Touch system and method for increasing report rate of the touch system

Abstract

A touch system includes a touch panel, an analog front-end circuit, a storage unit, and a computing unit. The analog front-end circuit sequentially performs a mutual capacitance scan on the touch panel to generate first mutual capacitance sensing information and a self-capacitance scan to generate self-capacitance sensing information. The storage unit stores the first mutual capacitance sensing information and the self-capacitance sensing information. The computing unit includes a machine learning model that generates second mutual capacitance sensing information based on the first mutual capacitance sensing information and the self-capacitance sensing information. The computing unit generates first touch coordinate information based on the first mutual capacitance sensing information and second touch coordinate information based on the second mutual capacitance sensing information.

Description

[Technical Field]

[0001] This invention relates to a touch system, and more particularly to a method for improving the return rate of a touch system. [Previous Technology]

[0002] Figure 1 is a schematic diagram of a touch panel. A touch panel 10 includes m first directional electrodes X1~Xm and n second directional electrodes Y1~Yn. The m parallel first directional electrodes X1~Xm and the n parallel second directional electrodes Y1~Yn are isolated by an insulating layer (not shown in the figure). The m first directional electrodes and the n second directional electrodes intersect to form m×n intersection points. A conventional scanning method involves performing a mutual capacitance scan and a self-capacitive scan on the touch panel 10 during a scan. The mutual capacitance scan can generate mutual capacitance sensing information including the sensing values ​​of the m×n intersection points, which can be used to determine the contact position or coordinates of an object. The self-capacitive scan can generate self-capacitive sensing information including the sensing values ​​of the m first directional electrodes and the n second directional electrodes, which can be used to achieve specific functions, such as waterproofing.

[0003] In current touch applications, there is a desire to increase the polling rate (hereinafter referred to as polling rate) of touch coordinates. Currently, the way to increase the polling rate is to use high-performance processors to reduce the computation time of the algorithm. [Summary of the Invention]

[0004] One of the objectives of this invention is to provide a touch system and method for improving the rate of return.

[0005] According to the present invention, a touch system includes a touch panel, an analog front-end circuit, a storage unit, and a computing unit. The analog front-end circuit is used to sequentially perform a mutual capacitance scan on the touch panel to generate first mutual capacitance sensing information and a self-capacitance scan to generate self-capacitance sensing information. The storage unit is used to store the first mutual capacitance sensing information and the self-capacitance sensing information. The computing unit includes a machine learning model, which generates second mutual capacitance sensing information based on the first mutual capacitance sensing information and the self-capacitance sensing information. The computing unit generates first touch coordinate information based on the first mutual capacitance sensing information and generates second touch coordinate information based on the second mutual capacitance sensing information.

[0006] According to the present invention, a method for improving the polling rate of a touch system includes the following steps: A. sequentially performing a mutual capacitance scan on a touch panel to generate a first mutual capacitance sensing information and a self-capacitance scan to generate a self-capacitance sensing information; B. generating a first touch coordinate information based on the first mutual capacitance sensing information; C. inputting the first mutual capacitance sensing information and the self-capacitance sensing information into a machine learning model to generate a second mutual capacitance sensing information; and D. generating a second touch coordinate information based on the second mutual capacitance sensing information.

[0007] According to the present invention, a touch system includes a touch panel, an analog front-end circuit, a storage unit, and a computing unit. The analog front-end circuit is used to sequentially perform mutual capacitance scanning on the touch panel to generate mutual capacitance sensing information and self-capacitance scanning to generate self-capacitance sensing information. The storage unit is used to store the mutual capacitance sensing information and the self-capacitance sensing information. The computing unit generates first touch coordinate information based on the mutual capacitance sensing information. The computing unit includes a machine learning model, which generates second touch coordinate information based on the first touch coordinate information and the self-capacitance sensing information.

