can be the maximum allowable coordinate recognition error limit, which can be used to determine whether the positioning performance meets the standard. For example, the preset positioning tolerance threshold can be set to a fixed value or an adaptively changing value according to the accuracy requirements of the application scenario. Further, the preset positioning tolerance threshold can be compared with the coordinate offset data to trigger a partition adjustment mechanism. Refined adjustment can be a process of comparing the coordinate offset data with the preset positioning tolerance threshold to determine whether the initial partition boundary needs to be modified. Further, refined adjustment can be achieved by separately splitting or reclassifying areas exceeding the tolerance threshold into higher-precision groups, thereby ensuring that the final partition meets the positioning accuracy requirements.

[0048] For example, in the scenario of flexible screen gesture control devices, the self-made glass sensor capacitance grid layout optimization method in this embodiment can address the issue of non-uniform capacitance response in edge regions due to structural deformation during multi-finger pressure scaling operations on curved flexible displays. The system first constructs a complexity matrix based on electrode density and signal-to-noise ratio, generating a directed perceptual graph to reveal the signal attenuation paths between the edge and center regions. By applying uniform pressure, the response amplitude of each region is obtained and clustered, initially dividing it into a high-response center region and a low-response edge region. Subsequently, a significant coordinate offset exceeding a preset tolerance threshold is detected in the edge region, leading to further subdivision into compensation and correction sub-regions. The resulting multi-sensory region layout effectively improves edge touch accuracy.

[0049] This embodiment constructs a quantification matrix of the complexity of the capacitance grid by acquiring the electrode density, unit capacitance base value, mutual coupling coefficient, and signal-to-noise ratio of each sub-region in the capacitance grid layout. Based on this matrix, a directed sensing graph is established, and clustering operations are performed by combining the mean and peak values ​​of capacitance change amplitude under pressure load to achieve preliminary partitioning. Then, the capacitance response grouping results are refined and adjusted by comparing the touch point coordinate offset data with the preset positioning tolerance threshold. The physical basis of the region division is enhanced by multi-dimensional electrical characteristic modeling. The clustering method is used to merge sub-regions with consistent response behavior, and the partition boundaries are optimized by the actual positioning error feedback mechanism. This achieves the technical effect of making the final multi-sensing region layout simultaneously take into account dynamic response consistency and spatial positioning accuracy.

[0050] In one embodiment, the base capacitance response threshold includes the baseline capacitance value, the lower limit of the dynamic range, and the upper limit of saturation; the coordinate positioning accuracy threshold includes the minimum resolvable distance and the linearity error limit; and the pressure change rate sensing threshold includes the minimum perceptible rate gradient and the maximum following rate.

[0051] The basic capacitance response threshold serves as a benchmark value for determining whether a capacitance signal is abnormal. It defines the effective operating range of the capacitance response, preventing misjudgment of weak signals and sensor failure caused by strong signal overflow. In this embodiment, the basic capacitance response threshold can be determined through historical data statistics and system noise analysis. It is composed of a baseline capacitance value, a lower limit of the dynamic range, and an upper limit of saturation, representing the initial capacitance in the no-touch state, the lowest boundary of the effective response, and the upper limit of signal saturation, respectively. Furthermore, the basic capacitance response threshold can be used in conjunction with real-time capacitance change data to detect signal deviation and trigger a parameter adjustment mechanism. For example, the basic capacitance response threshold may include, but is not limited to, one or more of the following: baseline capacitance value, lower limit of the dynamic range, and upper limit of saturation. The coordinate positioning accuracy threshold can be the maximum allowable positioning deviation limit, used to ensure spatial resolution and trajectory accuracy in multi-touch scenarios. In an exemplary embodiment, the coordinate positioning accuracy threshold is obtained based on spatial resolution testing and fitting error analysis of historical touch point coordinate data. It includes the minimum distinguishable distance and linearity error limit. The former refers to the closest distance between two touch points that can be distinguished, while the latter reflects the nonlinear deviation tolerance between the coordinate output and the actual position. Furthermore, the coordinate positioning accuracy threshold can be combined with real-time touch point coordinate data to determine whether positioning distortion or crosstalk exists. For example, the coordinate positioning accuracy threshold may include, but is not limited to, one or more of the following: minimum resolvable distance, linearity error limit, etc. The pressure change rate sensing threshold can be a minimum resolvable rate standard for dynamic pressure changes, used to ensure good sensing capability for both slow light touches and rapid presses. In this embodiment, the pressure change rate sensing threshold is calibrated based on dynamic response tests using historical pressure change rate data, including a minimum resolvable rate gradient and a maximum following rate, representing the lower limit of the pressure rise slope that the system can reliably detect and the upper limit of the system's distortion-free response when the user rapidly increases pressure, respectively. Furthermore, the pressure change rate sensing threshold can be used in conjunction with real-time pressure change rate data to evaluate the integrity of the pressure sensing function. For example, the pressure change rate sensing threshold may include, but is not limited to, one or more of the following: minimum resolvable rate gradient, maximum following rate, etc.

[0052] In one embodiment, determining the basic capacitance response threshold, coordinate positioning accuracy threshold, and pressure change rate sensing threshold for each sensing area map based on a sensing feature database includes: The baseline capacitance sampling value, capacitance saturation value and nonlinear distortion point of each sensing area in the non-touch state are extracted from the sensing feature database. The static response curve of capacitance-pressure is fitted by Gaussian process regression to determine the baseline capacitance value, lower limit of dynamic range and upper limit of saturation of each sensing area. The baseline capacitance sampling value can be the initial capacitance measurement value of each sensing area of ​​the capacitance grid in a non-touch state, and can be used to determine the baseline capacitance value in the basic capacitance response threshold. In this embodiment, the baseline capacitance sampling value can be obtained as a benchmark by performing multiple capacitance measurements on the capacitance grid under no external force. Further, the baseline capacitance sampling value can be used as the basic input data for fitting the capacitance-pressure static response curve using Gaussian process regression. The capacitance saturation value can be the maximum capacitance response value that the sensing area of ​​the capacitance grid can achieve under maximum pressure, and can be used to determine the upper limit of saturation in the basic capacitance response threshold. In this embodiment, the capacitance saturation value can be obtained by applying gradually increasing pressure until the capacitance change tends to stabilize, and recording the capacitance value in this state. Further, the capacitance saturation value can participate in the modeling of the capacitance-pressure response curve together with the nonlinear distortion point. The nonlinear distortion point can be the inflection point in the capacitance-pressure response curve that deviates from the linear relationship, and can be used to define the effective working interval between the lower limit of the dynamic range and the upper limit of the saturation. In this embodiment, the nonlinear distortion point can be identified by analyzing the curve of capacitance changing with pressure to determine its nonlinear inflection point. Furthermore, the nonlinear distortion point can participate in setting the capacitance response threshold along with the capacitance saturation value. Gaussian process regression, a nonparametric regression method, can be used to infer the function distribution from a finite sample and can be used to fit the capacitance-pressure static response curve, improving the accuracy of the threshold setting. In this embodiment, Gaussian process regression can be based on a Bayesian framework, using a kernel function to calculate the similarity between samples and construct a confidence interval for the capacitance-pressure relationship. Furthermore, Gaussian process regression can be applied to the baseline capacitance sample value, capacitance saturation value, and nonlinear distortion point to generate a response model.

