Wearable device employing gesture controlling and system for detecting gesture thereof

Abstract

The present invention relates to a wearable device for gesture control and a gesture recognition system using the same. A silicone adhesive layer attached to the skin of a human arm; a first polyurethane foam-absorbing layer located above the silicone adhesive layer, which absorbs heat and reduces trouble when the silicone adhesive layer is attached to the skin; a module and sensing layer located above the first polyurethane foam-absorbing layer, which detects and transmits EMG of the skin through the first polyurethane foam-absorbing layer; and a second polyurethane foam-absorbing layer located above the module and sensing layer, which absorbs heat and reduces trouble caused by the silicone adhesive layer and the module and sensing layer. and a polyurethane film-high permeability layer located on top of the second polyurethane foam-absorbing layer, which allows water vapor of heat absorbed from the second polyurethane foam-absorbing layer to pass through; wherein the silicone adhesive layer - the high permeability layer is made of an artificial skin-based EMG packaging layer; characterized by.The present invention is characterized by comprising: a square-shaped artificial skin-type packaging layer that adheres to the skin; 8 to 20 electromyography (EMG) array sensors formed in a circular shape and arranged in two rows at equal intervals to measure micro-signals of muscles involved in body movements in real time on the packaging layer; a communication module that transmits electromyography signals through the electromyography array sensors via Bluetooth communication; a server that receives and amplifies detection signals from the electromyography array sensors in conjunction with the communication module and provides gesture analysis information about the user to a smart terminal; a smart terminal that provides information consisting of voltage or current values ​​representing the electromyography sensors to be applied to the user's body from the electromyography array sensors in conjunction with the communication module to an artificial intelligence module, and has a gateway formed therein capable of transmitting and receiving data to receive information by being connected to the server; a rechargeable battery module formed on the electromyography array sensors to supply power for operation; and an artificial intelligence module that checks the recognition type of gesture through the smart terminal and classifies it by gesture.

Description

Technology Field

[0001] The contents disclosed in this specification relate to an artificial intelligence-based array-type EMG (electromyography) system, and more specifically, to an artificial intelligence-based array-type EMG system that detects movement by utilizing electromyography sensors to identify the locations of muscles requiring electromyography signal measurement for each gesture (movement). Background Technology

[0002] Unless otherwise indicated in this specification, the contents described in this section are not prior art for the claims of this application, and are not to be recognized as prior art simply because they are included in this section.

[0003] Generally, electromyography refers to the process of recording the electrical activity of muscles as a graph. Normal muscles have no electrical activity while at rest, but when they contract or are stimulated, electrical activity appears and an electric current is generated. The resulting action potential appears as a wave on the screen of an oscilloscope and is recorded. This recording is called an electromyogram, and since it is equipped with a sound-producing device, the sound can also be heard. However, when there is a disease in the muscle itself, the electromyogram shows reduced amplitude or decreased duration of the action potential, and the shape of the wave becomes more complex.

[0004] In other words, there are methods for recording muscle action potentials as a graph, such as the surface effleurage method, which records by attaching electrodes to the surface of the skin, and the needle electrode method, which records by inserting needles into the muscles, to diagnose abnormalities in motor function.

[0005] Furthermore, electromyography (EMG) is a diagnostic method that uses an electromyograph to check the electrical activity of muscles. By inducing and recording the action potentials of skeletal muscles using extracellular electrodes, it is possible to diagnose various neuromuscular diseases, such as radiculopathy, peripheral nerve disorders, and myopathy. This enables the detection of electrophysiological abnormalities that cannot be diagnosed through imaging tests or blood tests, thereby allowing for the diagnosis, prognosis, and monitoring of the disease's course.

[0006] The recent global COVID-19 pandemic has brought about significant changes to the daily lives of humanity.

[0007] In other words, due to the COVID-19 pandemic, non-face-to-face meetings or gatherings became the mainstream rather than face-to-face, and consequently, an Ontact (a combination of untact and online) related market emerged.

