CN111964821A - Pressure touch prediction method and pressure touch prediction model for electronic skin - Google Patents
Pressure touch prediction method and pressure touch prediction model for electronic skin Download PDFInfo
- Publication number
- CN111964821A CN111964821A CN202010777643.2A CN202010777643A CN111964821A CN 111964821 A CN111964821 A CN 111964821A CN 202010777643 A CN202010777643 A CN 202010777643A CN 111964821 A CN111964821 A CN 111964821A
- Authority
- CN
- China
- Prior art keywords
- prediction model
- electronic skin
- pressure
- neurons
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01L—MEASURING FORCE, STRESS, TORQUE, WORK, MECHANICAL POWER, MECHANICAL EFFICIENCY, OR FLUID PRESSURE
- G01L5/00—Apparatus for, or methods of, measuring force, work, mechanical power, or torque, specially adapted for specific purposes
- G01L5/0028—Force sensors associated with force applying means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B82—NANOTECHNOLOGY
- B82Y—SPECIFIC USES OR APPLICATIONS OF NANOSTRUCTURES; MEASUREMENT OR ANALYSIS OF NANOSTRUCTURES; MANUFACTURE OR TREATMENT OF NANOSTRUCTURES
- B82Y30/00—Nanotechnology for materials or surface science, e.g. nanocomposites
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B82—NANOTECHNOLOGY
- B82Y—SPECIFIC USES OR APPLICATIONS OF NANOSTRUCTURES; MEASUREMENT OR ANALYSIS OF NANOSTRUCTURES; MANUFACTURE OR TREATMENT OF NANOSTRUCTURES
- B82Y40/00—Manufacture or treatment of nanostructures
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B7/00—Measuring arrangements characterised by the use of electric or magnetic techniques
- G01B7/003—Measuring arrangements characterised by the use of electric or magnetic techniques for measuring position, not involving coordinate determination
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/03—Arrangements for converting the position or the displacement of a member into a coded form
- G06F3/041—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means
- G06F3/0414—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means using force sensing means to determine a position
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/03—Arrangements for converting the position or the displacement of a member into a coded form
- G06F3/041—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means
- G06F3/045—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means using resistive elements, e.g. a single continuous surface or two parallel surfaces put in contact
Abstract
The invention provides a pressure touch prediction method and a pressure touch prediction model of electronic skin, wherein the pressure touch prediction model comprises the following steps: s1, dividing N virtual lattices on the electronic skin according to sensitivity requirements, and connecting the electronic skin with K probe terminals; s2, applying different forces in the N virtual grids under different states of the electronic skin, and obtaining data detected by the K probe terminals; and S3, inputting the detected data into a deep learning model for training, and obtaining a pressure touch prediction model. The model can accurately identify the state of the electronic skin, such as a conventional state, a stretching state, a folding state and the like, and also can accurately identify the information of the position, the size and the like of the pressure contact force, thereby laying a foundation for subsequent contact application and providing possibility.
Description
Technical Field
The invention relates to the technical field of computer artificial intelligence, in particular to a pressure touch prediction method and a pressure touch prediction model of an electronic skin.
Background
Human skin can sense various external information, for example, when ants crawl on the skin, people can feel itchy feeling; when rainwater drops on the skin, people can quickly obtain a cool feeling; when the automobile touches a wall, the touched body part, the collision strength and the like can be known; when the skin is stretched or twisted by the outside world, the skin can be felt to be stretched or twisted, which are all functions naturally derived from the skin, and the functions mainly depend on the tactile nerves on the skin, and external signals can be acquired through the tactile nerves and transmitted to the brain of people.
As can be seen from the above, an important function of our skin is to sense whether the skin is contacted by the outside and judge the type of the outside contact (conventional press contact, stretching and twisting), and at the same time, the press contact position of the skin and the outside contact surface and the size of the press contact generated during the contact can be positioned, so as to help us to make corresponding response. Of course, the information we perceive on the skin is not limited to the above-mentioned ones, and it can also distinguish more detailed sensory information for each kind of contact, such as a feeling of cold when rainwater drops on our skin, a feeling of itching when mosquitoes bite the skin, and a feeling of pain when hitting a wall.