[0053] Extracting baseline capacitance samples can be achieved by filtering capacitance data under non-touch conditions from a sensor feature database. Furthermore, extracting baseline capacitance samples can be achieved by setting a time window to filter stable-state data and eliminate transient interference, thus obtaining reliable baseline capacitance values. Extracting capacitance saturation values ​​and nonlinear distortion points can be achieved by filtering capacitance response data under high pressure conditions from a database and identifying nonlinear feature points. Furthermore, extracting capacitance saturation values ​​and nonlinear distortion points can be achieved by using a curve inflection point detection algorithm to identify nonlinear distortion points, thereby defining the effective dynamic range of the capacitance response. Gaussian process regression fitting can be achieved by constructing a nonlinear response model based on historical capacitance-pressure data. Furthermore, Gaussian process regression fitting can be improved by selecting an appropriate kernel function, thereby generating a high-precision capacitance-pressure static response curve.

[0054] Multiple coordinate sampling data of standard test points for each sensing area map are extracted from the sensing feature database. The root mean square error and linear fitting residual between the map and the theoretical coordinates are calculated. The minimum resolvable distance and linearity error limit of each sensing area map are determined by confidence interval estimation. The standard test point can be a preset known coordinate location used to evaluate the positioning performance of the capacitive grid. In this embodiment, the standard test point can be achieved by applying standard pressure at a fixed position and collecting the corresponding capacitive response and coordinate output. Further, the standard test point can include, but is not limited to, one or more of the following: center test point, edge test point, diagonal test point, etc. The root mean square error (RMSE) can be the square root of the mean square of the deviations between the measured coordinates and the theoretical coordinates, and can be used to evaluate the repeatability and stability of coordinate positioning. In this embodiment, the RMSE can be achieved by taking the square root of the average of the squared differences between multiple measurement results and the theoretical value. Further, the RMSE can be combined with the linear fitting residual to determine the positioning accuracy threshold through confidence interval estimation. The linear fitting residual can be the deviation between the measured coordinates and the ideal linear model, and can be used to evaluate the systematic deviation of coordinate positioning. In this embodiment, the linear fitting residual can be achieved by fitting the relationship between the theoretical coordinates and the measured coordinates using the least squares method and calculating the residual distribution. Further, the linear fitting residual can be used in conjunction with the RMSE to participate in setting the positioning accuracy threshold. Confidence interval estimation can quantify the uncertainty range of parameter estimates using statistical methods, and can be used to determine reasonable ranges for minimum resolvable distance and linearity error limits. In this embodiment, confidence interval estimation can calculate the upper and lower confidence limits of parameters using sample data, reflecting the reliability of the estimated values. Furthermore, confidence interval estimation can be applied to the root mean square error and linear fitting residuals to generate a coordinate positioning accuracy threshold.

[0055] Extracting standard test point coordinate data can be achieved by retrieving multiple coordinate sampling results of preset test points from a database. Furthermore, extracting standard test point coordinate data can be achieved by classifying and extracting data according to test point numbers to support regional analysis, thus providing a data foundation for positioning accuracy assessment. Calculating the root mean square error and linear fitting residuals can be achieved by performing statistical analysis on measured and theoretical coordinates. Furthermore, calculating the root mean square error and linear fitting residuals can be achieved by using the least squares method for linear fitting, thereby quantifying the repeatability and systematic errors of coordinate positioning. Confidence interval estimation can be achieved by estimating the positioning error based on statistical methods. Furthermore, confidence interval estimation can be improved by employing the Bootstrap method to enhance estimation stability, thereby allowing for the setting of reasonable minimum resolvable distance and linearity error limits.

[0056] Capacitive response delay data and rate tracking error of each sensing region under a preset pressure ramp signal are extracted from the sensing feature database. The measurement noise is smoothed by using a Kalman filter and the system response bandwidth is identified in order to determine the minimum perceptible rate gradient and maximum following rate of each sensing region.

[0057] The pressure ramp signal can be a pressure input signal that varies linearly with time, and can be used to test the dynamic performance of the capacitive response. In this embodiment, the pressure ramp signal can be implemented by applying continuously varying pressure to the capacitive grid through a control device to simulate dynamic touch behavior. Furthermore, the pressure ramp signal, along with capacitive response delay data and rate tracking error, can constitute the basis for setting the pressure rate sensing threshold. The capacitive response delay data can be the time lag of the capacitance change relative to the pressure input, and can be used to evaluate the real-time performance of the pressure change rate sensing. In this embodiment, the capacitive response delay data can be the time difference between the pressure input and the capacitive response, reflecting the system response speed. Furthermore, the capacitive response delay data can be used in conjunction with the rate tracking error to set the pressure rate sensing threshold. The rate tracking error can be the deviation between the pressure change rate sensed by the system and the actual rate, and can be used to evaluate the accuracy of the dynamic pressure response. In this embodiment, the rate tracking error can be calculated by comparing the pressure input slope and the capacitive response slope. Furthermore, the rate tracking error can be used in conjunction with the capacitive response delay data to set the pressure rate sensing threshold.

