Touch screen multi-touch mode switching method, device and equipment

Abstract

The present application relates to the technical field of touch screens, and discloses a method, device and equipment for switching multiple touch modes of a touch screen, the method comprising: performing minimum error processing on a touch signal collected by a touch screen to obtain a target feature component; inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain current touch mode features; performing attenuation characteristic analysis and adaptive threshold adjustment on the target feature component to obtain a state transition parameter; performing delayed response modulation and regularized zero interference on a touch point response time, response intensity and response area to obtain a screen response change feature; performing fuzzy logic decision-making for continuous sampling cycles to obtain a feasibility determination result of touch mode switching; performing closed-loop control on a touch screen response parameter to obtain a mode switching delay compensation control amount, thereby improving the reliability of touch mode switching, avoiding the occurrence of false triggering, and significantly reducing the response delay during the mode switching process.

Description

Technical Field

[0001] The present application relates to the field of touch screen technology, and in particular to a touch screen multi-touch mode switching method, device and equipment. Background Art

[0002] As an important interface for human-computer interaction, touch screens and their multi-touch functions are widely used in modern smart devices. However, in actual use, switching between multi-touch modes often causes problems such as delays and false touches, which not only affects the user experience but may also lead to operational errors. Traditional touch mode switching methods mainly rely on simple contact detection and threshold judgment, which cannot effectively handle complex touch scenarios and fast mode conversion requirements.

[0003] The current touch mode switching technology has three main technical bottlenecks: the first is the problem of accurate recognition of touch signals. Due to environmental interference and hardware limitations, touch signals often contain noise and artifacts, resulting in inaccurate mode discrimination; the second is the delay problem of mode switching. Traditional switching algorithms require multiple sampling and verification to confirm mode conversion, resulting in obvious response delays; the last is the stability problem of the switching process. In continuous and rapid touch operations, frequent mode switching may lead to unstable system response. Summary of the invention

[0004] The present application provides a method, device and equipment for switching multiple touch modes of a touch screen, thereby improving the reliability of touch mode switching, avoiding the occurrence of false triggering, and significantly reducing the response delay during the mode switching process.

[0005] A first aspect of the present application provides a touch screen multi-touch mode switching method, the touch screen multi-touch mode switching method comprising:

[0006] Perform minimum error processing on the touch signal collected by the touch screen to obtain the target characteristic component;

[0007] Inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain current touch mode features;

[0008] Performing time dimension attenuation characteristic analysis and adaptive threshold adjustment on the target characteristic component to obtain a state transfer parameter;

[0009] Based on the current touch mode characteristics, delayed response modulation and regularized zero interference are performed on the touch point response time, response intensity and response area to obtain screen response change characteristics;

[0010] Perform fuzzy logic decision making in continuous sampling cycles according to the state transfer parameters and the screen response change characteristics to obtain a touch mode switching feasibility determination result;

[0011] Based on the touch mode switching feasibility determination result, a closed-loop control is performed on the touch screen response parameters to obtain a mode switching delay compensation control amount.

[0012] A second aspect of the present application provides a touch screen multi-touch mode switching device, the touch screen multi-touch mode switching device comprising:

[0013] A processing module, used for performing minimum error processing on the touch signal collected by the touch screen to obtain a target characteristic component;

[0014] A mapping module, used for inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain a current touch mode feature;

[0015] An adjustment module, used for performing time dimension attenuation characteristic analysis and adaptive threshold adjustment on the target characteristic component to obtain a state transfer parameter;

[0016] A modulation module, used to perform delayed response modulation and regularized zero interference on the touch point response time, response intensity and response area based on the current touch mode characteristics, so as to obtain screen response change characteristics;

[0017] A decision module, used to make fuzzy logic decisions in continuous sampling cycles according to the state transfer parameters and the screen response change characteristics, and obtain a feasibility determination result of the touch mode switching;

[0018] The control module is used to perform closed-loop control on the touch screen response parameters based on the touch mode switching feasibility determination result to obtain the mode switching delay compensation control amount.

[0019] The third aspect of the present application provides an electronic device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes the above-mentioned touch screen multi-touch mode switching method.

[0020] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and when the computer-readable storage medium is run on a computer, the computer executes the above-mentioned touch screen multi-touch mode switching method.

[0021] Compared with the prior art, the present application has the following beneficial effects: through the combination of minimum error processing and multi-layer touch mode classifiers, the accuracy of feature extraction of touch signals is improved, and the error rate of touch mode recognition is effectively reduced. The time dimension attenuation characteristic analysis and adaptive threshold adjustment mechanism are adopted to solve the problem of signal instability during touch mode switching and enhance the system's resistance to environmental interference. The introduction of delayed response modulation and regularized zero interference technology effectively eliminates interference signals in the touch point response process and ensures the accuracy of touch response. The continuous sampling cycle judgment mechanism based on fuzzy logic decision-making improves the reliability of touch mode switching and avoids the occurrence of false triggering. The closed-loop control strategy and delay compensation mechanism are adopted to significantly reduce the response delay during mode switching and optimize the user operation experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0024] Figure 1 It is a schematic diagram of the flow of a method for switching multiple touch modes of a touch screen provided by an embodiment of the present invention;

[0025] Figure 2 is a schematic block diagram of the structure of a touch screen multi-touch mode switching device provided by an embodiment of the present invention;

[0026] Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. Figure 1 In the embodiment of the present application, an embodiment of the touch screen multi-touch mode switching method includes:

[0031] Step 100: performing minimum error processing on the touch signal collected by the touch screen to obtain a target characteristic component;

[0032] It is understandable that the execution subject of the present application may be a touch screen multi-touch mode switching device, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0033] Specifically, the touch point position is collected in real time through the sensor of the touch screen, and an initial sampling matrix containing the two-dimensional coordinates (x, y) of the touch point, pressure value and timestamp is generated to record key information such as the spatial position distribution, pressure intensity and occurrence time of the touch operation. In order to ensure the continuity and effectiveness of the data, the initial sampling matrix is ​​updated at a fixed sampling frequency to capture the dynamic changes of the touch point and generate sufficient timing information. The initial sampling matrix is ​​input into the sliding window segmentation module to divide the continuous touch data into multiple windows of fixed length according to the time dimension, reduce the random error of the single sampling data and enhance the ability to extract local signal features. In each window, the multi-dimensional signal feature sequence of the touch point is extracted, including the coordinate trajectory of the touch point, the pressure value change and the time interval characteristics, effectively capturing the spatial and temporal characteristics of the touch behavior. The touch pressure value is calculated by the least squares method to extract the principal component characteristics of the touch pressure. Based on the relationship between the change of the touch pressure value and time, a linear model is constructed, and the principal component trend of the pressure value is obtained by fitting the least squares method. The principal component features reflect the global change pattern of touch pressure and can significantly reduce the errors caused by short-term fluctuations or noise in touch data. According to the principal component features of touch pressure, they are applied to the multidimensional touch signal feature sequence to achieve dimensionality reduction mapping. Through dimensionality reduction methods such as principal component analysis, the redundant information of the multidimensional feature sequence is compressed and mapped to a low-dimensional feature space to obtain the spatial distribution characteristics of the touch points and reveal the distribution pattern of the touch points on the screen. According to the spatial distribution characteristics of the touch points, dynamic time warping is performed to align and regularize the temporal changes of the touch points to eliminate the deformation of the touch signals caused by speed differences or uneven time distribution in different sampling periods. Through the dynamic time warping algorithm, the change trajectory of the touch points can be standardized in the time dimension to obtain the temporal change characteristics that describe the dynamic behavior of the touch points. The principal component features of touch pressure and the temporal change features of the touch points are fused to generate a multimodal feature vector. The methods of multimodal feature fusion include weighted averaging, feature concatenation, or feature sharing layers based on deep learning. The specific choice depends on the task requirements and hardware capabilities. The fused feature vector integrates the multi-dimensional information of the touch signal in space, time and intensity, and has stronger expressive ability.