[0008] Therefore, in line with this, there has been a demand for the development of systems using electromyography devices that can increase work efficiency and improve the quality of education.

[0009] When using a remote meeting platform, users should be able to easily use functions such as raising hands, controlling meeting PPTs, capturing the screen, and recording the screen via hand gestures without using a keyboard or mouse. This allows even first-time users to control the platform smoothly, reducing wasted time and enabling seamless communication through beneficial features, thereby improving work efficiency and the quality of education.

[0010] The present invention aims to develop an API interface that can be utilized therein.

[0011] In other words, we aim to develop an AI-based on-contact multimodal interface EMG array armband product that can control a non-face-to-face meeting platform using gestures. Prior art literature

[0012] Reference 1 Republic of Korea Patent Publication No. 2000-0072178

[0013] Reference 2 Republic of Korea Patent Publication No. 2005-0040669

[0014] Reference 3 Republic of Korea Patent Publication No. 2017-0008105

[0015] Reference 4 U.S. Patent No. 11481030 The problem to be solved

[0016] The disclosed content relates to an AI-based on-contact multimodal interface EMG array armband product capable of controlling a non-face-to-face meeting platform using gestures, designed to be worn on the arm, which has the most muscle and distinct movements, in order to accurately recognize gestures.

[0017] We aim to provide a wearable device for gesture control that has an artificial skin-based packaging layer that is not harmful to the skin even when worn for a long time and is easily deformable, and an EMG array sensor module that measures fine signals of gesture-related muscles using EMG signals. means of solving the problem

[0018] A wearable device for gesture control according to an embodiment is,

[0019] First, it provides EMG signals for gesture recognition, sets an appropriate peel strength level to minimize skin irritation when the product is attached, and incorporates moisture permeability control into the packaging layer to minimize itching.

[0020] In addition, to improve the sensitivity of the product's electromyography sensor, increase adhesion to the skin, and prevent the sensor's position from changing with muscle movement, a bio-packaging layer for attaching an EMG array based on artificial skin is provided.

[0021] Specifically, it is as follows.

[0022] 1. Artificial skin-based packaging layer

[0023] It provides improved sensitivity of electromyography sensors and a skin-friendly artificial skin-based packaging layer.

[0024] (1) To minimize skin irritation when the product is applied to the skin, the peel strength is set to an appropriate level and the moisture permeability is controlled in the packaging layer to minimize itching.

[0025] (2) It is important to accurately detect changes in electromyography signals to distinguish gestures. To increase adhesion to the skin and prevent the sensor's position from changing with muscle movement, a biocompatible packaging layer based on artificial skin is provided.

[0026] 2. Electromyography sensor

[0027] (1) Since the electromyography sensor will be attached to the skin, it must be flexible and capable of natural deformation similar to the skin, and ultimately provide an array-type sensor.

[0028] (2) To satisfy the conditions for an electromyography sensor that is attached to the arm for measurement, an elastic structure is applied.

[0029] (3) Considering this, the present invention applies a serpentine structure to a modified spider silk structure that can be applied to a kirigami structure and ensures high areal coverage.

[0030] 3. Fabrication of Biosignal Sensors and Process Optimization via Laser-Based Patterning Process

[0031] (1) A biosignal sensor array is fabricated using a laser-based patterning process based on the designed flexible structure.

[0032] (2) The material is made of polymer materials such as polyimide and PET, which are used as flexible substrates, so that it can be deformed to be similar to the curvature of the skin.

[0033] 4. Array-type electromyography sensor

[0034] (1) It is expected that multiple muscles will move when the user makes a gesture, and a single electromyography sensor is used to optimize the measurement location and range to select the position range of the muscles that need to be measured.

[0035] (2) An array-type electromyography sensor is designed based on the optimized range using a single electromyography sensor.