The flexible material which is similar to the skin and can sense the pressure touch and the corresponding pressure touch position and pressure touch size has a very wide application scene, for example, a flexible keyboard can be manufactured, and a flexible glove can be manufactured to realize a remote control machine. In the field of the existing flexible electronic skin, most of the electronic skin sensors with flexible expression and operation such as bending, folding and stretching of the sensors are in a laboratory stage. The common characteristic of the electronic skin material is that under an unconventional state (such as folding and stretching), the resistance (or capacitance) can linearly change along with the stretching degree/folding degree; meanwhile, in a conventional state and under a certain pressure, the resistance (or capacitance) of the pressure sensor also changes linearly with the pressure. The existing electronic skin material cannot distinguish the state (such as a conventional state, a folding state or a stretching state) of the electronic skin when contacting with the outside, cannot judge the press-contact position, and cannot realize the simulation of real skin due to the fact that the press-contact force is not accurate enough.
Disclosure of Invention
In order to solve the above problems, the present invention provides a pressure touch prediction method and a pressure touch prediction model for an electronic skin, which can accurately predict the electronic skin state, the pressure touch force position, and the pressure touch force magnitude.
The invention provides a pressure touch prediction model of electronic skin, which comprises the following components: s1, dividing N virtual lattices on the electronic skin according to sensitivity requirements, and connecting the electronic skin with K probe terminals; s2, applying different forces in the N virtual grids under different states of the electronic skin, and obtaining data detected by the K probe terminals; and S3, inputting the detected data into a deep learning model for training, and obtaining a pressure touch prediction model.
Preferably, the pressure touch prediction model includes a prediction model for predicting an electronic skin state, a pressure touch force position, and a pressure touch force magnitude.
The invention also provides a pressure touch prediction method of the electronic skin, which comprises the step of predicting by using the pressure touch prediction model.
The invention has the beneficial effects that: according to the invention, N virtual lattices are divided on the electronic skin according to the requirement of sensitivity, and the electronic skin does not need to be modified; meanwhile, a probe terminal is connected to detect the pressure and touch force, and data obtained through detection are input into a deep learning model to be trained to obtain a pressure and touch prediction model; the state of the electronic skin, such as a conventional state, a stretching state, a folding state and the like, can be accurately identified through the prediction model, and information such as the position, the size and the like of the pressure contact force can also be accurately identified, so that a foundation is laid for subsequent contact application, and the possibility is provided.
Drawings
Fig. 1 is a flowchart illustrating steps of obtaining a pressure-touch prediction model of an electronic skin according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of data acquisition according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a prediction model for predicting an e-skin condition according to an embodiment of the present invention.
FIG. 4 is a schematic diagram of a prediction model for predicting a position of a contact force according to an embodiment of the present invention.
FIG. 5 is a schematic diagram of a prediction model for predicting magnitude of a compressive touch force according to an embodiment of the present invention.
Detailed Description
The present invention is described in further detail below with reference to specific embodiments and with reference to the attached drawings, it should be emphasized that the following description is only exemplary and is not intended to limit the scope and application of the present invention.
The electronic skin adopted in the embodiment is a flexible and stretchable material, and is mainly an elastomer nano composite material formed by mixing and solidifying silicon rubber and carbon nano tubes, or the electronic skin containing conductive hydrogel. It is characterized in that under the stretching state, the resistance (or capacitance) can be linearly changed along with the stretching degree; meanwhile, under a certain pressure, the resistance (or capacitance) of the pressure sensor can also change linearly with the pressure. The electronic skin has the characteristic of low price, is very suitable for popularization and use, and simultaneously can be suitable for a plurality of application scenes due to the advantage of flexibility and stretchability. For example, the conventional keyboard is made of a hard material, and each key needs to be matched with a corresponding circuit board to sense whether the key is in a pressed state, so that the keyboard is inconvenient to store, the production and manufacturing costs are relatively high, and the number of the keys needs to be determined in advance (touch sensitivity). The flexible keyboard manufactured by utilizing the characteristics of flexibility, simplicity in manufacturing, low price, adjustable sensitivity and the like of the electronic skin is low in price, can be normally used after being kneaded into a ball and unfolded, and is very convenient to store. In addition, the electronic skin can be manufactured into wearable gloves, the machine can be remotely controlled by sleeving the wearable gloves on hands, workers do not need to visit the site, and great help is provided for construction in some areas with complex geographic environments.