[0058] A Kalman filter can be a recursive algorithm used to estimate the system state from noisy measurement data, and can be used to extract effective signal components from pressure rate sensing. In this embodiment, the Kalman filter can be implemented by smoothing measurement noise through a prediction-update mechanism to improve signal stability. Furthermore, the Kalman filter can be applied to capacitor response delay data and rate tracking error to identify the system response bandwidth. The system response bandwidth can be the frequency range within which the system can accurately track changes in the input signal, and can be used to determine the limiting capability of pressure change rate sensing. In this embodiment, the system response bandwidth can be identified by analyzing the filtered capacitor response data to identify its frequency response characteristics. Furthermore, the system response bandwidth can serve as the basis for setting the minimum perceptible rate gradient and the maximum following rate. Extracting pressure ramp signal response data can be achieved by retrieving capacitor response records under a preset pressure ramp from a database. Furthermore, extracting pressure ramp signal response data can be achieved by grouping data according to pressure change rate to support multi-rate analysis, thereby providing basic data for dynamic response analysis. Kalman filter processing can be implemented by performing noise smoothing and state estimation on the capacitor response data. Furthermore, Kalman filter processing can be achieved by dynamically adjusting filter parameters to adapt to different pressure change rates, thereby improving the signal-to-noise ratio and accuracy of pressure rate sensing. The system response bandwidth can be determined by analyzing the system's frequency response characteristics based on the filtered data. Furthermore, the system response bandwidth can be identified by using spectral analysis methods to determine the upper limit of the response bandwidth, thus defining the dynamic limits of pressure rate sensing.

[0059] Taking an industrial-grade high-precision pressure drawing device as an example, the self-made glass sensor capacitance grid layout optimization method in this embodiment can be as follows: On a professional drawing board, the user uses a pressure-sensitive pen to draw fine lines. The system automatically identifies whether the current pen pressure is within the dynamic range by using the preset capacitance-pressure response curve in the perception feature database. When the pressure is detected to be close to the upper limit of saturation, the system triggers a Gaussian process regression model to refit the local response curve and optimizes the capacitance response data through a Kalman filter to prevent signal distortion. At the same time, the system continuously monitors the minimum resolvable distance and linearity error limit to ensure that the pen touch coordinates maintain high-precision positioning during complex gesture operations.

[0060] In one embodiment, the impact of changes in grid parameters in each sensing region's layout on the accuracy of multi-touch point pressure sensing is analyzed based on a sensing feature database to obtain the grid parameter sensitivity of each sensing region's layout, including: Historical touch point coordinates, historical pressure change rate data, and corresponding historical grid parameter change records of each sensing area are extracted from the sensing feature database and used as training samples to train a temporal convolutional network, thereby obtaining a response model that represents the dynamic relationship between grid parameter changes and multi-dimensional sensing performance in each sensing area.

[0061] The temporal convolutional network (TCNN) can be a deep neural network structure specifically designed for processing time-series data. It extracts local features from the sequence through convolutional layers and models their temporal dependencies, enabling it to model the dynamic correlation between grid parameter changes and multi-dimensional sensing performance. In this embodiment, the TCNN extracts features from the input sequence by stacking multiple convolutional layers and activation functions, combined with a sliding window mechanism in the time dimension. Furthermore, the TCNN can be used in conjunction with a sensing feature database to train a response model. For example, the TCNN can be one or more of, including but not limited to, one-dimensional TCNNs, dilated TCNNs, and residual TCNNs. The response model can be a mathematical model characterizing the dynamic mapping relationship between grid parameter changes and multi-dimensional sensing performance in each sensing area, and can be used to predict the impact of specific parameter changes on sensing performance. In this embodiment, the response model is obtained by training the TCNN based on training samples (historical touch point coordinate data, historical pressure change rate data, and historical grid parameter change records). For example, the response model can provide predictive outputs for parameter-performance correlation strength analysis.

[0062] Training a temporal convolutional network can be done using historical touch point coordinate data, historical pressure change rate data, and historical grid parameter change records as input-output pairs. The network model is trained to capture the dynamic relationship between parameter changes and perception performance. Furthermore, training a temporal convolutional network can be achieved by using a sliding window to construct training samples, introducing regularization to prevent overfitting, and using cross-validation to select the optimal model, thereby constructing a high-precision response model for performance prediction.

[0063] Based on the response model, the changes in coordinate positioning error and pressure rate tracking error caused by the changes in specific grid parameters under a preset pressure mode sequence are statistically analyzed for each sensing area. Spearman rank correlation analysis is then performed on the error changes according to the pressure change rate interval to obtain the parameter-performance correlation strength index for each sensing area.

[0064] The parameter-performance correlation strength index can be a comprehensive score that quantifies the impact of specific grid parameter changes on multi-dimensional sensing performance, reflecting the sensitivity of parameters to performance. In this embodiment, the parameter-performance correlation strength index is based on the response model prediction results, statistically analyzing the changes in coordinate positioning error and pressure rate tracking error caused by parameter changes under a preset pressure mode sequence, and derived through Spearman rank correlation analysis. Furthermore, the parameter-performance correlation strength index can serve as the input basis for a sensitivity grading table. The statistical error change can be based on the response model simulation of changes in coordinate positioning error and pressure rate tracking error caused by specific parameter changes under a preset pressure mode sequence. Furthermore, the statistical error change can be achieved by setting multiple sets of parameter perturbation experiments, recording error responses, and classifying them statistically according to pressure change rate intervals, thereby quantifying the specific impact of parameter changes on sensing performance. The Spearman rank correlation analysis can be a non-parametric correlation analysis of the ranking relationship between error change and parameter change. Furthermore, Spearman rank correlation analysis can be achieved by calculating the rank correlation coefficient and assessing the strength of the monotonic relationship between parameters and performance indicators, thereby eliminating the limitations of the linear assumption and revealing the nonlinear correlation strength between parameters and performance.

[0065] By using the preset functional weights of each sensing region map in the global sensing task, the corresponding parameter-performance correlation strength index is weighted and fused. The fusion result is then divided into intervals according to the preset sensitivity grading threshold to obtain a grid parameter sensitivity grading table to quantitatively determine the grid parameter sensitivity of each sensing region map.