[0036] More specifically, the wearable device for gesture control described above is,

[0037] A silicone adhesive layer attached to the skin of a human arm;

[0038] A first polyurethane foam-absorbing layer located on top of the above silicone adhesive layer, which reduces trouble when the silicone adhesive layer is attached to the skin and absorbs heat;

[0039] A module and a sensing layer located on top of the first polyurethane foam-absorbing layer, which detect and transmit EMG of the skin through the first polyurethane foam-absorbing layer;

[0040] A second polyurethane foam absorption layer located on top of the above module and sensing layer, which reduces trouble caused by the silicone adhesive layer - the module and sensing layer and absorbs heat; and

[0041] A polyurethane film-high moisture permeability layer located above the second polyurethane foam-absorbing layer and permeating water vapor of heat absorbed from the second polyurethane foam-absorbing layer; comprising

[0042] The above silicone adhesive layer - the above high moisture permeability layer is characterized by being made of an artificial skin-based EMG packaging layer.

[0043] In addition, a gesture recognition system using a wearable device for a gesture controller according to an embodiment is,

[0044] The above-described wearable device for gesture control, that is, an EMG sensor array, primarily selects which features are important for identifying muscle movements through analysis. Then, it is characterized by removing unnecessary noise from the selected features, rapidly extracting the features, and recognizing gestures based on these features.

[0045] Specifically, a gesture recognition system using a wearable device for the gesture controller described above is,

[0046] Wearable device for a gesture controller; and

[0047] Includes a management information processing device; and

[0048] The wearable device for the gesture controller above is

[0049] A silicone adhesive layer attached to the skin of a human arm;

[0050] A first polyurethane foam-absorbing layer located on top of the above silicone adhesive layer, which reduces trouble when the silicone adhesive layer is attached to the skin and absorbs heat;

[0051] A module and a sensing layer located on top of the first polyurethane foam-absorbing layer, which detect and transmit EMG of the skin through the first polyurethane foam-absorbing layer;

[0052] A second polyurethane foam absorption layer located on top of the above module and sensing layer, which reduces trouble caused by the silicone adhesive layer - the module and sensing layer and absorbs heat; and

[0053] A polyurethane film-high moisture permeability layer located above the second polyurethane foam-absorbing layer and permeating water vapor of heat absorbed from the second polyurethane foam-absorbing layer; comprising

[0054] The above silicone adhesive layer - the above high moisture permeability layer is made of an artificial skin-based EMG packaging layer, and

[0055] The above management information processing device is

[0056] Step 1: Defining and registering movement patterns for each gesture, based on the positional range of major muscles when performing a gesture with the aforementioned human arm;

[0057] In the first step above, a second step of collecting movement signals according to the positional range of the major muscles in the wearable device for the gesture controller;

[0058] A third step of calculating a movement pattern using the movement signal of the second step above;

[0059] A fourth step of comparing the movement pattern of the third step above with the registered movement pattern above;

[0060] In the above 4th step, a 5th step of extracting a registered movement pattern corresponding to the movement pattern of the above 3rd step and extracting a gesture of the said movement pattern; and

[0061] It includes a sixth step of recognizing the gesture of the fifth step as a user gesture. Effects of the invention

[0062] According to the embodiments, the product relates to an AI-based on-contact multimodal interface EMG array armband capable of controlling a non-face-to-face meeting platform using gestures, and is designed to be worn on the arm, which has the most muscle and distinct movement, in order to accurately recognize gestures.

[0063] A wearable device for gesture control is provided, having an artificial skin-based packaging layer that is not harmful to the skin even when worn for a long time and is easily deformable, and an EMG array sensor module that measures fine signals of gesture-related muscles using EMG signals.

[0064] In addition, through the above-mentioned wearable device for gesture control, namely the EMG sensor array, features that are important for identifying muscle movements are primarily selected through analysis. Then, unnecessary noise is removed from the selected features, the corresponding features are quickly extracted, and a gesture recognition system using a wearable device for gesture control is provided, which recognizes gestures based on these features.