Preparation of electronic skin
Firstly, preparing an electronic skin, and preparing a square mould with the side length of 2cm and the height of about 0.3 cm; of course, the size of the mold can be set according to the size of the electronic skin to be prepared, such as 8cm by 8cm mold. Then pouring the silicon rubber into the mould, mixing the carbon nano tubes or the carbon fiber tubes, and simultaneously, covering the mould, sealing and then putting the mould into a mixer for uniform mixing in order to uniformly mix the carbon nano tubes or the carbon fiber tubes into the silicon rubber. The cover above the mould was then opened and placed in a vacuum pump to degas for 3 minutes and finally placed in a greenhouse for curing. Preferably, the cured product can be placed in a vacuum oven for further curing.
It should be noted that, in this embodiment, only the basic method of the electronic skin is illustrated, and actually, due to different layouts of the carbon fiber tubes of the electronic skin which is randomly manufactured, the expressed attributes such as the resistance are different, which means that for each manufactured electronic skin, an AI model for each electronic skin needs to be trained through a subsequent complex neural network model training process, and the larger the electronic skin is, the higher the accuracy requirement is, the higher the time complexity for training the AI model for each electronic skin is, the higher the training cost is, and the cost is increased invisibly. Therefore, to avoid this problem, the standardized production can be used to realize that the structure of the produced electronic skin is the same, and the AI model obtained by training one of the electronic skins produced in the same batch has universality and can be correctly applied to other electronic skins in the same batch. For standardized production, this example is not described in detail.
Training of pressure touch prediction model
Because the relationship between the resistance signal output of each probe terminal and the pressure contact position/pressure contact force of the pressure contact force on the electronic skin cannot be obtained through a simple mathematical model, an artificial intelligence prediction model is established by utilizing an AI deep learning algorithm.
Based on the electronic skin obtained as above, the present embodiment provides a pressure-touch prediction model of the electronic skin, as shown in fig. 1, which specifically includes the following steps:
101. according to the sensitivity requirement, N virtual grids are divided on the electronic skin, and the electronic skin is connected with K probe terminals.
According to different application scenes, the sensitivity/precision requirements are different. For example, in a scenario with low accuracy requirement, fewer (e.g., 2 × 2) virtual grids may be arranged on a unit of electronic skin, and in a scenario with higher accuracy requirement, more (8 × 8 or 16 × 16) virtual grids may be arranged on the electronic skin, so as to meet the accuracy requirement of different application scenarios. There is a case where: the size of the virtual grid is estimated according to the minimum area of the human touch area, for example, a finger click action is simulated, and the size of the virtual grid is determined according to the minimum touch area of the finger. In addition, a plurality of probe terminals are connected to the edge of the electronic skin to detect resistance signals of the probe terminals when the electronic skin is correspondingly pressed. The probe terminal used herein is an electrode made of carbon fiber for detecting a resistance signal around the electrode.
As shown in fig. 2, on the prepared electronic skin of 2 × 2cm, 2 probe terminals are inserted at regular intervals on each side, 2 × 4 — 8 probe terminals are used for detecting resistance signals around the electronic skin, all the probe terminals are connected to one data collector for resistance data collection, the number of the probe terminals can be increased or decreased appropriately according to the size of the electronic skin and the accuracy requirement of the electronic skin, if the accuracy requirement of the electronic skin of a unit is very high, 4 (or 8) probe terminals can be inserted at regular intervals on each side, but the effect is not better when the number of the probe terminals is larger, the probe terminals are too large, the training time of the model is not increased, and the effectiveness of the model is not greatly facilitated.
102. Under different electronic skin conditions, different forces are applied to the N virtual grids, and resistance data detected by K probe terminals are obtained.
As shown in fig. 2, a total of 4 virtual cells of 2 × 2 are marked in the e-skin, and then 1000 points are randomly selected in each virtual cell by using an electronic universal material testing machine, and the e-skin is pressed and touched at each point with 1KPa and 2KPa. Each experiment recorded the current electronic skin state (regular M1, equilibrium tension M2, oriented fold M3), and the resistance signals of the K probe terminals at different states.