[0066] The sensitivity grading threshold can be a numerical boundary used to divide the parameter-performance correlation strength index into different sensitivity levels, and can be used to achieve quantitative grading of grid parameter sensitivity. In this embodiment, the sensitivity grading threshold is set according to the overall system performance requirements and is used to divide the fused index into intervals. For example, the sensitivity grading threshold can be combined with the parameter-performance correlation strength index to generate a sensitivity grading table. The grid parameter sensitivity grading table can be a lookup table that records the sensitivity levels corresponding to different grid parameters in each sensing area, and can be used to provide parameter priority basis for subsequent optimization. In this embodiment, the grid parameter sensitivity grading table is generated by weighted fusion and interval division of the parameter-performance correlation strength index. For example, the grid parameter sensitivity grading table can guide the generation of the set of grid parameters to be optimized. Weighted fusion can be a weighted average of the parameter-performance correlation strength index based on the functional importance of each sensing area in the global task. Furthermore, weighted fusion can be achieved by adopting a task priority weight allocation strategy, such as assigning higher weights to the central area, thereby improving the global consistency of sensitivity assessment. Interval partitioning can be achieved by dividing the fused indicators into different level intervals according to a preset sensitivity grading threshold. Furthermore, interval partitioning can be implemented using methods such as equal width, equal frequency, or distribution-based adaptive partitioning strategies, thereby enabling visualization and operational representation of sensitivity.

[0067] For example, in scenarios involving multi-point pressure interaction during touchscreen gaming, the self-made glass sensor capacitive grid layout optimization method in this embodiment allows users to control character movement and attacks by rapidly alternating pressure with two fingers in competitive games. The system uses a self-made glass sensor to sense the pressure intensity and rate of change at each point in real time. When the pressure perception error increases in a certain area due to continuous rapid clicking, the system calls a response model to predict the error change after parameter adjustment in that area and determines key influencing parameters through Spearman's rank correlation analysis. Combining the functional weight of that area in the interactive task, the system weights and fuses the parameter-performance correlation strength index, and determines it to be a high-sensitivity area based on a sensitivity grading table, thereby triggering a fine-tuning mechanism to quickly restore perception accuracy and ensure stable and accurate operation response.

[0068] In one embodiment, the multiple correlations between capacitive response, coordinate positioning, and pressure rate sensing of each sensing region are analyzed based on a sensing feature database. Based on these multiple correlations, grid parameters are extracted from a grid parameter adjustment table to calculate sensing importance, thereby generating a set of grid parameters to be optimized for each sensing region, including: Historical capacitance response sequences, historical coordinate positioning data, and historical sensing performance evaluation data of each sensing area map under different pressure change rates are extracted from the sensing feature database. The similarity of sensing behavior under different pressure scenarios is calculated by dynamic time warping algorithm. Based on the similarity calculation results, each sensing area map is clustered under pressure scenarios to obtain a set of typical pressure sensing scenarios. The historical capacitance response sequence can be time-series data of capacitance values ​​changing over time in each sensing area of ​​the capacitance grid under different pressure change rates, which can be used to characterize the temporal adaptability of the capacitance response to dynamic pressure. In this embodiment, the historical capacitance response sequence can be obtained by recording the continuous fluctuation trajectory of the capacitance between electrodes during the touch process using a high sampling rate sensor circuit. For example, the historical capacitance response sequence can be combined with historical coordinate positioning data and historical perception performance evaluation data to construct perception behavior features. The historical coordinate positioning data can be a time-series record of the two-dimensional spatial coordinates of the touch point output by the system under different pressure loads, which can be used to evaluate positioning stability and trajectory tracking consistency. In an exemplary embodiment, the historical coordinate positioning data can be spatially solved based on the capacitance differential algorithm to generate a continuous frame coordinate sequence. Furthermore, the historical coordinate positioning data can be used in conjunction with the historical capacitance response sequence for perception behavior similarity calculation. The historical perception performance evaluation data can be a comprehensive score or label of capacitance response, coordinate accuracy, and pressure rate tracking effect in historical touch events, which can be used as a supervisory signal or reference benchmark for pressure scene clustering. In one specific embodiment, historical perception performance evaluation data can be obtained through expert annotation or by weighted performance scoring based on multiple indicators to reflect the quality of perception. Furthermore, historical perception performance evaluation data can be used to guide the rationality of scene classification results after dynamic time warping.

[0069] Dynamic time warping (RTW) algorithms can be used as sequence matching methods to align time-series data of different lengths or speeds to calculate their similarity. They can be used to eliminate interference from differences in pressure change rates on the comparison of perceived behaviors. In this embodiment, RTD can minimize the cumulative distance between two sets of time-series data by nonlinearly stretching or compressing the time axis. For example, RTD can introduce weighted Euclidean distance or Mahalanobis distance to enhance robustness to abnormal fluctuations, or combine sliding window segmented matching to improve local pattern recognition capabilities, thereby achieving semantic alignment of perceived behaviors under different pressure rhythms and eliminating temporal asynchronous interference. Pressure scenario clustering can classify historical pressure events into several typical interaction patterns based on the similarity of perceived behaviors, and can be used to identify high-frequency, high-impact typical pressure interaction types. In this embodiment, pressure scenario clustering can be achieved by grouping the dynamically time-warped behavioral feature vectors using a clustering algorithm. For example, pressure scenario clustering can employ K-means, hierarchical clustering, DBSCAN, or Gaussian mixture models to extract representative typical interaction patterns and reduce the complexity of subsequent modeling. For example, pressure scenario clustering may include one or more of the following: light touch swipe scenario, heavy pressure click scenario, and multi-finger pressure alternation scenario.

[0070] The mean, variance, extreme values, and covariance matrix of capacitive response characteristic values, coordinate stability index, and rate tracking accuracy within each typical pressure sensing scenario set are calculated to construct a multimodal sensing feature tensor. The partial least squares regression method is used to quantify the complex correlation between the layout mesh parameters and the multimodal sensing performance index, and the multivariate coupling relationship matrix of each sensing area layout is obtained. The multimodal sensing feature tensor can be a three-dimensional structured data volume composed of statistics of three types of indicators: capacitance response, coordinate stability, and rate tracking accuracy. It can be used to comprehensively characterize the statistical distribution and variable relationships of sensing performance under various scenarios. In this embodiment, the multimodal sensing feature tensor can calculate the mean, variance, extreme values, and covariance matrices of each indicator within each typical pressure scenario set and stack them into a tensor. Furthermore, the multimodal sensing feature tensor can introduce a sliding statistical window to dynamically update the variance and covariance, or embed higher-order moment features to enhance nonlinear expressive power, thereby forming a high-dimensional performance descriptor that can be analyzed by a regression model. The multivariate coupling relationship matrix can be a weight matrix describing the nonlinear dependence between grid parameters and multimodal sensing performance indicators, and can be used to reveal the synergistic influence path of parameters on multi-target performance. In this embodiment, the multivariate coupling relationship matrix can be achieved by fitting the latent variable relationship between the grid parameter vector and the multimodal sensing feature tensor using partial least squares regression. For example, multivariate coupling relationship matrices can introduce cross-validation to select the optimal latent variable dimension, or combine sparse constraints to avoid overfitting, thereby revealing the synergistic driving mechanism of parameters on multi-objective performance and breaking through the limitations of univariate analysis.