[0065] Therefore, when using a remote meeting platform, users can easily utilize functions such as raising hands, controlling meeting PPTs, capturing the screen, and recording the screen through hand gestures without using a keyboard or mouse. Even first-time users can control the platform smoothly, reducing wasted time and enabling seamless communication through beneficial features, thereby improving work efficiency and the quality of education. Brief explanation of the drawing

[0066] FIG. 1 is a conceptual diagram of a wearable device for gesture control and a gesture recognition system using the same according to a first embodiment of the present invention. FIG. 2 is a structural diagram of an EMG sensor according to a first embodiment of the device of FIG. 1. Figure 3 is a fabrication diagram of the EMG sensor of Figure 2. FIG. 4 is a drawing of a second embodiment of the EMG sensor of FIG. 2. Fig. 5 is an array diagram of the EMG sensor of Fig. 2. Specific details for implementing the invention

[0067] FIG. 1 is a conceptual diagram of a wearable device for gesture control and a gesture recognition system using the same according to a first embodiment of the present invention.

[0068] Referring to FIG. 1, the wearable device (100) for gesture control according to the first embodiment is an AI-based on-contact multimodal interface EMG array armband product that can control a non-face-to-face meeting platform using gestures. This is a method of wearing it on the arm, which has the most muscles and clear movements, in order to accurately recognize gestures.

[0069] In particular, it has an artificial skin-based packaging layer that is not harmful to the skin even when worn for a long time and is easily deformable, and an EMG array sensor module that measures minute signals of gesture-related muscles using EMG signals.

[0070] To this end, the wearable device (100) for gesture control according to the first embodiment measures an EMG signal for gesture recognition and provides it to a management information processing device (200). In particular, to minimize the occurrence of skin irritation when the product is attached to the skin, the design incorporates a packaging layer that sets an appropriate level of peel strength and minimizes the induction of itching, and controls the moisture permeability.

[0071] In addition, to improve the sensitivity of the product's electromyography sensor, increase adhesion to the skin, and prevent the sensor's position from changing with muscle movement, a bio-packaging layer for attaching an EMG array based on artificial skin is provided.

[0072] Specifically, the wearable device (100) for gesture control according to the first embodiment is as follows.

[0073] 1. Artificial skin-based packaging layer

[0074] It provides improved sensitivity of electromyography sensors and a skin-friendly artificial skin-based packaging layer.

[0075] (1) To minimize skin irritation when the product is applied to the skin, the peel strength is set to an appropriate level and the moisture permeability is controlled in the packaging layer to minimize itching.

[0076] (2) It is important to accurately detect changes in electromyography signals to distinguish gestures. To increase adhesion to the skin and prevent the sensor's position from changing with muscle movement, a biocompatible packaging layer based on artificial skin is provided.

[0077] 2. Electromyography sensor

[0078] (1) Since the electromyography sensor will be attached to the skin, it must be flexible and capable of natural deformation similar to the skin, and ultimately provide an array-type sensor.

[0079] (2) To satisfy the conditions for an electromyography sensor that is attached to the arm for measurement, an elastic structure is applied.

[0080] (3) Considering this, the present invention applies a serpentine structure to a modified spider silk structure that can be applied to a kirigami structure and ensures high areal coverage.

[0081] 3. Fabrication of Biosignal Sensors and Process Optimization via Laser-Based Patterning Process

[0082] (1) A biosignal sensor array is fabricated using a laser-based patterning process based on the designed flexible structure.

[0083] (2) The material is made of polymer materials such as polyimide and PET (Polyethylene Terephalate: saturated polyester resin) which are used as flexible substrates, so that it can be deformed to be similar to the curvature of the skin.

[0084] 4. Array-type electromyography sensor

[0085] (1) It is expected that multiple muscles will move when the user makes a gesture, and a single electromyography sensor is used to optimize the measurement location and range to select the position range of the muscles that need to be measured.

[0086] (2) An array-type electromyography sensor is designed based on the optimized range using a single electromyography sensor.