For a scene with high precision requirement, it is not necessary to re-produce the electronic skin, and all that needs to be done is to re-divide the virtual grids, for example, on the same electronic skin, four virtual grids, namely 2 × 2, originally divided are re-divided into 16 virtual grids, namely 4 × 4, or 256 virtual grids, then according to the precision requirement, the number of probe terminals is correspondingly increased, and finally, data of corresponding probe terminals are tested and collected on each virtual grid again. The probe terminal K and the virtual grid N are generally in a direct proportion relationship, but a strict equal proportion growth relationship does not exist, the probe terminal is not suitable for too many, the calculated amount is greatly increased, and the final prediction effect is not obviously improved. If the number of virtual lattices is increased, it is also possible that the number of probe terminals is not increased.
103. And inputting the detected resistance data into a deep learning model for training, and obtaining a pressure touch prediction model for predicting the electronic skin.
For some applications, it may only be necessary to predict the pressure contact position of the pressure contact force, for example, a flexible keyboard, and it is not necessary to predict the magnitude of the pressure contact force, and it is only necessary to predict the position of the pressure contact force applied on the electronic skin to determine which key is pressed by the user; when the electronic skin sensing device is applied to the skin of a future robot, the robot needs to sense the information such as the contact position, force and the like of the outside to the body part of the robot through the electronic skin so as to correctly make corresponding response, and therefore, the position and the magnitude of the pressure contact force need to be predicted; in addition, for different application scenarios, it is also necessary to determine what state the e-skin is in, such as: normal state, stretched state, folded state, etc. Therefore, three prediction models are established and trained in the embodiment, namely the prediction models for predicting the electronic skin state, the pressure contact force position and the pressure contact force magnitude.
The three AI models established in the embodiment are artificial neural network models based on DNN, and in the model training process, the problems of under-fitting or over-fitting are often encountered, the under-fitting indicates that data and scenes are not deeply known, the training data amount is too small or an inappropriate model is used, so that the model does not well understand actual data and has calculation deviation, and the over-fitting refers to the phenomenon of using an excessively complex model to form a class, so that the complex model has no universality.
In order to avoid the problems of under-fitting and over-fitting in the model training process, for different application scenarios and different prediction objects, when the model is designed, the number of neurons in the input layer, the number of hidden layers, the number of neurons in each hidden layer, and the number of neurons in the output layer need to be determined. The number of neurons in the input layer, the number of hidden layers, the number of neurons in each hidden layer, and the number of neurons in the output layer are different in each model. Meanwhile, in the process of model training, certain neurons (Dropout) are discarded randomly, and certain characteristics of data are learned selectively, so that the generalization capability of the model is improved, and the over-fitting condition is avoided.
The number of hidden layers in the three models and the number of neurons in the three models can be adjusted according to the actual effect, and the adjustment process is as follows:
in one case, the accuracy of the algorithm model for predicting the electronic skin state reaches 98%, but the accuracy of the algorithm model in actual use is only 65%, which is called as overfitting, and in this case, the number of hidden layers can be reduced, and the number of neurons in one or more hidden layers can be reduced, so that the complexity of the model is reduced, and the 'three-to-one' capability of the model is improved.
In another case, the accuracy of the model for predicting the electronic skin state in the experimental stage is only 70%, which is called under-fitting, and at this time, it may be tried to increase the number of hidden layers, and at the same time, the number of neurons in one or more hidden layers is increased, so as to increase the complexity of the model, so as to enhance the learning ability of the model.
The process is called as the process of adjusting the parameters, and a new electronic prediction skin state model is finally obtained through training of ' adjusting the parameters ' in prediction ' for a plurality of times.
The prediction model for predicting the electronic skin state has four network layers, as shown in fig. 3, including an input layer, two hidden layers and an output layer; the number of the neurons of the input layer is 8, and the number of the corresponding probe terminals is K; the number of neurons in the first hidden layer is 1560, and the number of neurons in the second hidden layer is 195; the number of neurons in the output layer is 3, corresponding to a conventional state, a balanced stretching state and an oriented folding state. If the electronic skin state is subdivided or other states exist, the number of the neurons of the output layer changes correspondingly, and the number of the neurons of the output layer corresponds to the total number of the types of the electronic skin state.