[0034] Step 200: Input the target feature component into the multi-layer touch mode classifier for feature mapping to obtain the current touch mode feature;

[0035] Specifically, the target feature component is subjected to a touch response mapping conversion, and a feature vector including a touch point coordinate matrix, a pressure value sequence, and a timestamp mark is constructed. The core of the touch response mapping conversion is to uniformly format the original touch signal so that it can adapt to the input requirements of the subsequent classifier. The touch point coordinate matrix describes the distribution position of the touch point on the screen, while the pressure value sequence reflects the pressure change applied by each touch point, and the timestamp mark records the dynamic time characteristics of the signal. The feature vector is input into the first layer classifier of the multi-layer touch pattern classifier, which focuses on the maximum response transfer mapping processing based on the touch point. The first layer classifier adopts a convolutional neural network structure, including 3 convolutional layers and 2 pooling layers. Each convolutional layer adopts the ReLU activation function, and the nonlinear activation method can effectively enhance the expression ability of the model while maintaining the efficiency of calculation. The convolution layer constructs the distribution feature map of the touch point by extracting the spatial local pattern of the feature vector, while the pooling layer improves the stability of the feature and the generalization ability of the model by reducing the feature dimension and denoising. Through the processing of this layer, a feature map containing the touch point distribution information is generated. The touch point distribution feature map is input into the second-layer classifier of the multi-layer touch pattern classifier for zero interference elimination, thereby extracting a purer feature matrix. The second-layer classifier adopts a deep residual network structure enhanced by the attention mechanism. Its design includes 4 residual blocks, each of which consists of two convolutional layers and a short-circuit connection. In the connection mode of the residual blocks, the short-circuit connection avoids the gradient vanishing problem by directly transmitting the input information, while retaining the original information of the input features, thereby enhancing the training performance of the network. A self-attention layer is added between each residual block, so that the model can dynamically focus on the most important information area in the feature matrix, thereby effectively suppressing the interference components in the signal and generating an interference-free feature matrix. The non-interference feature matrix is ​​subjected to temporal feature extraction to extract the temporal features of the touch pattern. The temporal features of the touch pattern capture the dynamic change law of the touch signal, such as the movement trajectory of the touch point over time, the pressure change trend and other key information. The temporal features are input into the third-layer classifier of the multi-layer touch pattern classifier, which focuses on multimodal feature fusion and adopts a multi-head self-attention mechanism to achieve more efficient information integration. The classifier contains 4 attention heads, each with a dimension of 64. The multi-head structure processes information of different dimensions in parallel to enhance the model's ability to capture complex touch signals. Generate a fused feature vector that reflects the spatial, temporal, and intensity information of the touch signal. Perform dimensionality reduction mapping on the fused feature vector, reduce redundant information through the dimensionality reduction process, and generate a probability distribution of the touch pattern. Based on the probability distribution, calculate the confidence score and duration feature to quantify the certainty and stability of the touch pattern. By setting a reasonable threshold, determine the category of the touch pattern, and combine the touch pattern category with its corresponding confidence score and duration feature to describe the current touch pattern characteristics.

[0036] Step 300: Analyze the time dimension attenuation characteristics of the target characteristic component and adjust the adaptive threshold to obtain the state transfer parameter;

[0037] It should be noted that the target characteristic component is subjected to time series analysis to construct an attenuation characteristic matrix including touch signal strength, attenuation rate and steady-state value. The touch signal strength is calculated based on the change in the pressure applied by the touch point, and the attenuation rate is obtained by analyzing the slope of the signal strength over time to describe the dynamic change trend of the signal. The steady-state value is the characteristic value of the touch signal tending to be stable under long-term observation, which is used to reflect the stability characteristics of the signal. By integrating this information into the attenuation characteristic matrix, a multi-dimensional feature representation is constructed. The signal strength in the attenuation characteristic matrix is ​​gradient calculated to obtain the attenuation curve parameters of the touch signal, which are used to characterize the change form of the signal in the time dimension. Through numerical gradient analysis based on time series, the rapid change area and change trend of the signal strength are captured. These gradient information are input into the attenuation fitting unit, which uses high-order polynomial fitting or exponential attenuation model to construct a complete attenuation curve. The key to the fitting process is to extract the inflection points in the attenuation curve, which represent the critical value of the signal from rapid change to gradual stability, called the attenuation critical value. The attenuation critical value is an important dividing point for the dynamic change of the signal. According to the attenuation critical value, the attenuation feature matrix is ​​divided into time periods to generate a segmented attenuation feature sequence. This division strategy can accurately distinguish different change stages of the touch signal, such as the rapid change stage, the stable change stage, and the steady-state stage. In each time period, a sliding average calculation is performed on the steady-state value sequence to smooth the signal noise and extract a stable dynamic threshold adjustment benchmark. The sliding average calculation is based on a sliding operation with a fixed window length. By calculating the average of multiple data points in the sequence, the impact of short-term fluctuations on the overall trend is reduced. Based on the dynamic threshold adjustment benchmark, the upper and lower threshold ranges and the adjustment step length are calculated to generate an adaptive threshold parameter. The calculation of the adjustment range comprehensively considers the overall change amplitude of the signal and the fluctuation of the steady-state value, while the adjustment step length is adjusted according to the dynamic change rate of the touch signal. Through this dynamic adjustment method, the adaptive threshold can adapt to the characteristic changes of the touch signal in real time, improve the flexibility and accuracy of the touch mode switching, and the generated adaptive threshold parameters can provide personalized signal processing strategies for different touch operations. The adaptive threshold parameters are matched with the attenuation feature matrix to extract the state feature vector. Dynamic thresholds are applied to different dimensions in the attenuation feature matrix to filter out signal states that meet specific conditions. The state feature vector is normalized to eliminate the dimensional differences between different dimensions and ensure the consistency and stability of the expression of state features. The normalized state feature vector is defined as a state transition parameter, which can accurately describe the change law of the touch signal in the time dimension and provide a reliable decision-making basis for the real-time switching of the touch mode.