[0087] First, regarding '1. Artificial Skin-Based Packaging Layer,' that is, firstly, to minimize skin irritation when the product is applied, an appropriate level of peel strength is set, and to minimize itching, it is manufactured using a packaging layer method to regulate moisture permeability, thereby achieving the desired effect.

[0088] To this end, the wearable device (100) for gesture control comprises, largely, a silicone adhesive layer (111), a first polyurethane foam (PU Foam)-absorbing layer (112), a module and sensing layer (113), a second polyurethane foam-absorbing layer (114), and a polyurethane film-high moisture permeability layer (115).

[0089] The above silicone adhesive layer (111) is attached to the skin of a human arm. The adhesive has the property of being able to be repeatedly attached and detached like a post. The silicone adhesive layer (111) has good airtightness and reduces the need to attach with water. Additionally, it can reduce adhesive residue when reattaching and can be easily replaced.

[0090] The first polyurethane foam (PU Foam) absorbent layer (112) is positioned on top of the silicone adhesive layer (111) to absorb heat and reduce skin irritation when the silicone adhesive layer (111) is attached to the skin. Polyurethane foam has low thermal conductivity, which provides an insulating effect. In addition, it is lightweight and reduces the weight of the product, thereby reducing the load and increasing stability. In particular, it has excellent waterproof and moisture-proof effects due to its non-absorbent nature, and is widely used in places sensitive to moisture. Considering that it is applied to the skin, that is, to prevent skin irritation, an open material that allows internal air to flow well may also be used.

[0091] The above module and sensing layer (113) are located on the upper part of the first polyurethane foam-absorbing layer (112) to detect and transmit EMG of the skin through the first polyurethane foam-absorbing layer (112). To this end, the above module and sensing layer (113) includes a wireless communication module. An embedded rechargeable lithium-ion battery may be used for weight reduction. EMG signals measured from a subject wearing an armband may be transmitted via Bluetooth communication, and a Bluetooth 5.0-based low-power protocol (BLE; Bluetooth Low Energy) may be applied.

[0092] The second polyurethane foam-absorbing layer (114) is positioned on top of the module and sensing layer (113) to reduce trouble caused by the silicone adhesive layer (111) - the module and sensing layer (113) and to absorb heat.

[0093] The above polyurethane film-high moisture permeability layer (115) is located on top of the second polyurethane foam-absorbing layer (114) and allows the water vapor of the heat absorbed from the second PU Foam-absorbing layer (114) to pass through to the polyurethane film.

[0094] Next, regarding '1. Artificial Skin-Based Packaging Layer,' secondly, it is important to accurately detect changes in electromyographic signals to distinguish gestures. To increase adhesion to the skin and prevent the sensor's position from shifting with muscle movement, it is made into a biocompatible packaging layer based on artificial skin.

[0095] That is, the above silicone adhesive layer (111) - the above high moisture permeability layer (115) is made of an artificial skin-based EMG packaging layer (110).

[0096] The remaining '2. Electromyography sensor', '3. Fabrication of biosignal sensor and process optimization through laser-based patterning process', and '4. Array-type electromyography sensor' are explained with reference to Fig. 2-4.

[0097] Meanwhile, the system according to the first embodiment largely includes a wearable device (100) for gesture control and a management information processing device (200).

[0098] The system of the first embodiment primarily selects which features are important for identifying muscle movements through analysis using the wearable device (100) for gesture control, i.e., the EMG sensor array. Then, it removes unnecessary noise from the selected features, quickly extracts the features, and recognizes gestures based on these features.

[0099] The above-described wearable device (100) for gesture control is as described above, and when a gesture is performed, it measures movement signals according to the positional range of major muscles and transmits them to the management information processing device (200).