The prediction model for predicting the position of the pressure contact force has five network layers, as shown in fig. 4, including an input layer, three hidden layers and an output layer; the number of the neurons of the input layer is that the number of the K probes corresponds to the number of the K probes and the electronic skin state data (K +1) predicted by a model; the number of neurons in the first hidden layer is 1560, the number of neurons in the second hidden layer is 780, and the number of neurons in the third hidden layer is 78; the number of neurons in the output layer is 4, corresponding to 2 × 2 virtual lattices. The number of neurons in the output layer corresponds to the number of virtual lattices N. The prediction model for predicting the magnitude of the pressure contact force has four network layers in total, as shown in fig. 5, including an input layer, two hidden layers and an output layer; the number of the neurons of the input layer is 10, and the corresponding number is the total level of the magnitude of the pressure contact force or the sensitivity required by practical application; the first hidden layer has 1560 neurons, and the second hidden layer has 195 neurons; the output layer has 10 neurons, corresponding to 10 levels of pressure contact force. The number of the neurons of the input layer and the output layer is the total number of the levels of the magnitude of the pressure contact force.
In the DNN-based artificial neural network model, commonly used activation functions are sigmod, tanh, relu and the like, nonlinear characteristics are introduced into the neural network, and if no activation function exists, matrix multiplication is performed on each layer of the neural network. In the three prediction models in this embodiment, softMax is used as an activation function of a full link layer, and the three prediction models are abstracted to a multi-class problem. If the classification model obtains the value of each cell by softmax at the full connection layer when the press touch (cell) position is performed, for example, the values obtained for A, B, C, D four cells are 0.5, 0.2, 0.1 and 0.2 respectively, then the cell which we finally obtained the pre-press touch is A.
Dividing the collected resistance data into 3:1 total fours according to the 3 prediction models established in the above way, wherein 3 parts of pressure-touch experimental data are used as training data, 1 part of pressure-touch experimental data are used as prediction data, the pressure-touch experimental data are input into the three prediction models established in the above way for learning, the corresponding trained prediction models are obtained after training, and then the trained prediction models are verified through the prediction data.
For example, for a prediction model of a predicted pressure contact force position, each pressure contact of each group (a group of pressure values are generated from the contact start to the contact separation of each pressure contact) is predicted in a verification set, and a virtual grid position with the highest prediction frequency is selected from each group as a result of the pressure contact force position, so that the prediction accuracy is finally obtained and is 100%. In practical application, ten instantaneous data are collected for each press touch, then the ten instantaneous data are predicted, calculation is carried out according to the predicted position result, the result of the prediction is considered to be the result of the current time when the position with the largest number of times is predicted, specifically, if 10 instantaneous data are collected, the 10 data are input into a prediction model for predicting the press touch force position, wherein 7 times of prediction are the first grid, 2 times of prediction are the second grid, and 1 time of prediction is the third grid, the grid pressed at the current time is considered to be the first grid (because 7>2> 1). For the prediction model for predicting the magnitude of the pressure touch force, certain pressure touch is fixedly applied to the electronic skin, resistance signals of eight probe terminals are collected, a plurality of groups of resistance signals obtained by multiple pressure touches are put into the prediction model for predicting the magnitude of the pressure touch force for prediction, and finally the accuracy rate of a test result also reaches 100%. In general, three trained AI models have very good prediction effect.
In practical application, the state, the pressure contact force position and the pressure contact force magnitude of the electronic skin can be accurately predicted by inputting signals obtained by detecting the probe terminals into the prediction models respectively established for predicting the state, the pressure contact force position and the pressure contact force magnitude of the electronic skin. The accuracy of the three models in the experimental stage and the prediction stage reaches more than 98%.
The state of the electronic skin can be accurately predicted through the prediction model, such as a conventional state, a balanced stretching state (the whole electronic skin is not locally stretched), an oriented folding state (an inherent folding mode exists) and the like, and information such as the position and the size of the pressure contact force can be accurately identified, so that a foundation is laid for subsequent contact application, and the possibility is provided. Meanwhile, the prediction model is fast in training speed and prediction speed.