[0038] Step 400: Based on the current touch mode characteristics, delayed response modulation and regularized zero interference are performed on the touch point response time, response intensity and response area to obtain screen response change characteristics;

[0039] Specifically, the current touch mode feature is feature deconstructed to separate a response feature set including a response time series, a response strength value, and a response area data. The response time series records the trigger delay of the touch point at different times, while the response strength value reflects the pressure or signal strength applied by the touch point on the screen. The response area data describes the contact range of the touch point on the screen. Feature deconstruction is to parse the complex touch mode features into basic data units that are easy to analyze separately. The response feature set is input into the delay response modulation unit to perform delay modeling and generate delay model parameters including response delay parameters and response compensation coefficients. In the delay modeling process, the delay model of the touch point response is established by analyzing the dynamic characteristics of the response time series. The model can describe the time difference from the touch point signal triggering to the system response, and extract the key response delay parameters by analyzing the transmission path and processing bottleneck of the touch signal. At the same time, in order to compensate for the touch error caused by the delay, the corresponding response compensation coefficient is calculated in combination with the touch strength and area characteristics to reduce the inconsistency caused by the delay at the system level. The maximum response transmission mapping based on the touch point is performed on the response time series to generate a touch point delay distribution map. The touch point delay distribution map reflects the delay characteristics of the touch signal in different touch areas by mapping the relationship between the spatial position of the touch point and its response time. Based on this distribution map, the response time compensation value is calculated to generate the delay compensation parameter. The core function of the delay compensation parameter is to dynamically adjust the response behavior of the touch point so that the touch signal is more consistent and accurate in the time dimension and optimize the user experience. At the same time, the response intensity value and response area data are subjected to regularization zero-interference operation to eliminate random errors and environmental noise in the signal. Regularization processing maps the signal to a unified numerical interval by standardizing the data range, ensuring that touch signals of different intensities or areas can be compared and processed on the same scale. Zero-interference operation uses filtering algorithms and suppression strategies to remove abnormal signals or unnecessary fluctuations and generate interference-free response features. Based on the delay compensation parameters and interference-free response features, feature reconstruction is performed to generate a response modulation matrix. Time series sampling is performed on the response modulation matrix to generate a response change sequence. Time series sampling constructs the change trajectory of the touch signal in the time dimension by extracting key data points in the modulation matrix within a fixed time interval. The response change gradient is calculated based on the response change sequence, and the dynamic characteristics of the touch response are quantified using the gradient change. This gradient reflects the change rate and direction of the touch point in different time periods. The above processing steps generate the screen response change characteristics, reflecting the dynamic change rules of the touch signal in time, intensity and area.

[0040] Step 500: Perform fuzzy logic decision making for continuous sampling cycles according to the state transfer parameters and the screen response change characteristics to obtain a feasibility determination result of the touch mode switching;

[0041] Specifically, the state transition parameters and the screen response change characteristics are input into the fuzzy logic system to form a fuzzy rule input matrix including the signal attenuation ratio, the response change rate, and the current mode confidence. The state transition parameters describe the dynamic changes of the touch signal in the time dimension, such as the stability and mutation characteristics of the signal, while the screen response change characteristics reflect the real-time response performance of the touch point, such as the response speed and intensity change. After combining these features, a high-dimensional input matrix is ​​constructed to describe the changes in the current touch environment and the potential demand for mode switching. The membership function calculation is performed on the signal attenuation ratio in the fuzzy rule input matrix to generate an attenuation ratio membership map. The membership function is the core part of the fuzzy logic system. By assigning different membership values ​​to the signal attenuation ratio, the continuous signal characteristics are mapped to the fuzzy set. For example, the attenuation ratio is divided into fuzzy sets such as "low", "medium", and "high", and each type is described by a trapezoidal or triangular membership function. At the same time, a similar membership function calculation is performed on the response change rate in the input matrix to generate a response change rate membership map, revealing the dynamic characteristics of the touch point response performance. According to the attenuation ratio membership mapping and the response change rate membership mapping, a fuzzy inference rule base is constructed to generate fuzzy decision values. The fuzzy inference rule base consists of a set of predefined rules, such as "If the signal attenuation ratio is 'high' and the response change rate is 'low', the switching requirement is 'unnecessary'", or "If the signal attenuation ratio is 'medium' and the response change rate is 'high', the switching requirement is 'necessary'". The rule base performs fuzzy inference on the input membership value, converting the complex signal characteristics into easy-to-understand fuzzy decision values, indicating the degree of need for touch mode switching. The stability of the fuzzy decision value is calculated in continuous sampling cycles to evaluate the change trend and consistency of the signal at different times. The stability calculation is to generate a stability score by analyzing the fluctuation of the fuzzy decision value in multiple sampling cycles, quantifying its degree of stability over time. The higher the stability score, the smoother the signal change and the stronger the system's confidence in mode switching. The stability score is input into a predefined time filter with a 10ms time constant for stability processing to filter out noise and fluctuations in a short time. The time filter converts the rapidly changing decision value into a more stable switching index by smoothing the stability score, thereby improving the robustness of the system and the reliability of decision making. The filtered switching index can more accurately reflect the real changes in the touch environment. The feasibility judgment calculation is performed on the filtered switching index to generate a switching feasibility score. The feasibility judgment is to convert the switching index into a specific feasibility score by comprehensively analyzing the numerical range and trend of the switching index to quantify the possibility and suitability of the touch mode switching. The switching feasibility score is compared with a preset threshold to generate a touch mode switching feasibility judgment result. If the switching feasibility score exceeds the threshold, it is determined that the mode switching is feasible; otherwise, it is determined that the current mode is maintained.

[0042] Step 600: Based on the feasibility determination result of the touch mode switching, a closed-loop control is performed on the touch screen response parameter to obtain a mode switching delay compensation control amount.

[0043] Specifically, the feasibility determination result of the touch mode switching is input into the fuzzy predefined time response controller to complete the nonlinear feature modeling and generate the feature model parameters including the touch state variable and the quantized touch signal. The feasibility determination result of the touch mode switching reflects the degree of demand for mode switching in the current touch environment, and through the modeling process of the fuzzy controller, the complex nonlinear relationship is abstracted into a feature parameterized expression. Among them, the touch state variable is used to describe the global characteristics of the touch operation, such as the number, distribution and change trend of the touch points; the quantized touch signal characterizes the local characteristics of the touch point response, including the response intensity and time change. The feature model parameters are adaptively recursively operated to generate a control benchmark sequence including performance function values ​​and response time constraints. In the adaptive recursive operation process, the performance function value is continuously optimized according to the real-time touch signal changes by dynamically adjusting the calculation parameters, and the response time is ensured to meet the preset constraints. The performance function value is used to quantify the optimal degree of system response, and the response time constraint ensures the real-time and smoothness of the touch operation by controlling the time range. According to the control benchmark sequence, a feedforward compensation unit based on fuzzy logic is constructed to generate a feedforward control quantity including a compensation direction vector and a compensation amplitude coefficient. In the feedforward compensation unit, the current touch state and quantization signal are analyzed in combination with fuzzy logic reasoning rules to determine the appropriate compensation direction vector to adjust the response trend of the system; at the same time, the compensation amplitude coefficient is calculated to ensure that the size of the compensation amount can adapt to the change amplitude of the current touch signal. The core role of the feedforward control amount is to eliminate potential response delay problems through predictive adjustment, so as to achieve parameter optimization before the mode switch occurs. The feedforward control amount is input into the predefined time filter and repeated operations are eliminated to generate an optimized compensation sequence. The time filter removes redundant signals and short-term fluctuations by smoothing the feedforward control amount and ensures that the compensation sequence has high stability. Closed-loop differential calculation is performed on the optimized compensation sequence and the real-time response parameters of the touch screen to extract the feedback correction amount containing real-time deviation and cumulative error. The real-time deviation represents the instantaneous difference between the current touch screen response and the expected response, while the cumulative error reflects the overall deviation of the system over a period of time. The error of the system response is quantified through differential calculation. Based on the feedback correction amount, the feedforward control amount is iteratively optimized to generate closed-loop control parameters. In the iterative optimization process, the feedforward control strategy is dynamically adjusted in combination with the feedback correction, so that the system can gradually reduce the response error and approach the optimal state. The closed-loop control parameters are input into the delay compensation controller to complete the response characteristic prediction and generate the prediction compensation sequence. By modeling and predicting the delay characteristics of the touch signal, the delay problem is estimated in advance and actively compensated. Stability constraints are also applied to the prediction compensation sequence to ensure that its output meets the real-time and reliability requirements of the system, and finally the mode switching delay compensation control amount is generated.