[0008] According to the present invention, a method for improving the polling rate of a touch system includes the following steps: A. sequentially performing a mutual capacitance scan on a touch panel to generate mutual capacitance sensing information and a self-capacitance scan to generate self-capacitance sensing information; B. generating a first touch coordinate information based on the mutual capacitance sensing information; and C. inputting the first touch coordinate information and the self-capacitance scan information into a machine learning model to generate a second touch coordinate information. The mutual capacitance scan information includes K mutual capacitance sensing values, the self-capacitance scan information includes L self-capacitance sensing values, and the first touch coordinate information includes Q first coordinate groups, each of the first coordinate groups including a first X coordinate and a first Y coordinate, where K, L, and Q are positive integers greater than or equal to 1.

Implementation Method

[0010] FIG. The touch system 20 of FIG. The architecture of the touch panel 21 can be referred to FIG. 1 , where the touch panel 21 includes m first-direction electrodes X1 ~ The first direction electrodes X1~Xm in FIG. The controller 22 is used to drive the touch panel 21 to obtain sensing information. The controller 22 of FIG. The computing unit 223 may be, but is not limited to, a Microcontroller Unit (MCU) or a Micro-Processor. Analog front-end circuit 221l coupling contacts all electrodes of the control panel 21 . During a scan, the computing unit 223 sends a scan command to the analog front-end circuit 221 to enable the analog front-end circuit 221 to perform a mutually accommodating scan and a self-containable scan to the touch panel 21 sequentially. After performing a mutually capacitive scan of the touch panel 21 , the analog front-end circuit 221 can obtain a mutual capacitance sensing information MD1 from the touch panel 21 , wherein the mutual capacitance sensing information MD1 includes the mutual capacitance sensing at the intersection of each first-direction electrode in the touch panel 21 and each second-direction electrode, i.e. After performing a self-accommodating scan of the touch panel 21 , the analog front-end circuit 221 can obtain a self-capacitating sensing information SD1 from the touch panel 21 , wherein the self-capacitating sensing information SD1 includes the self-capacitating sensing of m first-direction electrodes X1~Xm and n second-direction electrodes Y1~Yn in the touch panel 21 . The specific operating procedures of mutually accommodating scanning and self-accommodating scanning are known technical means and are not repeated here. The storage unit 222 stores the mutually accommodating sensing information MD1 output by the analog front-end circuit 221 as well as the self-accommodating sensing information SD1. Computing unit 223 obtains mutually accommodating sensing information MD1 as well as self-accommodating sensing information SD1 from storage unit 222, and generates one-touch coordinate information TD1 and one-touch coordinate information TD2 based on mutually accommodating sensing information MD1 and self-accommodating sensing information SD1.

[0011] In the embodiment of FIG. 2, the storage unit 222 and the computing unit 223 are disposed in the controller 22, but the present invention is not limited thereto. FIG. 3 shows a second embodiment of the touch system of the present invention. The touch system 30 of FIG. 3, like the touch system 20 of FIG. 2, includes a touch panel 21, an analog front-end circuit 221, a storage unit 222, a computing unit 223, and a host 23. The difference between the touch system 30 of FIG. 3 and the touch system 20 of FIG. 2 is that the storage unit 222 and the computing unit 223 of FIG. 3 are disposed in the host 23. The host 23 can be, for example, a central processing unit of a computer. Compared with the controller 22, the host 23 can usually be configured with a higher-performance computing unit 223, thus reducing the computation time of the touch coordinate algorithm and further improving the return rate of the touch coordinate.

[0012] Figure 4 shows a first embodiment of the computing unit of Figures 2 and 3. The computing unit 223 of Figure 4 includes a data processing unit 2231 and a machine learning model 2232, wherein the data processing unit 2231 is connected to the storage unit 222 and the machine learning model 2232, and the machine learning model 2232 is connected to the storage unit 222. The data processing unit 2231 generates touch coordinate information TD1 based on the mutual capacitance sensing information MD1 in the storage unit 222. The machine learning model 2232 generates mutual capacitance sensing information MD2 based on the mutual capacitance sensing information MD1 and the self-capacitance sensing information SD1 in the storage unit 222. The mutual capacitance sensing information MD2 generated by the machine learning model 2232 is provided to the data processing unit 2231. The data processing unit 2231 generates touch coordinate information TD2 based on the mutual capacitance sensing information MD2. In one embodiment, the machine learning model 2232 may be implemented as a convolutional neural network (CNN) or a multilayer perceptron (MLP), but the present invention is not limited thereto.