[0100] The above management information processing device (200) identifies what the features for gesture recognition are through the above gesture control wearable device (100) and recognizes various gestures through this. Specifically, the above management information processing device (200) determines and registers movement patterns for each gesture, based on the position range of major muscles when performing a gesture with the human arm. In this state, the above gesture controller wearable device (100) collects movement signals based on the position range of major muscles. A movement pattern is calculated using the collected movement signals. The calculated movement pattern is compared with the above registered movement pattern. As a result of the comparison, the registered movement pattern corresponding to the movement pattern of the third step is extracted, and the gesture of the corresponding movement pattern is extracted. Thus, that gesture is recognized as a user gesture.

[0101] The above management information processing device (200) can select hand gestures that are easiest to use in a non-face-to-face meeting and connect them to each function.

[0102] When using a non-face-to-face meeting platform, you can easily use functions such as raising your hand, controlling meeting PPTs, capturing the screen, and recording the screen through hand gestures without using a keyboard or mouse.

[0103] The above-mentioned management information processing device (200) extracts features from the EMG sensor through a Continuous Wavelet Transform. Wavelet transforms can construct a model of a signal, system, or process as a set of special signals, and these special signals are called wavelets. A single small wavelet existing locally is used as a pattern to represent an arbitrary waveform through a scale of transition, enlargement, or reduction.

[0104] The widely used Fourier transform is sensitive to singularities because its basis is similar, resulting in reduced sensitivity to singularities. To address this, a method was devised to perform time-frequency analysis without compromising the similarity of the basis; this is the wavelet transform. The characteristics of time-frequency analysis using the wavelet transform include high temporal resolution in the high-frequency range and high frequency resolution in the low-frequency range.

[0105] For this reason, the wavelet transform is more effective than the STFT in time-frequency analysis, and can be utilized for extracting key features for gesture recognition. Features extracted through the Continuous Wavelet Transform can be processed through a noise removal deep learning model, such as an Autoencoder, to remove noise and obtain a clearer signal.

[0106] Since the features extracted from this are obtained in the form of one image feature per time window of approximately 260ms, the input features given sequentially in chronological order can be classified using a CNN model to determine the final type of gesture.

[0107] The proposed model consists of a total of 4 CNN layers, and a Slow-fusion ConvNet architecture can be used to learn features given in chronological order simultaneously.

[0108] Dropout regularization is used in all layers except the final layer, and training is performed using the ADAM optimization methodology.

[0109] Figure 2 is a drawing showing the EMG sensor (113) of the device of Figure 1.

[0110] Referring to FIG. 2, the EMG sensor (113) has the following characteristics.

[0111] (1) The above EMG sensor (113), that is, the electromyography sensor, will be attached to the skin, so it must be flexible and capable of natural deformation similar to the skin, and ultimately provide an array-type sensor.

[0112] Accordingly, the size of a single sensor is also manufactured considering the size and signal sensitivity of the final array.

[0113] (2) To satisfy the conditions for an electromyography sensor that is attached to the arm for measurement, an elastic structure is applied.

[0114] (3) In consideration of this, the present invention combines various elastic structures and applies a serpentine structure to a modified spider silk structure that allows for the application of a kirigami structure and ensures high areal coverage.

[0115] To this end, the EMG sensor (113), that is, the module and sensing layer (113), is a flexible structure formed by combining, in order, a deformed spider web structure (1) to a kirigami structure (2), each deformed structure (3) from the kirigami structure (2), a serpentine structure (4) from each deformed structure (3), and a stretchable structure (5) from the serpentine structure (4) with respect to the EMG sensor part (113-1).

[0116] Figure 3 is a manufacturing process diagram of the EMG sensor (113) of Figure 2.

[0117] Referring to FIG. 3, the EMG sensor (113) creates the aforementioned flexible structure through a laser-based patterning process, and the fabrication of the biosignal sensor and process optimization accordingly have the following characteristics.

[0118] (1) A biosignal sensor array is fabricated using a laser-based patterning process based on the designed flexible structure.

[0119] To prevent electrode damage caused by contact with the skin, it may be necessary to design an embedded structure or a method to enhance the adhesion between the electrode and the substrate.