[0044] In the embodiment of the present application, by combining minimum error processing and multi-layer touch mode classifiers, the accuracy of feature extraction of touch signals is improved, and the error rate of touch mode recognition is effectively reduced. The time dimension attenuation characteristic analysis and adaptive threshold adjustment mechanism are adopted to solve the problem of signal instability during touch mode switching and enhance the system's resistance to environmental interference. The introduction of delayed response modulation and regularized zero interference technology effectively eliminates interference signals in the touch point response process and ensures the accuracy of touch response. The continuous sampling cycle judgment mechanism based on fuzzy logic decision-making improves the reliability of touch mode switching and avoids the occurrence of false triggering. The closed-loop control strategy and delay compensation mechanism are adopted to significantly reduce the response delay during mode switching and optimize the user operation experience.

[0045] In a specific embodiment, the process of executing step 100 may specifically include the following steps:

[0046] Perform touch signal acquisition on the touch point position of the touch screen to obtain an initial sampling matrix including the touch point xy coordinates, pressure value and time stamp;

[0047] Sliding window segmentation is performed on the initial sampling matrix to obtain a multi-dimensional touch signal feature sequence, and the pressure values ​​in the multi-dimensional touch signal feature sequence are calculated by least square method to obtain the touch pressure principal component feature;

[0048] According to the principal component characteristics of the touch pressure, the multi-dimensional touch signal feature sequence is mapped to a dimension reduction to obtain the spatial distribution characteristics of the touch points, and the spatial distribution characteristics of the touch points are dynamically time-warped to obtain the temporal change characteristics of the touch points;

[0049] The touch pressure principal component features and the touch point temporal change features are fused to obtain a multimodal feature vector, and the multimodal feature vector is iteratively calculated by error minimization to obtain the target feature component.

[0050] Specifically, the touch point position on the touch screen is used to perform signal acquisition and construct an initial sampling matrix For each touch event, collect the touch point coordinate, Coordinates, pressure values , and the timestamp of when the touch occurred . Assume that during a sampling process, touch points, the initial sampling matrix It is expressed as:

[0051] ;

[0052] in, and It is The screen coordinates of the touch point, is the corresponding pressure value, is the timestamp of the touch point. This information constitutes the basic data set of the touch operation. Perform sliding window segmentation. Divide the continuous touch data into several time windows of fixed length, and process the data in each window independently. Suppose the length of the sliding window is , the sliding step length is , then each segment of data generated by the sliding window is expressed as ,in is the sequence number of the window, expressed as:

[0053] ;

[0054] Through this operation, a multi-dimensional touch signal feature sequence in multiple windows is generated to capture the local dynamic characteristics of touch behavior. The pressure values ​​in the multi-dimensional touch signal feature sequence are calculated by the least squares method to extract the principal component characteristics of the touch pressure. Suppose the pressure value sequence in each window is , by fitting a linear model Describes the relationship between pressure and time. The goal of the least squares method is to minimize the following error function:

[0055] ;

[0056] Through Respectively and Find the partial derivative and set it to zero to get the optimal solution:

[0057] ;

[0058] in, is the slope of the pressure change, is the intercept. These two parameters together constitute the principal component characteristics of touch pressure and describe the changing trend of pressure signal. Based on the extracted principal component characteristics of touch pressure, the multi-dimensional touch signal feature sequence is mapped to generate the spatial distribution characteristics of touch points. The core of dimensionality reduction is to remove redundant information in the data and retain only key features. For example, the original high-dimensional touch data is projected into a low-dimensional space through principal component analysis. Let the spatial distribution feature matrix of touch points be , by calculating the covariance matrix And find its eigenvalue and eigenvector, map the original data to the principal component direction corresponding to the eigenvector, so as to achieve dimensionality reduction. Dynamic time warping (DTW) is performed on the generated touch point spatial distribution features to align the time series changes of the touch points. The goal of DTW is to find two time series and The optimal matching path between them is minimized by the dynamic programming algorithm:

[0059] ;

[0060] This matching path ensures the nonlinear alignment of the time series and generates the temporal variation characteristics of the touch points. The touch pressure principal component features are fused with the touch point temporal variation features to generate a multimodal feature vector. Feature fusion is achieved through feature concatenation, weighted averaging or deep learning methods. For example, the pressure features are combined into a and timing characteristics Concatenate into joint feature vector , in order to retain the key information of the two modes. Perform error minimization iterative calculation on the multimodal feature vector to optimize the final target feature component. Define the error function through the gradient descent method Represents the error between the feature vector and the expected target, adjusting the parameters To minimize the error:

[0061] ;

[0062] in, is the learning rate. After multiple iterations, the optimized target feature component is finally obtained to describe the core mode of touch behavior.

[0063] In a specific embodiment, the process of executing step 200 may specifically include the following steps:

[0064] Performing touch point response mapping conversion on the target feature component to obtain a feature vector including a touch point coordinate matrix, a pressure value sequence and a timestamp mark;

[0065] The feature vector is input into the first layer classifier of the multi-layer touch pattern classifier for maximum response transfer mapping based on the touch point. The first layer classifier adopts a convolutional neural network structure, which includes 3 convolutional layers and 2 pooling layers. Each convolutional layer uses a ReLU activation function to obtain a touch point distribution feature map.

[0066] The touch point distribution feature map is input into the second-layer classifier of the multi-layer touch pattern classifier for zero interference elimination. The second-layer classifier adopts a deep residual network structure enhanced by the attention mechanism, which includes 4 residual blocks. Each residual block contains two convolutional layers and a short-circuit connection. A self-attention layer is added between the residual blocks to obtain a non-interference feature matrix.

[0067] The non-interference feature matrix is ​​subjected to temporal feature extraction to obtain the touch mode temporal feature, and the touch mode temporal feature is input into the third layer classifier of the multi-layer touch mode classifier for multi-modal feature fusion. The third layer classifier adopts a multi-head self-attention mechanism, including 4 attention heads, each with a dimension of 64, to obtain a fusion feature vector.

[0068] The fused feature vector is mapped to a dimension reduction mode to obtain the touch pattern probability distribution, and the confidence score and duration feature are calculated based on the touch pattern probability distribution. The touch pattern category is obtained through threshold judgment, and the touch pattern category, confidence score and duration feature are combined to obtain the current touch pattern feature.