[0071] Candidate adjustment parameters are extracted from the grid parameter adjustment table of each sensing area based on the multivariate coupling relationship matrix. The contribution of the extracted candidate grid parameters to the overall sensing performance improvement is quantified by the random forest algorithm. The sensing importance of the extracted candidate grid parameters is quantified based on the contribution. Candidate grid parameters whose importance calculation results exceed the importance threshold are selected to form the set of grid parameters to be optimized for each sensing area.

[0072] The candidate adjustment parameters can be a subset of adjustable parameters that may affect perception performance, initially selected from the grid parameter adjustment table. This can be used to narrow the optimization search space and avoid interference from irrelevant parameters. In this embodiment, candidate adjustment parameters can be obtained by filtering based on all operable parameters in the adjustment table and significant related terms in the multivariate coupling relationship matrix. Furthermore, candidate adjustment parameters can serve as the input feature set for the random forest algorithm. Perception importance can be a quantitative score that measures the contribution of candidate grid parameters to the overall improvement of perception performance, and can be used to determine which parameters are most worthy of optimization. In this embodiment, perception importance can be achieved by calculating the information gain or Gini importance of each parameter in the regression tree split using the random forest algorithm. For example, perception importance can be enhanced by using permutation importance or SHAP value to improve interpretability, or by combining multiple sampling to improve stability, thereby achieving a nonlinear, interactive contribution assessment of the parameter's impact and avoiding linear assumption bias. The set of grid parameters to be optimized can be a set of parameters identified as high-value optimization targets after being screened by perception importance. This can be used to accurately focus optimization resources and improve parameter reconstruction efficiency. In this embodiment, the set of grid parameters to be optimized can be formed by selecting candidate parameters whose importance exceeds a preset threshold. Furthermore, the set of grid parameters to be optimized can be configured with a dynamic threshold mechanism to adaptively adjust the selection criteria based on the current system load or scene activity, thereby ensuring that optimization resources are concentrated on the most promising parameters and improving system response efficiency and convergence speed.

[0073] Taking multi-pressure level interaction with a touchscreen stylus as an example, the self-made glass sensor capacitive grid layout optimization method in this embodiment allows users to perform calligraphy writing with a pressure-sensitive pen, rapidly switching between light, medium, and heavy pressure, while simultaneously using multi-finger assisted positioning. The system collects the capacitance timing, coordinate trajectory, and performance score during this process. By dynamically warping the behavior patterns under different writing speeds, it clusters and identifies typical scenarios such as "rapid heavy pressure stroke" and "continuous light swipe transition." A multimodal feature tensor is constructed for each scenario, and partial least squares regression is used to find that electrode spacing and edge capacitance coupling have the greatest impact on pressure rate tracking. Random forest further confirms that only three parameters significantly contribute to the overall perception quality, while the remaining parameters can be frozen. Finally, an optimization set containing only these three parameters is generated, enabling the system to complete local reconstruction within milliseconds, ensuring that the thickness of the handwriting and ink smudges change naturally with pressure, without delay or distortion.

[0074] In one embodiment, the layout of each sensing region is grouped by sensitivity based on the sensitivity of each grid parameter to obtain a multi-level sensitivity layout region. A regression tree is then used to construct a mapping relationship between the grid parameter adjustment amount and the improvement in multi-touch point pressure sensing performance for each level of sensitivity layout region. A progressive layout adjustment rule is generated to optimize the set of grid parameters to be optimized for each sensing region layout. Based on the optimization results, the layout of the self-made glass sensor capacitor grid is then parametrically reconstructed, including: The performance index time series data of the self-made glass sensor with optimized layout parameters were obtained under various pressure change rate test modes. The performance index time series data was used as input features to train the conditional generative adversarial network. The trained network was used to generate high-fidelity virtual pressure touch signals in real time and simulate sensing response to obtain virtual perception response data. The performance index time-series data can be dynamic performance sequence data output by the sensor under various pressure change rate test modes after layout optimization. This data can be used as conditional input to drive generative adversarial network (GAN) training, simulating sensing behavior under extreme conditions. In this embodiment, the performance index time-series data can be a continuous data stream of indicators such as capacitance response, coordinate offset, and pressure rate tracking error over time, collected by a high-sampling-rate sensing circuit. Furthermore, the performance index time-series data can be the input features of the conditional GAN, forming a training loop with the virtual sensing response data. The conditional GAN ​​can be a deep generative model capable of generating realistic time-series data based on input conditions, used to generate high-fidelity virtual pressure touch signals and sensing responses without physical testing. In an exemplary embodiment, the conditional GAN ​​can be trained adversarially to allow the generator to learn the distribution characteristics of the performance index time-series data, while the discriminator distinguishes between real and generated data. Furthermore, the conditional GAN ​​can be enhanced by employing a time-series encoder to improve temporal dependency modeling and introducing an attention mechanism to focus on pressure abrupt change segments, thereby achieving a high-fidelity virtual extension of sensing behavior under extreme pressure rates.

[0075] Virtual sensing response data can be sensor output signals generated by a conditional generative adversarial network under simulated pressure touch scenarios. It can be used to expand extreme condition datasets for boundary analysis and parameter scanning. In one specific embodiment, virtual sensing response data can be generated by using time-series performance index data as conditions to produce capacitance, coordinate, and rate response sequences corresponding to the pressure change rate. Furthermore, virtual sensing response data can serve as the input basis for parameter tolerance boundary calculations and virtual layout simulation.

[0076] Obtaining parameter tolerance boundaries can be achieved by applying extreme pressure change rates to virtual sensing response data to identify performance collapse thresholds. For example, parameter tolerance boundaries can be obtained by using a sliding window statistical method to detect inflection points of sudden error increases, defining these as tolerance boundaries. This allows for the establishment of physical feasibility constraints before manufacturing, preventing the design from entering unstable regions.