[0013] In one embodiment, the machine learning model 2232 can generate K virtual mutual capacitance sensing values ​​to form mutual capacitance scanning information MD2 according to the following formula 1: Formula 1 Where x is a positive integer from 1 to K, 0 is K+L, d2x is the xth virtual mutual capacitance sensing value of mutual capacitance sensing information MD2, d11~d1K are K mutual capacitance sensing values ​​of mutual capacitance sensing information MD1, S1~SL are L self-capacitance sensing values ​​of self-capacitance sensing information SD1, W1x~WOx are the weight values ​​of the xth virtual mutual capacitance sensing value, b1x~bOx are the offset values ​​of the xth virtual mutual capacitance sensing value, and the symbol "." represents the Adama product.

[0014] In one embodiment, the machine learning model 2232 can first perform convolution operations on the K mutual capacitance sensing values ​​of the mutual capacitance sensing information MD1 to generate P convolution result values, where P is a positive integer greater than 1 and less than K. Then, it generates K virtual mutual capacitance sensing values ​​to form the mutual capacitance scanning information MD2 according to the following formula 2: Formula 2 Where x is a positive integer from 1 to K, Q is P+L, d2x is the xth virtual mutual capacitance sensing value of the mutual capacitance sensing information MD2, c1~cP are convolution result values, S1~SL are the L self-capacitance sensing values ​​of the self-capacitance sensing information SD1, W1x~WQx are the weight values ​​of the xth virtual mutual capacitance sensing value, b1x~bQx are the offset values ​​of the xth virtual mutual capacitance sensing value, and the symbol "." represents the Adama product.

[0015] The calculation unit 223 in Figure 4 uses self-capacitance sensing information SD1 and mutual capacitance sensing information MD1 to generate mutual capacitance sensing information MD2, and then obtains the second touch coordinate information TD2 based on the mutual capacitance sensing information MD2. Therefore, the touch system 20 of the present invention can generate two touch coordinate information TD1 and TD2 in one scan period, which has a higher return rate.

[0016] From the descriptions of Figures 2 and 4 above, it can be understood that the first embodiment of the method for improving the return rate of a touch system according to the present invention can be expressed as shown in Figure 5, including the following steps: Step S10: Sequentially performing a mutual capacitance scan on the touch panel 21 to generate mutual capacitance sensing information MD1 and a self-capacitance scan to generate self-capacitance sensing information SD1; Step S11: Generating touch coordinate information TD1 based on the mutual capacitance sensing information MD1; Step S12: Inputting the mutual capacitance sensing information MD1 and the self-capacitance sensing information SD1 into the machine learning model 2232 to generate mutual capacitance sensing information MD2; and Step S13: Generating touch coordinate information TD2 based on the mutual capacitance sensing information MD2. The order of steps S11 and S12 is not limited to that shown in Figure 5. In different embodiments, it is also possible to perform steps S11 and S12 simultaneously.

[0017] The weight values ​​and offset values ​​of the aforementioned formulas 1 to 2 can be obtained by training the machine learning model. The material for training the machine learning model is the continuous multiple sensing information obtained by actually scanning the touch panel, such as mutually accommodating sensing information MD11, self-accommodating sensing information SD11, mutually accommodating sensing information MD12, self-accommodating sensing information SD12…. The training method is to input, for example, mutually accommodating sensing information MD11 as well as self-accommodating sensing information SD11 into a machine learning model that generates mutually accommodating sensing information MD2 based on mutually accommodating sensing information MD11 as well as self-accommodating sensing information SD11. The mutual capacitance sensing information MD2 was then compared with the ground truth information GT1. When the error value of the mutually accommodating sensing information MD2 and the benchmark truth information GT1 is greater than a first default value, the parameters of the machine learning model (i.e., weight values ​​​​and offset values) are updated based on the error value and an optimization algorithm. In contrast, the training of the machine learning model is ended when the error value of the mutual compatibility sensing information MD2 and the benchmark truth information GT1 is less than or equal to that first default value or the number of iterations reaches one second default value. In one embodiment, the optimization algorithm may be, but is not limited to, Adam (Adaptive Moment Estimation) or SGD (Stochastic Gradient Descent).