[0120] (2) The material is made of polymer materials such as polyimide and PET (Polyethylene Terephalate: saturated polyester resin) which are used as flexible substrates, so that it can be deformed to be similar to the curvature of the skin.

[0121] Specifically, the method of manufacturing the above EMG sensor unit (113-1) is as follows.

[0122] First, a primary PI layer is formed by first spin-coating polyimide (PI) on top of a flexible substrate.

[0123] The flexible substrate on which the above-mentioned first PI layer was made is baked on a hot plate.

[0124] Ti / Au is deposited on top of the primary PI layer obtained by baking above.

[0125] With respect to the above-deposited Au, the flexible structure of the EMG sensor part (113-1) is patterned by laser ablation.

[0126] The above EMG sensor unit (113-1) is moved to the glass.

[0127] In this state, dipping is done in hot water.

[0128] Cut through PI / Au / Ti / PI.

[0129] A second PI layer is created by performing a second spin coating of PI again.

[0130] So, a stretchable patch is placed on this EMG sensor part (113-1).

[0131] By excluding that patch, only the elastic structure of the EMG sensor part (113-1) is extracted.

[0132] FIG. 4 is a second embodiment of the EMG sensor (113) of FIG. 2.

[0133] Referring to FIG. 4, the EMG sensor (113) according to the second embodiment, that is, the module and the sensing layer (113), may have a resist structure with respect to the EMG sensor part (113-1).

[0134] Even if it is not a flexible Kirigami-Serpentine structure, it can be used in gesture controllers and can have a more stable Resid structure.

[0135] The method of manufacturing the EMG sensor part (113-1) is as follows.

[0136] First, a primary PI layer is formed by first spin-coating polyimide (PI) on top of a flexible substrate.

[0137] The flexible substrate on which the above-mentioned first PI layer was made is baked on a hot plate.

[0138] Ti / Au is deposited on top of the primary PI layer obtained by baking above.

[0139] For the above Au, the lizard structure of the EMG sensor part (113-1) is patterned by laser ablation.

[0140] The above EMG sensor unit (113-1) is moved to the glass.

[0141] And, dip in hot water.

[0142] Cut through PI / Au / Ti / PI.

[0143] Figure 5 is an array diagram of the EMG sensor (113) of Figure 2.

[0144] Referring to FIG. 5, the array EMG sensor (113), that is, the array EMG sensor (113) according to the first embodiment, is an array-type electromyography sensor and has the following features.

[0145] (1) It is expected that multiple muscles will move when the user makes a gesture, and a single electromyography sensor is used to optimize the measurement location and range to select the position range of the muscles that need to be measured.

[0146] (2) An array-type electromyography sensor is designed based on the optimized range using a single electromyography sensor.

[0147] To this end, the array EMG sensor (113), that is, the EMG sensor part (113-1), is an array EMG sensor (113-1') to which the elastic structure or ligament structure of the EMG sensor part (113-1) is applied, corresponding to the positional range of the major muscles when the human arm makes a gesture.

[0148] The number of sensors in the array sensor above is about 5 to 10 per column, totaling 10 to 20, but 16 to 20 would be appropriate.

[0149] This packaging layer is similar in shape to an armband-style arm sleeve and has a size large enough to wrap around the user's wrist and a certain elasticity.

[0150] Therefore, since the above packaging layer must come into direct contact with the human skin, it is formed in the form of artificial skin. This is because it is intended to distinguish gestures and accurately detect changes in electromyography, and to ensure that the position of the electromyography sensor does not change according to muscle movement.

[0151] In addition, the electromyography (EMG) array sensor formed on the packaging layer is a sensor that records the electrical activity of muscles as a graph, and detects the locations of muscles requiring EMG signal measurement for each gesture, or movement.

[0152] The above electromyogram array sensor detects electromyogram signals indicating the degree of muscle activity through electrodes having different polarities of the sensor.