[0077] The virtual layout simulation environment can be a digital simulation platform containing a four-dimensional parameter space including electrode width, spacing, shape factor, and interlayer coverage area. It can be used to support large-scale virtual iterations before multi-parameter collaborative optimization. In this embodiment, the virtual layout simulation environment can construct adjustable electrode geometry using parametric modeling tools and couple electromagnetic field and mechanical stress simulation engines. Furthermore, the virtual layout simulation environment can automatically switch geometric configurations using parametric scripts and improve simulation efficiency by combining mesh adaptive algorithms, thereby enabling large-scale, high-degree-of-freedom exploration of the virtual parameter space.

[0078] The optimization objective is to minimize the weighted sum of the comprehensive accuracy of multi-touch pressure sensing, which includes coordinate error, pressure value error, and rate tracking error. Multi-objective Bayesian optimization is used to collaboratively optimize the virtual mesh parameters of each sensing area in the virtual map simulation environment, resulting in the Pareto optimal set of virtual parameter combinations for each sensing area. The comprehensive accuracy weighted sum can be a comprehensive performance evaluation index that weights and fuses coordinate error, pressure value error, and rate tracking error. It can be used as a unified objective function for multi-objective optimization to balance multi-dimensional performance requirements. In this embodiment, the comprehensive accuracy weighted sum can be a linear weighted sum of the three types of errors, with weight coefficients set according to the priority of the perception task. Furthermore, the comprehensive accuracy weighted sum can be the optimization objective of multi-objective Bayesian optimization, driving the virtual parameter combination search. Multi-objective Bayesian optimization can use the comprehensive accuracy weighted sum as the optimization objective to search for a Pareto optimal solution set in a virtual layout simulation environment. For example, multi-objective Bayesian optimization can balance exploration and development using expected improvement and Pareto front update strategies, and introduce a Gaussian process surrogate model to accelerate convergence, thereby efficiently locating the multi-objective equilibrium point in a high-dimensional parameter space and reducing the number of simulations. The Pareto optimal virtual parameter combination set can be a set of parameters in the virtual layout simulation environment that cannot be further optimized for comprehensive accuracy without degrading the performance of other dimensions; it can be used to provide optimal parameter candidate schemes before manufacturing. In this embodiment, the Pareto optimal virtual parameter set can be obtained by searching for a non-dominated solution set in the four-dimensional parameter space using a multi-objective Bayesian optimization algorithm. Furthermore, the Pareto optimal virtual parameter set can serve as the input basis for high-precision electromagnetic-structural coupling simulation and final reconstruction.

[0079] Based on virtual sensing response data, the parameter tolerance boundary of each sensing area layout under extreme pressure change rate is obtained as a stability constraint condition, and a virtual layout simulation environment containing four sets of virtual mesh parameters including electrode width, spacing, shape factor and interlayer coverage area is constructed through parametric scanning. The optimization objective is to minimize the weighted sum of the comprehensive accuracy of multi-touch pressure sensing, which includes coordinate error, pressure value error, and rate tracking error. Multi-objective Bayesian optimization is used to collaboratively optimize the virtual mesh parameters of each sensing area in the virtual map simulation environment, resulting in the Pareto optimal set of virtual parameter combinations for each sensing area. High-precision electromagnetic-structural coupling simulation is performed on the capacitor grid based on each Pareto optimal virtual parameter combination set. When the sensing performance index data obtained from the simulation is stable within the preset performance target range, the layout of the self-made glass sensor capacitor grid is controlled according to the optimized virtual parameter combination set to perform final parameterization reconstruction and output before manufacturing.

[0080] High-precision electromagnetic-structural coupling simulation can be a multi-physics simulation that simultaneously simulates the influence of electrode geometry on electric field distribution and mechanical deformation, and can be used to verify the feasibility of virtual optimization results in a real physical environment. In this embodiment, high-precision electromagnetic-structural coupling simulation can involve importing virtual parameter combinations into a finite element simulation platform to solve the coupling equations of capacitive coupling and stress distribution. Furthermore, high-precision electromagnetic-structural coupling simulation can monitor the coordinated response of capacitive fluctuations and structural deformation by setting multi-condition pressure loading paths, thereby ensuring that the virtual optimization results still meet performance targets under real physical constraints. The final parametric reconstruction and output can generate a manufacturing layout design file based on the Pareto optimal combination obtained through coupling simulation. For example, the final parametric reconstruction and output can output a GDSII format file and embed parameter tolerance boundary annotations for use in mask manufacturing and yield analysis, thereby achieving a seamless connection from virtual optimization to physical manufacturing and reducing trial-and-error risks.

[0081] Furthermore, this embodiment of the invention also proposes a storage medium storing a self-made glass sensor capacitor grid layout optimization program, which, when executed by a processor, implements the steps of the self-made glass sensor capacitor grid layout optimization method described above.

[0082] In addition, refer to Figure 3 This invention also proposes a self-made glass sensor capacitor grid layout optimization device, which includes: The data acquisition module 10 is used to divide the layout structure of the self-made glass sensor capacitor grid into regions using a multi-dimensional perception analysis method to obtain a multi-sensor region layout. It also collects historical capacitance change data, historical touch point coordinate data and corresponding historical pressure change rate data of each sensing region layout under various pressure loads to construct a sensing feature database. The parameter sensing module 20 is used to determine the basic capacitance response threshold, coordinate positioning accuracy threshold and pressure change rate sensing threshold of each sensing area based on the sensing feature database, and to analyze the influence of grid parameter changes on the multi-touch point pressure sensing accuracy based on the sensing feature database, so as to obtain the grid parameter sensitivity of each sensing area. The adjustment generation module 30 is used to acquire real-time capacitance change data, real-time touch point coordinate data and real-time pressure change rate data of each of the sensing area maps, and generate a grid parameter adjustment table for each of the sensing area maps when the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold or pressure change rate sensing threshold is exceeded. The correlation analysis module 40 is used to analyze the multiple correlations between the capacitance response, coordinate positioning and pressure rate perception of each of the sensing regions based on the sensing feature database, and extract grid parameters from the grid parameter adjustment table according to the multiple correlations to calculate the perception importance, so as to generate a set of grid parameters to be optimized for each of the sensing regions. The parameter adjustment module 50 is used to group the layout of each sensing region according to the sensitivity of each grid parameter to obtain a multi-level sensitivity layout region. It then constructs a mapping relationship between the grid parameter adjustment amount and the improvement amount of multi-touch point pressure sensing performance of each level sensitivity layout region through a regression tree, generates a progressive layout adjustment rule to optimize the set of grid parameters to be optimized for each sensing region layout, and controls the parameterized reconstruction of the layout of the self-made glass sensor capacitor grid based on the optimization results.