[0018] There are many ways to obtain the benchmark truth information. For example, the benchmark truth information GT1 can be determined by the average of the two strokes of mutual compatibility sensing information before and after. Specifically, it is assumed that a mutually accommodating sensing information MD11 as well as a self-accommodating sensing information SD11 are obtained during a first scan, and a mutually accommodating sensing information MD12 as well as a self-accommodating sensing information SD12 are obtained during the following second scan during the first scan. Then the average value of the mutual-accommodating sensing information MD11 and the mutual-accommodating sensing information MD12 can be used as the benchmark truth information GT1 to determine whether the mutual-accommodating sensing information MD2 generated according to the mutual-accommodating sensing information MD11 and the self-accommodating sensing information SD11 is qualified.

[0019] Another embodiment of the computing unit 223 is shown in FIG. 6. The computing unit 223 of FIG. 6 includes a data processing unit 2233 and a machine learning model 2234, wherein the data processing unit 2233 is connected to the storage unit 222 and the machine learning model 2234, and the machine learning model 2234 is also connected to the storage unit 222. The data processing unit 2233 generates touch coordinate information TD1 based on the mutual capacitance sensing information MD1 in the storage unit 222. The machine learning model 2234 generates touch coordinate information TD2 based on the self-capacitance sensing information SD1 in the storage unit 222 and the touch coordinate information TD1 provided by the data processing unit 2233. The touch coordinate information TD1 includes R first coordinate sets, and the touch coordinate information TD2 includes R second coordinate sets, where R is a positive integer greater than or equal to 1. The R first coordinate sets and the R second coordinate sets represent the positions of R objects on the touch panel 21. In one embodiment, the machine learning model 2234 may be implemented using a CNN or an MLP, but the invention is not limited thereto.

[0020] In one embodiment, the machine learning model 2234 can obtain the touch coordinate information TD2 according to the following formulas 3 and 4: Formula 3 Formula 4 Where i is a positive integer from 1 to R, t is 2R+L, X2i is the X coordinate of the i-th second coordinate group, Y2i is the Y coordinate of the i-th second coordinate group, X11~X1i are the X coordinates of R first coordinate groups, Y11~Y1i are the Y coordinates of R first coordinate groups, S1~SL are the L self-capacitive sensing values ​​of the self-capacitive sensing information SD1, WX1i~WXti are the weight values ​​of the X coordinates of the i-th second coordinate group, bX1i~bXti are the offset values ​​of the X coordinates of the i-th second coordinate group, WY1i~WYti are the weight values ​​of the Y coordinates of the i-th second coordinate group, bY1i~bYti are the offset values ​​of the Y coordinates of the i-th second coordinate group, and the symbol "." represents the Adama product.

[0021] The calculation unit 223 in Figure 6 uses the self-capacitance sensing information SD1 combined with the touch coordinate information TD1 to generate a touch coordinate information TD2. Therefore, the touch system 20 of the present invention can obtain two touch coordinate information TD1 and TD2 in one scanning period, which has a higher return rate.

[0022] From the above description of FIG Step S22: Input the touch coordinate information TD1 and self-containment scan information SD1 into the machine learning model 2234 to generate a touch coordinate information TD2.

[0023] The weight values ​​and offset values ​​of the foregoing formulas 3 to 4 can be obtained by training the machine learning model. The material for training the machine learning model is the continuous multiple touch coordinate information and self-contained sensing information obtained by actually scanning the touch panel. Among them, the touch coordinate information TD11 is obtained by performing coordinate operations based on the mutually accommodating sensing information MD11. In this embodiment, the machine learning model is trained by inputting the touch coordinate information TD11 as well as the self-contained sensing information SD11 into a machine learning model that generates the touch coordinate information TD2 based on the touch coordinate information TD11 as well as the self-contained sensing information SD11. The touch coordinate information TD2 was then compared with the ground truth information GT2. When the error value of the touch coordinate information TD2 and the benchmark truth information GT2 is greater than a first default value, the parameters of the machine learning model (i.e., the weight value and the offset value) are updated based on the error value and an optimization algorithm, and the touch coordinate information TD11 and the self-contained sensing information SD11 are again input into the updated machine learning model for the next iteration. In contrast, the training of the machine learning model is ended when the error value of the touch coordinate information TD2 and the benchmark truth information GT2 is less than or equal to that first default value or the number of iterations reaches a second default value. In one embodiment, the optimization algorithm may be, but is not limited to, Adam (Adaptive Moment Estimation) or SGD (Stochastic Gradient Descent).