[0069] Specifically, the target feature components are transformed into touch point response mapping to construct a feature vector containing a touch point coordinate matrix, a pressure value sequence, and a timestamp mark. Assume that the touch point data contains touch points, each of which contains two-dimensional coordinates , pressure value and timestamp By combining this information, we construct the feature vector Indicates The feature vector set of the entire touch event is represented as a matrix :

[0070] ;

[0071] in are the screen coordinates of the touch point, is the pressure value, is the timestamp. Through mapping, the original touch signal is standardized into a unified feature format. The feature vector The first layer classifier input to the multi-layer touch pattern classifier performs maximum response transfer mapping based on the touch point. This classifier adopts a convolutional neural network structure, which includes 3 convolutional layers and 2 pooling layers. Assume that the input feature vector The dimension is , through the first convolution layer, the local spatial features of the touch point are extracted. The output of the convolution operation is expressed as:

[0072] ;

[0073] in is the output of the first convolutional layer. eigenvalues, is the convolution kernel weight, is the bias term, The ReLU activation function is defined as Through the progressive processing of 3 convolutional layers and 2 pooling layers, the distribution feature map of touch points is obtained. , which is used to describe the global distribution of touch points on the screen. The second layer classifier of the multi-layer touch pattern classifier is input for zero interference elimination. The classifier adopts a deep residual network structure enhanced by the attention mechanism, which contains 4 residual blocks. The operation of each residual block is expressed as:

[0074] ;

[0075] in is the output of the residual block, is the input feature map, is the convolution kernel weight, is the bias term. The residual structure retains the original feature information through short-circuit connections and combines the attention mechanism to give higher weights to the key parts of the feature map. The weight calculation of the attention mechanism is:

[0076] ;

[0077] in is the attention weight, is the attention score, is the parameter of the attention layer. By combining the residual block and the attention layer, the interference-free feature matrix is ​​generated. , effectively remove noise and redundant information. For the non-interference feature matrix Perform time series feature extraction to capture the dynamic change characteristics of the touch pattern. Through time series analysis methods, such as recursive neural networks or long short-term memory networks, extract the dependencies in the time dimension and generate touch pattern time series features. ,in Indicates The touch pattern temporal features are input into the third layer classifier of the multi-layer touch pattern classifier for multimodal feature fusion. This classifier adopts a multi-head self-attention mechanism to process different feature subspaces in parallel through multiple attention heads. are query, key, and value matrices respectively, then the output of each attention head is:

[0078] ;

[0079] in is the dimension of the key matrix. Through 4 attention heads, each with a dimension of 64, we get the fused feature vector , fully expressing the spatial, temporal and intensity characteristics of the touch pattern. Perform dimensionality reduction mapping to generate touch mode probability distribution ,in Representation Mode Calculate the confidence score based on the probability distribution and duration characteristics , through the threshold Determine the touch mode category:

[0080] ;

[0081] The current touch pattern feature is finally generated by combining the pattern category, confidence score and duration features.

[0082] In a specific embodiment, the process of executing step 300 may specifically include the following steps:

[0083] Performing a time series analysis on the target characteristic component to obtain a decay characteristic matrix including touch signal strength, decay rate, and steady-state value;

[0084] Performing gradient calculation on the signal strength in the attenuation feature matrix to obtain the touch signal attenuation curve parameters, and inputting the touch signal attenuation curve parameters into the attenuation fitting unit to extract the inflection point to obtain the attenuation critical value;

[0085] The attenuation characteristic matrix is ​​divided into time periods according to the attenuation critical value to obtain a segmented attenuation characteristic sequence, and a sliding average calculation is performed on the steady-state value sequence in the segmented attenuation characteristic sequence to obtain a dynamic threshold adjustment benchmark;

[0086] Based on the dynamic threshold adjustment benchmark, the upper and lower threshold ranges and the adjustment step length are calculated to obtain the adaptive threshold parameters;

[0087] The adaptive threshold parameter is matched with the attenuation characteristic matrix to obtain the state characteristic vector, and the state characteristic vector is normalized to obtain the state transfer parameter.

[0088] Specifically, a time series analysis is performed on the target feature component to construct an attenuation feature matrix including the touch signal strength, attenuation rate, and steady-state value. The intensity sequence of the touch signal is: ,in Indicates The signal strength at the moment. By performing time dimension differential calculation on the signal strength, the decay rate sequence is obtained. , where each Defined as:

[0089] ;

[0090] in is the time interval between adjacent sampling points. The steady-state value is calculated by identifying the average intensity of the signal after a long period of stability, and is defined as:

[0091] ;

[0092] in is the number of samples in the steady-state interval. , decay rate and steady-state value Combination to form the attenuation feature matrix :

[0093] ;

[0094] right The signal strength in the signal is calculated by gradient to obtain the attenuation curve parameters of the touch signal. By taking the derivative of the signal strength sequence, the gradient value represents the rate of change of the signal strength. The gradient calculation formula is defined as:

[0095] ;

[0096] Using the gradient value , fitting the signal attenuation curve Assuming that the signal attenuation process conforms to the exponential attenuation model, the curve fitting formula is:

[0097] ;

[0098] in is the initial signal strength, is the decay rate, is the steady-state value. Fitting and optimizing parameters , describing the dynamic characteristics of the signal. The fitted attenuation curve parameters are input into the attenuation fitting unit to extract the key inflection point of signal attenuation. The inflection point is defined as the critical value where the signal changes from the rapid change phase to the stable phase, and its mathematical expression is the position where the second-order derivative is equal to zero:

[0099] ;

[0100] Calculated by this formula Determine the turning point in the signal attenuation process. According to the attenuation threshold , for the attenuation feature matrix Divide the time period and generate a segmented attenuation feature sequence. Assume The time axis is divided into a fast decay segment and a stable segment, and the signal features of these two stages are extracted for subsequent analysis. In each time period, a sliding average calculation is performed on the steady-state value sequence to eliminate random fluctuations and smooth the signal. The sliding average formula is defined as:

[0101] ;

[0102] in is the sliding window size, Indicates The average value of the window. Based on the dynamic threshold adjustment benchmark obtained by sliding average calculation, the upper and lower threshold ranges and adjustment steps are calculated to generate adaptive threshold parameters. The upper and lower threshold ranges are achieved by adding or subtracting a fixed ratio from the dynamic benchmark value. The formula is:

[0103] ;

[0104] in is the standard deviation of the signal strength, and is the adjustment coefficient. The adjustment step is calculated by the rate of change of the signal and is defined as:

[0105] ;

[0106] in is the step-size proportional coefficient. The adaptive threshold parameter is matched with the attenuation feature matrix to generate the state feature vector. The goal of the matching operation is to filter out the signal features that meet the threshold range. The formula is:

[0107] ;

[0108] The state eigenvector is normalized to generate the state transfer parameter. The purpose of normalization is to eliminate the influence of the eigenvalue range so that the features can be compared at the same scale. The normalization formula is:

[0109] ;

[0110] Normalized state transition parameters Used to describe the dynamic state of touch signals.

[0111] In a specific embodiment, the process of executing step 400 may specifically include the following steps:

[0112] Deconstruct the current touch mode features to obtain a response feature set including a response time series, a response intensity value, and a response area data;

[0113] Inputting the response feature set into the delayed response modulation unit for delay modeling, and obtaining delay model parameters including response delay parameters and response compensation coefficients;

[0114] Performing a maximum response transmission mapping based on the contact on the response time sequence to obtain a contact delay distribution diagram, and calculating a response time compensation value according to the contact delay distribution diagram to obtain a delay compensation parameter;

[0115] Regularized zero-interference operation is performed on the response intensity value and response area data to obtain interference-free response characteristics;

[0116] Based on the delay compensation parameters and the interference-free response characteristics, feature reconstruction is performed to obtain a response modulation matrix;

[0117] The response modulation matrix is ​​sampled in time series to obtain a response change sequence, and the response change gradient is calculated based on the response change sequence to obtain the screen response change characteristics.