[0083] Other embodiments or specific implementations of the self-made glass sensor capacitor grid layout optimization device described in this invention can be referred to the above-described method embodiments, and will not be repeated here.

[0084] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0085] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. In the module claims that list several devices, several of these devices may be embodied by the same hardware item. The use of the terms first, second, and third, etc., does not indicate any order and can be interpreted as names.

[0086] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as a read-only memory image (ROM) / random access memory (RAM), magnetic disk, optical disk), and includes several instructions to cause a terminal user device (which may be a mobile phone, computer, server, air conditioner, or network user device, etc.) to execute the methods described in the various embodiments of the present invention.

[0087] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for optimizing the capacitive grid layout of a self-made glass sensor, characterized in that, The method includes: The layout structure of the capacitor grid of the self-made glass sensor is divided into regions by multi-dimensional sensing analysis method to obtain a multi-sensing region layout. Historical capacitance change data, historical touch point coordinate data and corresponding historical pressure change rate data of each sensing region layout under various pressure loads are collected to construct a sensing feature database. Based on the perception feature database, the basic capacitance response threshold, coordinate positioning accuracy threshold, and pressure change rate perception threshold of each perception area are determined. Based on the perception feature database, the influence of grid parameter changes on the multi-touch point pressure perception accuracy of each perception area is analyzed to obtain the grid parameter sensitivity of each perception area. The system acquires real-time capacitance change data, real-time touch point coordinate data, and real-time pressure change rate data for each of the sensing area maps. When the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold, or pressure change rate sensing threshold is exceeded, a grid parameter adjustment table for each of the sensing area maps is generated. Based on the perception feature database, the multiple correlations between capacitance response, coordinate positioning and pressure rate perception of each perception area are analyzed, and the grid parameters are extracted from the grid parameter adjustment table according to the multiple correlations to calculate the perception importance, so as to generate a set of grid parameters to be optimized for each perception area. Based on the sensitivity of each grid parameter, the layout of each sensing region is grouped by sensitivity to obtain multi-level sensitivity layout regions. A regression tree is used to construct a mapping relationship between the grid parameter adjustment amount and the improvement amount of multi-touch point pressure sensing performance of each level of sensitivity layout region. A progressive layout adjustment rule is generated to optimize the set of grid parameters to be optimized in each sensing region layout. Based on the optimization results, the layout of the self-made glass sensor capacitor grid is parametrically reconstructed.

2. The method for optimizing the capacitor grid layout of a self-made glass sensor as described in claim 1, characterized in that, The method of dividing the layout structure of the self-made glass sensor capacitor grid into regions using a multi-dimensional sensing analysis method to obtain a multi-sensing region layout includes: The electrode density, cell capacitance base value, mutual coupling coefficient and signal-to-noise ratio of each sub-region in the capacitor grid layout are obtained to construct the layout complexity quantization matrix of each sub-region. Based on the complex quantification matrix of each of the aforementioned layouts, a perceptual directed graph is established, and clustering operation is performed on the mean and peak values ​​of capacitance change amplitude of each of the aforementioned sub-regions under pressure load to perform preliminary partitioning of each of the aforementioned sub-regions and obtain capacitance response grouping results. The touch point coordinate offset data corresponding to each sub-region is obtained, and the coordinate offset data is compared with the preset positioning tolerance threshold. The capacitive response grouping results are then refined and adjusted according to the comparison results to obtain a multi-sensing region map.

3. The method for optimizing the capacitive grid layout of a self-made glass sensor as described in claim 1, characterized in that, The basic capacitance response threshold includes the baseline capacitance value, the lower limit of the dynamic range, and the upper limit of saturation; the coordinate positioning accuracy threshold includes the minimum resolvable distance and the linearity error limit; the pressure change rate sensing threshold includes the minimum perceptible rate gradient and the maximum following rate.

4. The method for optimizing the capacitive grid layout of a self-made glass sensor as described in claim 3, characterized in that, The determination of the basic capacitance response threshold, coordinate positioning accuracy threshold, and pressure change rate sensing threshold for each of the sensing regions based on the sensing feature database includes: The baseline capacitance sampling value, capacitance saturation value and nonlinear distortion point of each sensing area in the non-touch state are extracted from the sensing feature database. The static response curve of capacitance-pressure is fitted by Gaussian process regression to determine the baseline capacitance value, lower limit of dynamic range and upper limit of saturation of each sensing area. Extract multiple coordinate sampling data of standard test points for each of the sensing area maps from the sensing feature database, calculate the root mean square error and linear fitting residual between the data and the theoretical coordinates, and determine the minimum resolvable distance and linearity error limit of each of the sensing area maps by confidence interval estimation. Capacitive response delay data and rate tracking error of each sensing region under a preset pressure ramp signal are extracted from the sensing feature database. The measurement noise is smoothed by a Kalman filter and the system response bandwidth is identified to determine the minimum perceptible rate gradient and maximum following rate of each sensing region.

5. The method for optimizing the capacitor grid layout of a self-made glass sensor as described in claim 1, characterized in that, The analysis of the impact of grid parameter changes on the multi-touch point pressure sensing accuracy in each sensing area based on the sensing feature database, to obtain the grid parameter sensitivity of each sensing area, includes: Historical touch point coordinate data, historical pressure change rate data and corresponding historical grid parameter change records of each of the perception area maps are extracted from the perception feature database and used as training samples to train the temporal convolutional network to obtain a response model that represents the dynamic relationship between grid parameter changes and multi-dimensional perception performance in each of the perception area maps. Based on the response model, the coordinate positioning error change and pressure rate tracking error change caused by the change of specific grid parameters in each of the sensing area maps under the preset pressure mode sequence are statistically analyzed. Spearman rank correlation analysis is performed on the error change according to the pressure change rate interval to obtain the parameter-performance correlation strength index of each of the sensing area maps. The corresponding parameter-performance correlation strength index is weighted and fused using the preset functional weights of each perception region map in the global perception task. The fusion result is then divided into intervals according to the preset sensitivity grading threshold to obtain a grid parameter sensitivity grading table to quantitatively determine the grid parameter sensitivity of each perception region map.