For the same/same batch of electronic skins, multiple sets of prediction models can be trained for the electronic skins at the beginning according to different accuracies, so that the same/same batch of electronic skins can be suitable for application scenes with various accuracy requirements, which is incomparable to a traditional sensor in an integrated circuit board mode.
The foregoing is a more detailed description of the invention in connection with specific/preferred embodiments and is not intended to limit the practice of the invention to those descriptions. It will be apparent to those skilled in the art that various substitutions and modifications can be made to the described embodiments without departing from the spirit of the invention, and these substitutions and modifications should be considered to fall within the scope of the invention.
Claims (10)
1. A pressure-touch predictive model of electronic skin, comprising:
s1, dividing N virtual lattices on the electronic skin according to sensitivity requirements, and connecting the electronic skin with K probe terminals;
s2, applying different forces in the N virtual grids under different states of the electronic skin, and obtaining data detected by the K probe terminals;
and S3, inputting the detected data into a deep learning model for training, and obtaining a pressure touch prediction model.
2. The pressure touch prediction model of claim 1, wherein the pressure touch prediction model comprises a prediction model for predicting an electronic skin condition, a pressure touch force location, and a pressure touch force magnitude.
3. The pressure-touch prediction model of claim 2, wherein the prediction model for predicting an e-skin condition comprises an input layer, two hidden layers and an output layer; the number of the neurons of the input layer is K; the number of neurons of the first hidden layer in the two hidden layers is 1560, and the number of neurons of the second hidden layer is 195; the number of neurons in the output layer is the total number of categories of the electronic skin state.
4. The press-touch prediction model of claim 2, wherein the prediction model for predicting press-touch force locations comprises an input layer, three hidden layers, and an output layer; the number of the neurons of the input layer is K + 1; the number of neurons of the first hidden layer in the three hidden layers is 1560, the number of neurons of the second hidden layer is 780, and the number of neurons of the third hidden layer is 78; the number of the neurons of the output layer is N.
5. The pressure touch prediction model of claim 2, wherein the prediction model for predicting the magnitude of the pressure touch force comprises an input layer, two hidden layers and an output layer; the number of the neurons of the input layer and the output layer is the total number of the levels of the magnitude of the pressure contact force; the number of neurons in the first hidden layer of the two hidden layers is 1560, and the number of neurons in the second hidden layer is 195.
6. The pressure touch prediction model of claim 2, wherein the pressure touch prediction model employs softMax as an activation function for a fully connected layer.
7. The pressure-contact prediction model of claim 1, wherein the electronic skin is prepared by mixing carbon fiber tubes or carbon nanotubes with silicone rubber.
8. The press-touch prediction model of claim 1, wherein the K probe terminals are evenly distributed at an edge of the electronic skin.
9. The pressure touch prediction model of claim 1, wherein step S2 includes applying different forces in the same virtual grid.
10. A method for predicting pressure contact of electronic skin, which is characterized in that the pressure contact prediction model according to any one of 1-9 is used for prediction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010777643.2A CN111964821A (en) | 2020-08-05 | 2020-08-05 | Pressure touch prediction method and pressure touch prediction model for electronic skin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010777643.2A CN111964821A (en) | 2020-08-05 | 2020-08-05 | Pressure touch prediction method and pressure touch prediction model for electronic skin |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111964821A true CN111964821A (en) | 2020-11-20 |
Family
ID=73363569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010777643.2A Pending CN111964821A (en) | 2020-08-05 | 2020-08-05 | Pressure touch prediction method and pressure touch prediction model for electronic skin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111964821A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112710415A (en) * | 2020-12-01 | 2021-04-27 | 南京理工大学 | High-precision planar piezoresistive sensor system and application method thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103335754A (en) * | 2013-06-19 | 2013-10-02 | 合肥工业大学 | Fully-flexible three-dimensional force tactile sensor |