[0118] Specifically, the current touch mode features are deconstructed to extract a response feature set including response time series, response intensity value and response area data. These features can fully reflect the dynamic characteristics of touch behavior. Assume that the touch mode feature matrix is ,in It is The response time of each touch point, is the corresponding response intensity value, is the touch area data. Decomposed into three independent series: response time series , a sequence of response intensity values , and the response area data series The response feature set is input into the delayed response modulation unit, and the time delay modeling is performed to generate delay model parameters including response delay parameters and response compensation coefficients. Response time series Modeled as ,in is the initial response time, It is a touch point Assuming that the delay shows an exponential decay trend over time, the following formula is used for fitting:

[0119] ;

[0120] in, is the initial delay amplitude, is the decay rate, is the steady-state delay. The above formula is fitted by the least squares method and the parameters are optimized. , β and γ, thereby describing the response delay characteristics of the touch point. In order to compensate for the impact of delay on touch operation, the response compensation coefficient is calculated , defined as:

[0121] ;

[0122] Response time series Perform the maximum response transmission mapping based on the touch point to generate the touch point delay distribution map. The touch point delay distribution map is a two-dimensional mapping of the spatial position of the touch point and its response time, defined as:

[0123] ;

[0124] in It is a touch point After generating the delay distribution graph, the response time compensation value is calculated by analyzing its gradient. , defined as:

[0125] ;

[0126] in and The distribution map is and Direction gradient. Indicates the demand intensity of global compensation. At the same time, the response intensity value sequence and the response area data series A regularization zero-interference operation is performed to remove noise and outliers. This operation is achieved using the standard regularization formula:

[0127] ;

[0128] in, and They are and The mean of and are their standard deviations respectively. Through regularization processing, the data distribution is uniform, which is convenient for subsequent analysis. Based on the delay compensation parameters and non-interference response characteristics , feature reconstruction is performed to generate the response modulation matrix. The reconstruction operation is completed by fusing the above features into a multi-dimensional matrix, defined as:

[0129]

[0130] This matrix combines the time, intensity, area, and compensation characteristics to describe the global characteristics of the touch response. Perform time series sampling to generate a sequence of response changes The goal of timing sampling is to extract the key change pattern in the modulation matrix, and the sampling operation is defined as:

[0131] ;

[0132] in is the size of the sampling window. The dynamic change law of touch response is captured. The response change gradient is calculated based on the response change sequence to generate the screen response change characteristics. The gradient is defined as:

[0133] ;

[0134] Indicates The rate of change of each sampling point. , quantify the response dynamic characteristics of the touch screen.

[0135] In a specific embodiment, the process of executing step 500 may specifically include the following steps:

[0136] Input the state transfer parameters and the screen response change characteristics into the fuzzy logic system to obtain a fuzzy rule input matrix including the signal attenuation ratio, the response change rate and the current mode confidence;

[0137] A membership function is performed on the signal attenuation ratio in the fuzzy rule input matrix to obtain an attenuation ratio membership map, and a membership function is performed on the response change rate in the fuzzy rule input matrix to obtain a response change rate membership map;

[0138] A fuzzy inference rule base is constructed according to the attenuation ratio membership mapping and the response change rate membership mapping to obtain a fuzzy decision value;

[0139] The stability of the fuzzy decision value is calculated in continuous sampling periods to obtain a stability score;

[0140] The stability score is input into a predefined time filter with a 10ms time constant for stability processing to obtain a filtered switching index;

[0141] The feasibility determination is calculated for the filtered switching index to obtain a switching feasibility score, and the switching feasibility score is compared with a preset threshold to obtain a touch mode switching feasibility determination result.

[0142] Specifically, the state transition parameters and screen response change characteristics are used as input to construct a fuzzy rule input matrix including signal attenuation ratio, response change rate, and current mode confidence. Assume that the state transition parameters are ,in represents the time dimension change characteristic of the touch signal, and the screen response change characteristic is ,in Describes the dynamic response behavior of the touch screen. Based on these inputs, defines the signal attenuation ratio and the response rate They are:

[0143] ;

[0144] in, and are the means of the state transition parameters and the screen response change characteristics respectively. Current mode confidence By selecting the maximum value of the probability distribution of the touch mode:

[0145] ;

[0146] in is the touch mode classification probability, is the output of the fused feature vector. , and Combined into fuzzy rule input matrix:

[0147] ;

[0148] Signal attenuation ratio in the fuzzy rule input matrix and the response rate Membership function calculations are performed to generate attenuation ratio membership maps and response change rate membership maps. A membership function is a mathematical function that maps a continuous variable to a fuzzy set. Assuming the fuzzy sets of the attenuation ratio are "low", "medium", and "high", their membership functions are:

[0149] ;

[0150] ;

[0151] ;

[0152] in is the parameter of the fuzzy membership function. Response change rate The membership function of is defined in a similar way. According to the attenuation ratio membership mapping and the response change rate membership mapping, combined with the current mode confidence , build a fuzzy reasoning rule base. The rule base contains a set of conditional statements, and the mathematical expression of each rule is:

[0153] ;

[0154] Generate fuzzy decision values ​​through fuzzy reasoning calculation :

[0155] ;

[0156] in is the weight of the rule, reflecting the importance of each rule. A stability calculation is performed over consecutive sampling periods to assess the temporal consistency of the decision value. The stability calculation is performed by analyzing the mean and variance of consecutive samples to define a stability score. for:

[0157] ;

[0158] in is the mean of the fuzzy decision values, is its variance. The stability score Enter a predefined time filter with a 10ms time constant to smooth out short-term fluctuations. The expression for the time filter is:

[0159] ;

[0160] in ms is the time constant. The feasibility decision calculation is performed on the filtered switching indicators to generate a switching feasibility score. The feasibility score is calculated as:

[0161] ;

[0162] in is the preset decision threshold. Compared with the threshold, the feasibility determination result of touch mode switching is finally generated:

[0163] .

[0164] In a specific embodiment, the process of executing step 600 may specifically include the following steps:

[0165] The feasibility determination result of the touch mode switching is input into the fuzzy predefined time response controller for nonlinear characteristic modeling to obtain characteristic model parameters including touch state variables and quantized touch signals;

[0166] Adaptively recursively operate the characteristic model parameters to obtain a control benchmark sequence including performance function values ​​and response time constraints;

[0167] According to the control reference sequence, a feedforward compensation unit based on fuzzy logic is constructed to obtain a feedforward control quantity including a compensation direction vector and a compensation amplitude coefficient;

[0168] The feedforward control quantity is input into a predefined time filter to eliminate repeated operations and obtain an optimized compensation sequence;

[0169] Perform closed-loop differential calculation on the optimized compensation sequence and the real-time response parameters of the touch screen to obtain a feedback correction amount including real-time deviation and cumulative error;

[0170] Based on the feedback correction amount, the feedforward control amount is iteratively optimized to obtain the closed-loop control parameters;

[0171] The closed-loop control parameters are input into the delay compensation controller to predict the response characteristics, obtain the predicted compensation sequence, and perform stability constraints on the predicted compensation sequence to obtain the mode switching delay compensation control quantity.