[0024] There are many ways to obtain the baseline truth information GT2. For example, the baseline truth information GT2 can be determined by averaging the values ​​of two consecutive touch coordinate information. Specifically, assume that mutual capacitance sensing information MD11 and self-capacitance sensing information SD11 are obtained during a first scan, and mutual capacitance sensing information MD12 and self-capacitance sensing information SD12 are obtained during a second scan following the first scan. Touch coordinate information TD11 can be calculated from the mutual capacitance sensing information MD11, and touch coordinate information TD12 can be calculated from the mutual capacitance sensing information MD12. The average value of touch coordinate information TD11 and touch coordinate information TD12 can be used as the baseline truth information GT2 to determine whether the touch coordinate information TD12 generated based on touch coordinate information TD11 and self-capacitance sensing information SD11 is qualified.

[0025] As can be seen from the above description, the touch system 20 of the present invention can generate two touch coordinate information during one scan period, thereby improving the return rate.

[0026] The above description is merely an embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above by way of embodiment, it is not intended to limit the present invention. Anyone skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention. [Simplified Explanation of the Diagram]

[0009] Figure 1 is a schematic diagram of a conventional touch panel. Figure 2 shows a first embodiment of the touch system of the present invention. Figure 3 shows a second embodiment of the touch system of the present invention. Figure 4 shows a first embodiment of the computing unit of the present invention. Figure 5 shows a first embodiment of the method for improving the return rate of the touch system of the present invention. Figure 6 shows a second embodiment of the computing unit of the present invention. Figure 7 shows a second embodiment of the method for improving the return rate of the touch system of the present invention.

Claims

1. A touch system, comprising: One touch panel; An analog front-end circuit connected to the touch panel is configured to sequentially perform a mutual capacitance scan on the touch panel to generate a first mutual capacitance sensing information and a self-capacitance scan to generate a self-capacitance sensing information; a storage unit connected to the analog front-end circuit is configured to store the first mutual capacitance sensing information and the self-capacitance sensing information; and a computing unit connected to the storage unit, the computing unit including a machine learning model, the machine learning model generating a second mutual capacitance sensing information based on the first mutual capacitance sensing information and the self-capacitance sensing information, the computing unit generating a first touch coordinate information based on the first mutual capacitance sensing information and generating a second touch coordinate information based on the second mutual capacitance sensing information.

2. The touch system as claimed in claim 1, wherein the computing unit further comprises: A data processing unit, connected to the storage unit and the machine learning model, is used to generate the first touch coordinate information based on the first mutual capacitance sensing information and to generate the second touch coordinate information based on the second mutual capacitance sensing information.

3. The touch system as claimed in claim 1, wherein the touch panel includes m first directional electrodes and n second directional electrodes, the m first directional electrodes and the n second directional electrodes are perpendicular, the first mutual capacitance sensing information and the second mutual capacitance sensing information both include K = m × n mutual capacitance sensing values, and the self-capacitive sensing information includes L = m + n self-capacitive sensing values, where K and L are positive integers greater than or equal to 1.

4. The touch system as claimed in claim 3, wherein the machine learning model generates the second mutual capacitance sensing information by obtaining K virtual mutual capacitance sensing values ​​to form the second mutual capacitance sensing information according to the following formula: , where x is a positive integer from 1 to K, O is K+L, d2x is the x-th virtual mutual capacitance sensing value of the second mutual capacitance sensing information, d11~d1K are the K mutual capacitance sensing values ​​of the first mutual capacitance sensing information, S1~SL are the L self-capacitance sensing values ​​of the self-capacitance sensing information, W1x~WOx are the weight values ​​of the x-th virtual mutual capacitance sensing value, and b1x~bOx are the offset values ​​of the x-th virtual mutual capacitance sensing value. The symbol represents the Adama product.