[0153] The effect of such gesture recognition is that, for example, when conducting a non-face-to-face meeting from a remote location, actions such as raising a hand, capturing the screen, and swiping through the screen can be easily recognized by detecting subtle muscle movements without the need to specifically prepare and use a keyboard and mouse in advance.

[0154] Through this, the management information processing device (200) identifies what the features for gesture recognition are through the gesture control wearable device (100), and recognizes various gestures through this.

[0155] That is, the above-mentioned management information processing device (200) first determines and registers a movement pattern for each gesture, based on the position range of the major muscles when performing a gesture with the human arm.

[0156] In this state, the wearable device (100) for the gesture controller collects movement signals according to the position range of the major muscles.

[0157] A movement pattern is calculated using the collected movement signal. The calculated movement pattern is compared with the registered movement pattern.

[0158] Based on the comparison result above, a registered movement pattern corresponding to the movement pattern of the third step is extracted, and a gesture of the corresponding movement pattern is extracted.

[0159] Therefore, that gesture is recognized as a user gesture. Explanation of the symbols

[0160] 100 : Wearable device for gesture control 200 : Management information processing unit 111: Silicone adhesive layer 112: First polyurethane foam (PU Foam) - absorbent layer 113: Module and sensing layer 114: Second polyurethane foam-absorbing layer 115 : Polyurethane film - high moisture permeability layer

Claims

Claim 1 A silicone adhesive layer (111) attached to the skin of a human arm; a first polyurethane foam (PU Foam) absorption layer (112) located above the silicone adhesive layer (111) to absorb heat and reduce trouble when the silicone adhesive layer (111) is attached to the skin; a module and sensing layer (113) located above the first PU Foam absorption layer (112) to detect and transmit EMG of the skin through the first PU Foam absorption layer (112); and a second PU Foam absorption layer (114) located above the module and sensing layer (113) to absorb heat and reduce trouble caused by the silicone adhesive layer (111) and the module and sensing layer (113). A wearable device for gesture control, comprising: a PU film-high permeability layer (115) positioned above the second PU foam-absorbing layer (114) and permeating water vapor of heat absorbed from the second PU foam-absorbing layer (114); wherein the silicone adhesive layer (111) and the high permeability layer (115) are made of an artificial skin-based EMG packaging layer (110). Claim 2 A wearable device (100) for a gesture controller; and a management information processing device (200); wherein the wearable device (100) for a gesture controller comprises: a silicone adhesive layer (111) attached to the skin of a human arm; a first PU foam-absorbing layer (112) located above the silicone adhesive layer (111) to reduce trouble and absorb heat when the silicone adhesive layer (111) is attached to the skin; a module and sensing layer (113) located above the first PU foam-absorbing layer (112) to detect and transmit EMG of the skin through the first PU foam-absorbing layer (112); and a second PU foam-absorbing layer (114) located above the module and sensing layer (113) to reduce trouble and absorb heat caused by the silicone adhesive layer (111) - the module and sensing layer (113). and a PU film-high permeability layer (115) positioned above the second PU foam-absorbing layer (114) and permeating water vapor of heat absorbed from the second PU foam-absorbing layer (114); wherein the silicone adhesive layer (111) and the high permeability layer (115) are made of an artificial skin-based EMG packaging layer (110), and the management information processing device (200) performs a first step of determining and registering a movement pattern for each position range of major muscles when performing a gesture with the human arm for each gesture; a second step of collecting movement signals for each position range of major muscles from the wearable device (100) for the gesture controller in the first step; a third step of calculating a movement pattern using the movement signals from the second step; a fourth step of comparing the movement pattern of the third step with the registered movement pattern; and in the fourth step, extracting the registered movement pattern corresponding to the movement pattern of the third step, and the corresponding A gesture recognition system using a wearable device for a gesture controller, comprising: a fifth step of extracting a gesture of a movement pattern; and a sixth step of recognizing the gesture of the fifth step as a user gesture.

Images