6. The method for optimizing the capacitive grid layout of a self-made glass sensor as described in claim 1, characterized in that, The process involves analyzing the multiple correlations between capacitive response, coordinate positioning, and pressure rate sensing of each sensing region based on the sensing feature database, and extracting grid parameters from the grid parameter adjustment table according to these multiple correlations to calculate sensing importance, thereby generating a set of grid parameters to be optimized for each sensing region. This includes: Historical capacitance response sequences, historical coordinate positioning data, and historical sensing performance evaluation data of each sensing region map under different pressure change rates are extracted from the sensing feature database. The similarity of sensing behavior under different pressure scenarios is calculated by dynamic time warping algorithm. Based on the similarity calculation results, each sensing region map is clustered for pressure scenarios to obtain a set of typical pressure sensing scenarios. The mean, variance, extreme values, and covariance matrix of the capacitive response characteristic values, coordinate stability index, and rate tracking accuracy within each set of typical pressure sensing scenarios are calculated to construct a multimodal sensing feature tensor. The partial least squares regression method is used to quantify the complex correlation between the layout mesh parameters and the multimodal sensing performance index, thereby obtaining the multivariate coupling relationship matrix of each sensing area layout. Based on the multivariate coupling relationship matrix, candidate adjustment parameters are extracted from the grid parameter adjustment table of each of the sensing regions. The contribution of the extracted candidate grid parameters to the overall sensing performance improvement is quantified by the random forest algorithm. The sensing importance of the extracted candidate grid parameters is quantified based on the contribution measure. Candidate grid parameters whose importance calculation results exceed the importance threshold are selected to form the set of grid parameters to be optimized for each of the sensing regions.

7. The method for optimizing the capacitive grid layout of a self-made glass sensor as described in claim 1, characterized in that, The process involves grouping the sensor area layout according to the sensitivity of each grid parameter to obtain multi-level sensitivity layout regions. A regression tree is then used to construct a mapping relationship between the grid parameter adjustment amount and the improvement in multi-touch point pressure sensing performance for each level of sensitivity layout region. A progressive layout adjustment rule is generated to optimize the set of grid parameters to be optimized for each sensor area layout. Based on the optimization results, the layout of the self-made glass sensor capacitor grid is then parametrically reconstructed. This process includes: The performance index time series data of the self-made glass sensor with optimized layout parameters are obtained under various pressure change rate test modes. The performance index time series data is used as input features to train the conditional generative adversarial network. The trained network is used to generate high-fidelity virtual pressure touch signals in real time and simulate sensing response to obtain virtual perception response data. Based on the virtual sensing response data, the parameter tolerance boundary of each sensing area layout under extreme pressure change rate is obtained as a stability constraint condition, and a virtual layout simulation environment containing four sets of virtual mesh parameters including electrode width, spacing, shape factor and interlayer coverage area is constructed by parametric scanning method. The optimization objective is to minimize the weighted sum of the comprehensive accuracy of multi-touch pressure sensing, which includes coordinate error, pressure value error, and rate tracking error. Multi-objective Bayesian optimization is used to collaboratively optimize the virtual mesh parameters of each sensing region in the virtual landscape simulation environment to obtain the Pareto optimal set of virtual parameter combinations for each sensing region. High-precision electromagnetic-structural coupling simulation is performed on the capacitor grid based on each Pareto optimal virtual parameter combination set. When the sensing performance index data obtained from the simulation is stable within the preset performance target range, the layout of the self-made glass sensor capacitor grid is controlled according to the optimized virtual parameter combination set for final parameterization reconstruction and output before manufacturing.

8. A self-made glass sensor capacitor grid layout optimization device, characterized in that, The self-made glass sensor capacitance grid layout optimization device includes: The data acquisition module is used to divide the layout structure of the capacitor grid of the self-made glass sensor into regions using a multi-dimensional perception analysis method, obtain a multi-sensor region layout, and collect historical capacitance change data, historical touch point coordinate data and corresponding historical pressure change rate data of each sensing region layout under various pressure loads, so as to construct a sensing feature database. The parameter sensing module is used to determine the basic capacitance response threshold, coordinate positioning accuracy threshold, and pressure change rate sensing threshold of each sensing area based on the sensing feature database, and to analyze the impact of grid parameter changes on the multi-touch point pressure sensing accuracy based on the sensing feature database, thereby obtaining the grid parameter sensitivity of each sensing area. The adjustment generation module is used to obtain real-time capacitance change data, real-time touch point coordinate data and real-time pressure change rate data of each of the sensing area maps, and generate a grid parameter adjustment table for each of the sensing area maps when the corresponding basic capacitance response threshold, coordinate positioning accuracy threshold or pressure change rate sensing threshold is exceeded. The correlation analysis module is used to analyze the multiple correlations between the capacitance response, coordinate positioning and pressure rate perception of each of the sensing regions based on the sensing feature database, and extract grid parameters from the grid parameter adjustment table according to the multiple correlations to calculate the perception importance, so as to generate a set of grid parameters to be optimized for each of the sensing regions. The parameter adjustment module is used to group the layout of each sensing region according to the sensitivity of each grid parameter to obtain a multi-level sensitivity layout region. It constructs a mapping relationship between the grid parameter adjustment amount and the improvement amount of multi-touch point pressure sensing performance of each level sensitivity layout region through regression tree, generates progressive layout adjustment rules to optimize the set of grid parameters to be optimized in each sensing region layout, and controls the parameterized reconstruction of the layout of the self-made glass sensor capacitor grid based on the optimization results.

9. A self-made glass sensor capacitance grid layout optimization device, characterized in that, The self-made glass sensor capacitor grid layout optimization device includes: a memory, a processor, and a self-made glass sensor capacitor grid layout optimization program stored in the memory and executable on the processor. The self-made glass sensor capacitor grid layout optimization program is configured to implement the steps of the self-made glass sensor capacitor grid layout optimization method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a self-made glass sensor capacitor grid layout optimization program, which, when executed by a processor, implements the steps of the self-made glass sensor capacitor grid layout optimization method as described in any one of claims 1 to 7.

Images