5. The touch system as described in claim 3, wherein the machine learning model generates the second mutual capacitance sensing information in the following manner: C1. Perform a convolution operation on the K mutual capacitance inductance values ​​to generate P convolution result values, where P is a positive integer greater than 1 and less than K; and C2. Obtain K virtual mutual capacitance inductance values ​​according to the following formula to form the second mutual capacitance sensing information: , where x is a positive integer from 1 to K, Q is P+L, d2x is the x-th virtual mutual capacitance inductance value of the second mutual capacitance sensing information, c1~cP are the convolution result values, S1~SL are the L self-capacitance inductance values ​​of the self-capacitance sensing information, W1x~WQx are the weight values ​​of the x-th virtual mutual capacitance inductance value, b1x~bQx are the offset values ​​of the x-th virtual mutual capacitance inductance value, and the symbol represents the Adama product.

6. A method for improving the polling rate of a touch system, comprising the following steps: A. sequentially performing a mutual capacitance scan on a touch panel to generate a first mutual capacitance sensing information and a self-capacitance scan to generate a self-capacitance sensing information; B. generating a first touch coordinate information based on the first mutual capacitance sensing information; C. inputting the first mutual capacitance sensing information and the self-capacitance sensing information into a machine learning model to generate a second mutual capacitance sensing information; and D. generating a second touch coordinate information based on the second mutual capacitance sensing information.

7. The method for improving the polling rate of a touch system as described in claim 6, wherein the touch panel includes m first directional electrodes and n second directional electrodes, the m first directional electrodes and the n second directional electrodes are perpendicular, the first mutual capacitance sensing information and the second mutual capacitance sensing information both include K = m × n mutual capacitance sensing values, and the self-capacitive sensing information includes L = m + n self-capacitive sensing values, where K and L are positive integers greater than or equal to 1.

8. The method for improving the polling rate of a touch system as described in claim 7, wherein step C includes obtaining K virtual mutual capacitance sensing values ​​to form the second mutual capacitance sensing information according to the following formula: , where, x is a positive integer from 1 to K, O is K+L, d2x is the x-th virtual mutual capacitance sensing value of the second mutual capacitance sensing information, d11~d1K are the K mutual capacitance sensing values ​​of the first mutual capacitance sensing information, S1~SL are the L self-capacitance sensing values ​​of the self-capacitance sensing information, W1x~WOx are the weight values ​​of the x-th virtual mutual capacitance sensing value, and b1x~bOx are the offset values ​​of the x-th virtual mutual capacitance sensing value. The symbol represents the Adama product.

9. The control method as described in claim 7, wherein step C includes: C1. Perform a convolution operation on the K mutual capacitance inductance values ​​to generate P convolution result values, where P is a positive integer greater than 1 and less than K; and C2. Obtain K virtual mutual capacitance inductance values ​​according to the following formula to form the second mutual capacitance sensing information: ' where x is a positive integer from 1 to K, Q is P+L, d2x is the x-th virtual mutual capacitance inductance value of the second mutual capacitance sensing information, c1~cP are the convolution result values, S1~SL are the L self-capacitance inductance values ​​of the self-capacitance sensing information, W1x~WQx are the weight values ​​of the x-th virtual mutual capacitance inductance value, and b1x~bQx are the offset values ​​of the x-th virtual mutual capacitance inductance value.

10. A touch system, comprising: One touch panel; An analog front-end circuit connected to the touch panel is used to sequentially perform a mutual capacitance scan on the touch panel to generate mutual capacitance sensing information and a self-capacitance scan to generate self-capacitance sensing information; a storage unit connected to the analog front-end circuit is used to store the mutual capacitance sensing information and the self-capacitance sensing information; and a computing unit connected to the storage unit is used to generate a first touch coordinate information based on the mutual capacitance sensing information, wherein the computing unit includes a machine learning model, and the machine learning model generates a second touch coordinate information based on the first touch coordinate information and the self-capacitance sensing information.