[0172] Specifically, the feasibility determination result of the touch mode switching is input into the fuzzy predefined time response controller for nonlinear characteristic modeling. Assume that the feasibility determination result of the touch mode switching is , whose value range is [0,1], indicates the necessity of switching. Through nonlinear feature modeling, Convert to include touch state variables and quantify touch signals The modeling process relies on a fuzzy logic system, in which the state variables The calculation formula is:

[0173] ;

[0174] in is a fuzzy membership function used to quantify The impact on the system status, is the touch signal sequence. Quantized touch signal The signal strength is defined as The weighted mean of :

[0175] ;

[0176] in is the weight, which depends on the importance of the touch signal. Through this modeling process, the feature model parameters Complete the quantification of the characteristics of touch mode switching. Perform adaptive recursive operations on the characteristic model parameters to generate performance function values and response time constraints The control benchmark sequence of the performance function It is defined as the sum of squares of the deviations between the current touch state and the expected state. The formula is:

[0177] ;

[0178] in is the target state variable, is the current state variable. Response time constraint is the real-time performance indicator of the touch system and is updated by the following recursive formula:

[0179] ;

[0180] in is the smoothing factor, is the current response time change. According to the generated control reference sequence, a feedforward compensation unit based on fuzzy logic is constructed to generate a compensation direction vector and compensation amplitude coefficient The goal of feedforward compensation is to reduce the delay of mode switching through predictive regulation. The compensation direction vector The calculation formula is:

[0181] ;

[0182] Where sign(·) represents the sign function, which is used to determine the adjustment direction. Compensation amplitude coefficient Defined as a combination of a performance function value and a response time constraint:

[0183] ;

[0184] The feedforward control quantity is input into a predefined time filter to eliminate repeated operations and short-term noise, generating an optimized compensation sequence The time filter is in the form of a first-order low-pass filter, and its recursive formula is:

[0185] ;

[0186] in is the filter coefficient, which is used to control the update speed. The optimized compensation sequence and the real-time response parameters of the touch screen are calculated by closed-loop difference to obtain the real-time deviation and cumulative error The feedback correction amount. The calculation formula for real-time deviation is:

[0187] ;

[0188] The cumulative error is calculated by integration:

[0189] ;

[0190] Based on the feedback correction, the feedforward control quantity is iteratively optimized to generate closed-loop control parameters. The goal of closed-loop optimization is to reduce the deviation by adjusting the compensation parameters, and the formula is:

[0191] ;

[0192] in is the proportional gain, is the integral gain. The closed-loop control parameters are input into the delay compensation controller to predict the response characteristics and generate a prediction compensation sequence Prediction compensation is achieved by modeling the response delay behavior of the touch screen, and the model form is:

[0193] ;

[0194] in is the delay attenuation coefficient. Stability constraints are enforced on the prediction compensation sequence to ensure output stability, and finally the mode switching delay compensation control quantity is generated. The stability constraint is ensured by the following conditions:

[0195]

[0196] in is the stability threshold.

[0197] The above describes the touch screen multi-touch mode switching method in the embodiment of the present application. The following describes the touch screen multi-touch mode switching device 10 in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the touch screen multi-touch mode switching device 10 includes:

[0198] The processing module 11 is used to perform minimum error processing on the touch signal collected by the touch screen to obtain the target characteristic component;

[0199] A mapping module 12, used for inputting the target feature component into the multi-layer touch mode classifier for feature mapping to obtain the current touch mode feature;

[0200] The adjustment module 13 is used to perform time dimension attenuation characteristic analysis and adaptive threshold adjustment on the target characteristic component to obtain the state transfer parameter;

[0201] The modulation module 14 is used to perform delayed response modulation and regularized zero interference on the touch point response time, response intensity and response area based on the current touch mode characteristics to obtain the screen response change characteristics;

[0202] A decision module 15 is used to make fuzzy logic decisions in continuous sampling cycles according to the state transfer parameters and the screen response change characteristics, and obtain a feasibility determination result of the touch mode switching;

[0203] The control module 16 is used to perform closed-loop control on the touch screen response parameters based on the touch mode switching feasibility determination result to obtain the mode switching delay compensation control amount.

[0204] Through the synergy of the above components, the combination of minimum error processing and multi-layer touch mode classifiers improves the accuracy of feature extraction of touch signals and effectively reduces the error rate of touch mode recognition. The time dimension attenuation characteristic analysis and adaptive threshold adjustment mechanism are used to solve the problem of signal instability during touch mode switching and enhance the system's resistance to environmental interference. The introduction of delayed response modulation and regularized zero interference technology effectively eliminates interference signals during touch point response and ensures the accuracy of touch response. The continuous sampling cycle judgment mechanism based on fuzzy logic decision-making improves the reliability of touch mode switching and avoids the occurrence of false triggering. The closed-loop control strategy and delay compensation mechanism are used to significantly reduce the response delay during mode switching and optimize the user operation experience.

[0205] See also Figure 3 , Figure 3 This is a schematic block diagram of the structure of an electronic device 300 provided in an embodiment of the present application. The electronic device 300 includes a processor 301 and a memory 302. The processor 301 and the memory 302 are connected via a device bus 303, wherein the memory 302 may include a non-volatile storage medium and an internal memory.

[0206] The non-volatile storage medium can store a computer program. The computer program includes program instructions. When the program instructions are executed by the processor 301, the processor 301 can execute any of the above-mentioned touch screen multi-touch mode switching methods.

[0207] The processor 301 is used to provide computing and control capabilities to support the operation of the entire electronic device 300 .

[0208] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 301, the processor 301 can execute any of the above-mentioned touch screen multi-touch mode switching methods.

[0209] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the present application scheme, and does not constitute a limitation on the electronic device 300 involved in the present application scheme. The specific electronic device 300 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0210] It should be understood that the processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0211] It should be noted that those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the electronic device 300 described above can refer to the corresponding process of the aforementioned touch screen multi-touch mode switching method, which will not be repeated here.

[0212] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement the touch screen multi-touch mode switching method provided in the embodiment of the present application.

[0213] The computer-readable storage medium may be an internal storage unit of the electronic device 300 in the aforementioned embodiment, such as a hard disk or memory of the electronic device 300. The computer-readable storage medium may also be an external storage device of the electronic device 300, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the electronic device 300.

[0214] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0215] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0216] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for switching multiple touch modes of a touch screen, characterized in that: The method comprises: Perform minimum error processing on the touch signal collected by the touch screen to obtain the target characteristic component; Inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain current touch mode features; Performing time dimension attenuation characteristic analysis and adaptive threshold adjustment on the target feature component to obtain state transfer parameters; specifically comprising: performing time series analysis on the target feature component to obtain an attenuation feature matrix including touch signal strength, attenuation rate and steady-state value; performing gradient calculation on the signal strength in the attenuation feature matrix to obtain touch signal attenuation curve parameters, and inputting the touch signal attenuation curve parameters into an attenuation fitting unit to extract inflection points to obtain attenuation critical values; dividing the attenuation feature matrix into time periods according to the attenuation critical values ​​to obtain a segmented attenuation feature sequence, and performing sliding average calculation on the steady-state value sequence in the segmented attenuation feature sequence to obtain a dynamic threshold adjustment benchmark; calculating the upper and lower threshold ranges and the adjustment step size based on the dynamic threshold adjustment benchmark to obtain an adaptive threshold parameter; performing matching operation on the adaptive threshold parameter and the attenuation feature matrix to obtain a state feature vector, and performing normalization operation on the state feature vector to obtain a state transfer parameter; Based on the current touch mode characteristics, delayed response modulation and regularized zero interference are performed on the touch point response time, response intensity and response area to obtain screen response change characteristics; Perform fuzzy logic decision making in continuous sampling cycles according to the state transfer parameters and the screen response change characteristics to obtain a touch mode switching feasibility determination result; Based on the feasibility determination result of the touch mode switching, the touch screen response parameters are closed-loop controlled to obtain the mode switching delay compensation control amount; specifically, the method comprises: inputting the feasibility determination result of the touch mode switching into a fuzzy predefined time response controller for nonlinear feature modeling to obtain feature model parameters including touch state variables and quantized touch signals; performing adaptive recursive operation on the feature model parameters to obtain a control reference sequence including performance function values ​​and response time constraints; constructing a fuzzy logic-based feedforward compensation unit according to the control reference sequence to obtain a feedforward control amount including a compensation direction vector and a compensation amplitude coefficient; inputting the feedforward control amount into a predefined time filter for repeated operation elimination to obtain an optimized compensation sequence; performing closed-loop difference calculation on the optimized compensation sequence and the real-time response parameters of the touch screen to obtain a feedback correction amount including a real-time deviation and a cumulative error; performing iterative optimization operation on the feedforward control amount based on the feedback correction amount to obtain a closed-loop control parameter; inputting the closed-loop control parameter into a delay compensation controller for response feature prediction to obtain a predicted compensation sequence, and executing stability constraints on the predicted compensation sequence to obtain a mode switching delay compensation control amount.