| CN106482874A (en) * | 2016-11-25 | 2017-03-08 | 中国科学院合肥物质科学研究院 | A kind of positive tetrahedron formula three-dimensional force flexible touch sensation sensor array |
| CN110325949A (en) * | 2016-12-29 | 2019-10-11 | 谷歌有限责任公司 | For predicting to touch the multitask machine learning explained |
| US20190321643A1 (en) * | 2017-01-09 | 2019-10-24 | Industry-Academia Cooperation Group Of Sejong University | Sensing system and sensing method using machine learning |
| US20200012382A1 (en) * | 2019-08-19 | 2020-01-09 | Lg Electronics Inc. | Method, device, and system for determining a false touch on a touch screen of an electronic device |
| CN111079333A (en) * | 2019-12-19 | 2020-04-28 | 华南理工大学 | Flexible touch sensor deep learning sensing method |
-
2020
- 2020-08-05 CN CN202010777643.2A patent/CN111964821A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103335754A (en) * | 2013-06-19 | 2013-10-02 | 合肥工业大学 | Fully-flexible three-dimensional force tactile sensor |
| CN106482874A (en) * | 2016-11-25 | 2017-03-08 | 中国科学院合肥物质科学研究院 | A kind of positive tetrahedron formula three-dimensional force flexible touch sensation sensor array |
| CN110325949A (en) * | 2016-12-29 | 2019-10-11 | 谷歌有限责任公司 | For predicting to touch the multitask machine learning explained |
| US20190321643A1 (en) * | 2017-01-09 | 2019-10-24 | Industry-Academia Cooperation Group Of Sejong University | Sensing system and sensing method using machine learning |
| US20200012382A1 (en) * | 2019-08-19 | 2020-01-09 | Lg Electronics Inc. | Method, device, and system for determining a false touch on a touch screen of an electronic device |
| CN111079333A (en) * | 2019-12-19 | 2020-04-28 | 华南理工大学 | Flexible touch sensor deep learning sensing method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112710415A (en) * | 2020-12-01 | 2021-04-27 | 南京理工大学 | High-precision planar piezoresistive sensor system and application method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ma et al. | An optimal electrical impedance tomography drive pattern for human-computer interaction applications | |
| Chorley et al. | Development of a tactile sensor based on biologically inspired edge encoding | |
| Wen et al. | Wearable multimode sensor with a seamless integrated structure for recognition of different joint motion states with the assistance of a deep learning algorithm | |
| CN106407237A (en) | An online study test question recommendation method and system | |
| WO2012050938A2 (en) | Wearable tactile keypad with stretchable artificial skin | |
| Qiu et al. | Nondestructive identification of softness via bioinspired multisensory electronic skins integrated on a robotic hand | |
| CN111964821A (en) | Pressure touch prediction method and pressure touch prediction model for electronic skin | |
| Wang et al. | A flexible tactile sensor with dual-interlocked structure for broad range force sensing and gaming applications | |
| EP3567529A1 (en) | Sensing system and sensing method using machine learning | |
| CN110414366A (en) | A kind of pressure drag array and pressure distribution matching process based on Dynamic Signal | |
| Van Volkinburg et al. | Development of a wearable controller for gesture-recognition-based applications using polyvinylidene fluoride | |
| Qiu et al. | A multisensory-feedback tactile glove with dense coverage of sensing arrays for object recognition | |
| Zhang et al. | Dig information of nanogenerators by machine learning | |
| So et al. | Epidermal piezoresistive structure with deep learning-assisted data translation | |
| CN113447073B (en) | Bionic flexible modular electronic skin system and control method thereof | |
| Shapiro et al. | Translating affective touch into text | |
| Ruan et al. | Bionic octopus-like flexible three-dimensional force sensor for meticulous handwriting recognition in human-computer interactions | |
| CN208283896U (en) | Detection device and Intelligent glove | |
| Rosset et al. | Leveraging machine learning for arrays of soft sensors | |
| Shih et al. | Classification of components of affective touch using rapidly-manufacturable soft sensor skins | |
| CN102831435B (en) | A kind for the treatment of apparatus of characteristic of button behavior and disposal route | |
| Wang et al. | Tactile sensory response prediction and design using virtual tests | |
| Wang et al. | A ring-shaped hand motion sensor based on triboelectricity | |
| Belk | Soft interdigitated capacitive e-skin enhanced by supervised machine learning | |
| Dai et al. | Smart Insoles for Gait Analysis Based on Meshless Conductive Rubber Sensors and Neural Networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20201120 |