11. The touch system as claimed in claim 10, wherein the computing unit further comprises: A data processing unit, connected to the storage unit and the machine learning model, generates the first touch coordinate information based on the mutual capacitance sensing information.

12. The touch system as claimed in claim 10, wherein the touch panel includes m first directional electrodes and n second directional electrodes, the m first directional electrodes being perpendicular to the n second directional electrodes, the mutual capacitance sensing information including K = m × n mutual capacitance values, the self-capacitance sensing information including L = m + n self-capacitance values, the first touch coordinate information including R first coordinate groups, each of the first coordinate groups including an X coordinate and a Y coordinate, the second touch coordinate information including R second coordinate groups, each of the second coordinate groups including an X coordinate and a Y coordinate, and K, L, and R being positive integers greater than or equal to 1.

13. The touch system as claimed in claim 12, wherein the machine learning model generates the second touch coordinate information by means of obtaining the second touch coordinate information according to the following formula: , , where, i is a positive integer from 1 to R, t is 2R+L, X2i is the X coordinate of the i-th second coordinate group, Y2i is the Y coordinate of the i-th second coordinate group, X11~X1i are the X coordinates of the R first coordinate groups, Y11~Y1i are the Y coordinates of the R first coordinate groups, S1~SL are the L self-capacitive sensing quantities of the self-capacitive sensing information, WX1i~WXti are the weight values ​​of the X coordinates of the i-th second coordinate group, bX1i~bXti are the offset values ​​of the X coordinates of the i-th second coordinate group, WY1i~WYti are the weight values ​​of the Y coordinates of the i-th second coordinate group, bY1i~bYti are the offset values ​​of the Y coordinates of the i-th second coordinate group, and the sign is not specified. This represents the Adama product.

14. A method for improving the polling rate of a touch system, comprising the following steps: A. sequentially performing a mutual capacitance scan on a touch panel to generate mutual capacitance sensing information and a self-capacitance scan to generate self-capacitance sensing information; B. generating a first touch coordinate information based on the mutual capacitance sensing information; and C. inputting the first touch coordinate information and the self-capacitance sensing information into a machine learning model to generate a second touch coordinate information; wherein, The mutual capacitance sensing information includes K mutual capacitance sensing values, the self-capacitance sensing information includes L self-capacitance sensing values, and the first touch coordinate information includes Q first coordinate groups, each of which includes a first X coordinate and a first Y coordinate, where K, L, and Q are positive integers greater than or equal to 1.

15. The method for improving the polling rate of a touch system as described in claim 14, wherein the touch panel includes m first directional electrodes and n second directional electrodes, the m first directional electrodes being perpendicular to the n second directional electrodes, the mutual capacitance sensing information including K = m × n mutual capacitance values, the self-capacitance sensing information including L = m + n self-capacitance values, the first touch coordinate information including R first coordinate groups, each of the first coordinate groups including an X coordinate and a Y coordinate, the second touch coordinate information including R second coordinate groups, each of the second coordinate groups including an X coordinate and a Y coordinate, and K, L, and R being positive integers greater than or equal to 1.

16. The control method as described in claim 15, wherein step C includes obtaining the second touch coordinate information according to the following formula: , , where, i is a positive integer from 1 to R, t is 2R+L, X2i is the X coordinate of the i-th second coordinate group, Y2i is the Y coordinate of the i-th second coordinate group, X11~X1i are the X coordinates of the R first coordinate groups, Y11~Y1i are the Y coordinates of the R first coordinate groups, S1~SL are the L self-capacitive sensing quantities of the self-capacitive sensing information, WX1i~WXti are the weight values ​​of the X coordinates of the i-th second coordinate group, bX1i~bXti are the offset values ​​of the X coordinates of the i-th second coordinate group, WY1i~WYti are the weight values ​​of the Y coordinates of the i-th second coordinate group, bY1i~bYti are the offset values ​​of the Y coordinates of the i-th second coordinate group, and the sign is not specified. This represents the Adama product.

Images