2. The touch screen multi-touch mode switching method according to claim 1, characterized in that: The step of performing minimum error processing on the touch signal collected by the touch screen to obtain the target characteristic component includes: Perform touch signal acquisition on the touch point position of the touch screen to obtain an initial sampling matrix including the touch point xy coordinates, pressure value and time stamp; Performing sliding window segmentation on the initial sampling matrix to obtain a multi-dimensional touch signal feature sequence, and performing least squares calculation on the pressure values ​​in the multi-dimensional touch signal feature sequence to obtain a touch pressure principal component feature; Performing dimension reduction mapping on the multi-dimensional touch signal feature sequence according to the touch pressure principal component feature to obtain the touch point spatial distribution feature, and performing dynamic time warping on the touch point spatial distribution feature to obtain the touch point time series change feature; The touch pressure principal component feature and the touch point temporal change feature are subjected to feature fusion to obtain a multimodal feature vector, and the multimodal feature vector is subjected to error minimization iterative calculation to obtain a target feature component.

3. The touch screen multi-touch mode switching method according to claim 2, characterized in that: The step of inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain the current touch mode feature includes: Performing a touch point response mapping conversion on the target feature component to obtain a feature vector including a touch point coordinate matrix, a pressure value sequence and a timestamp mark; The feature vector is input into the first layer classifier of the multi-layer touch pattern classifier for maximum response transfer mapping processing based on the touch point, wherein the first layer classifier adopts a convolutional neural network structure, including 3 convolutional layers and 2 pooling layers, and each convolutional layer uses a ReLU activation function to obtain a touch point distribution feature map; Inputting the touch point distribution feature map into the second layer classifier of the multi-layer touch pattern classifier for zero interference elimination, the second layer classifier adopts a deep residual network structure enhanced by an attention mechanism, including 4 residual blocks, each residual block includes two convolutional layers and a short-circuit connection, and a self-attention layer is added between the residual blocks to obtain a non-interference feature matrix; Performing temporal feature extraction on the non-interference feature matrix to obtain touch mode temporal features, and inputting the touch mode temporal features into the third layer classifier of the multi-layer touch mode classifier for multimodal feature fusion, wherein the third layer classifier adopts a multi-head self-attention mechanism, including 4 attention heads, each with a dimension of 64, to obtain a fused feature vector; The fused feature vector is subjected to dimensionality reduction mapping to obtain a touch pattern probability distribution, and a confidence score and a duration feature are calculated based on the touch pattern probability distribution. The touch pattern category is obtained through threshold judgment, and the touch pattern category, the confidence score and the duration feature are combined to obtain the current touch pattern feature.

4. The touch screen multi-touch mode switching method according to claim 1, characterized in that: The delay response modulation and regularization zero interference are performed on the touch point response time, response intensity and response area based on the current touch mode characteristics to obtain the screen response change characteristics, including: Deconstructing the current touch mode feature to obtain a response feature set including a response time series, a response intensity value, and a response area data; Inputting the response feature set into a delay response modulation unit for delay modeling to obtain delay model parameters including response delay parameters and response compensation coefficients; Performing a maximum response transmission mapping based on a contact on the response time sequence to obtain a contact delay distribution diagram, and calculating a response time compensation value according to the contact delay distribution diagram to obtain a delay compensation parameter; Regularized zero-interference operation is performed on the response intensity value and response area data to obtain interference-free response characteristics; Perform feature reconstruction based on the delay compensation parameter and the interference-free response feature to obtain a response modulation matrix; The response modulation matrix is ​​sampled in time series to obtain a response change sequence, and a response change gradient is calculated based on the response change sequence to obtain a screen response change feature.

5. The touch screen multi-touch mode switching method according to claim 4, characterized in that: The fuzzy logic decision of continuous sampling period is performed according to the state transfer parameter and the screen response change characteristic to obtain the feasibility determination result of touch mode switching, including: Inputting the state transfer parameter and the screen response change characteristic into a fuzzy logic system to obtain a fuzzy rule input matrix including a signal attenuation ratio, a response change rate and a current mode confidence; Performing membership function calculation on the signal attenuation ratio in the fuzzy rule input matrix to obtain an attenuation ratio membership map, and performing membership function calculation on the response change rate in the fuzzy rule input matrix to obtain a response change rate membership map; Constructing a fuzzy inference rule base according to the attenuation ratio membership mapping and the response change rate membership mapping to obtain a fuzzy decision value; Performing stability calculation on the fuzzy decision value within a continuous sampling period to obtain a stability score; Inputting the stability score into a predefined time filter including a 10 ms time constant for stability processing to obtain a filtered switching index; A feasibility determination calculation is performed on the filtered switching index to obtain a switching feasibility score, and the switching feasibility score is compared with a preset threshold to obtain a touch mode switching feasibility determination result.

6. A touch screen multi-touch mode switching device, characterized in that: Used to execute the touch screen multi-touch mode switching method according to any one of claims 1 to 5, the touch screen multi-touch mode switching device comprises: A processing module, used for performing minimum error processing on the touch signal collected by the touch screen to obtain a target characteristic component; A mapping module, used for inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain a current touch mode feature; An adjustment module, used for performing time dimension attenuation characteristic analysis and adaptive threshold adjustment on the target characteristic component to obtain a state transfer parameter; A modulation module, used to perform delayed response modulation and regularized zero interference on the touch point response time, response intensity and response area based on the current touch mode characteristics, so as to obtain screen response change characteristics; A decision module, used to make fuzzy logic decisions in continuous sampling cycles according to the state transfer parameters and the screen response change characteristics, and obtain a feasibility determination result of the touch mode switching; The control module is used to perform closed-loop control on the touch screen response parameters based on the touch mode switching feasibility determination result to obtain the mode switching delay compensation control amount.

7. An electronic device, characterized in that: The electronic device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instruction in the memory to enable the electronic device to execute the touch screen multi-touch mode switching method according to any one of claims 1-5.

8. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the touch screen multi-touch mode switching method according to any one of claims 1 to 5 is implemented.

Images