TWI657352B - Three-dimensional capacitive wear human-computer interaction device and method thereof - Google Patents

Abstract

A three-dimensional capacitive wearable human-machine interaction device and method. The three-dimensional capacitive wearable human-machine interactive device includes a capacitive sensing unit, a power unit, an identification unit and a transmission unit. The capacitance sensing unit can collect changes in the surrounding capacitance. The power unit supplies three-dimensional capacitive wearable human-machine interactive device power. The recognition unit records both the time unit and the sensing capacitance value. The personally designated motion trajectory uses real-time recognition. The algorithm obtains the identification purpose of trajectory characteristics to determine the identity of the user, and the transmission unit interfaces the calculation result of the identification unit to other devices. The three-dimensional capacitive wearable human-computer interactive device allows users to control the back-end device according to different gestures. Its remote capacitance sensing method detects the capacitance change generated by the palm in real time, and returns the current capacitance value to the recognition unit to complete The effect is identified, and the finger sleeve design is adopted to allow the user to intuitively control and facilitate the installation of the sensing position.

Description

Three-dimensional capacitive wearable human-machine interaction device and method

The invention relates to a non-contact capacitive amplification measurement method and real-time identification technology. In particular, the invention relates to a three-dimensional capacitive wearable human-machine interaction device and method.

With the advancement of information and communication technology, many devices or equipment in modern life have gradually been controlled by real-time calculation technology.

The conventional human-computer interaction interface makes relative control instructions for different gestures of the user. In the common gesture interaction technology of the conventional technology, real-time camera image recognition and myoelectric signal acquisition are common. The disadvantage is that the camera needs The identification and calculation processor in the specified space are expensive. Furthermore, the EMG signal is a non-linear time-varying signal, which requires a high-standard processor for signal processing, and the EMG signal needs to be fixed close to the skin to measure. There are many inconveniences to use, so the development of human-computer interaction interface will inevitably require more novel designs.

According to the Republic of China Announcement No. 201626166 "Gesture-based 3D Image Manipulation Technology" patent case, the embodiment of this patent is related to The three-dimensional image should be controlled based on user gestures. Examples include rendering three-dimensional images based on three-dimensional image data, receiving sensor data corresponding to a user's gesture, and identifying the use based on the sensor data. Gesture. Based on the identified user gesture and the three-dimensional image data, a function may be determined and executed to update the corresponding three-dimensional image data, and thus change the rendered three-dimensional image. In addition, the three-dimensional image data of this patent may be generated by a sensor coupled to one of the same computing devices used to render the three-dimensional images. As can be seen from the above, the patent proposes to recognize the gesture state by image processing to achieve accurate gesture analysis. However, the disadvantage of using the camera as a sensing element for the sensor cannot reduce the device cost.

In addition, according to the Republic of China Publication No. 201205404 "Three-dimensional Touch Sensor and Its Application" patent, this patent provides a three-dimensional touch sensor, which uses a two-dimensional capacitive touch sensor with a conductive layer and an elastic insulator or The insulating layer is constructed with an elastic conductive object. When the three-dimensional touch sensor is touched, the two-dimensional capacitive touch sensor locates the position of the contact point on the sensing plane. The elastic insulator or elastic conductive object responds to Deformation due to pressure causes a change in capacitance. From this change in capacitance, an induction value in the vertical direction can be obtained, which is related to the magnitude of the pressure. The patent proposes to use a two-dimensional capacitive touch sensor with a conductive layer and an elastic insulator or an insulating layer and an elastic conductive object to sense the touch position of a finger, and then push the two-dimensional matrix and the force perpendicular to the finger. The disadvantage is that the mechanism design needs to be equipped with an elastic conductive object to sense the vertical direction. Furthermore, it is a contact-type sensing method, in which the user needs to touch the device to perform sensing identification.

It can be seen that there are still many shortcomings in the above-mentioned customary methods. It is not a good design and needs to be improved. Therefore, how to find a real-time computing algorithm for human-computer interaction, in particular, it can reduce the existing contact restrictions and high-cost hardware Disadvantages such as performing calculations, and being able to immediately analyze the user's gestures will become the goal of those skilled in the art.

The purpose of the present invention is to use a capacitive effect detection structure to detect a change in capacitance, and through a real-time identification function, a low-cost, more flexible, convenient and intuitive human-machine interaction device is constructed.

The invention proposes a three-dimensional capacitive wearable human-machine interactive device, which includes: a power unit for providing power required by the three-dimensional capacitive wearable human-machine interaction device; and a capacitive sensing unit for detecting the three-dimensional capacitance. The capacitance change around the wearable human-machine interactive device; the identification unit is used to calculate the feature vector of the three-dimensional capacitive wearable human-machine interactive device based on the capacitance change detected by the capacitive sensing unit, and An analysis result of the feature vector is used to obtain an identification result about the movement trajectory of the three-dimensional capacitive wearable human-machine interactive device; and a transmission unit is used to transmit the identification result to the back-end device.

In one embodiment, the capacitive sensing unit is composed of one of a printed circuit board (PCB) hard circuit board and a soft circuit.

In another embodiment, the capacitive sensing unit is presented as a single metallic node or finger-type mechanism sensing design.

In yet another embodiment, between the capacitive sensing unit and the identification unit Wireless or wired interface is used as the communication interface.

In still another embodiment, the identification unit uses a microprocessor or a processor equipped with an operating system as the computing unit.

In yet another embodiment, the identification unit and the transmission unit use a wireless or wired interface as a communication interface.

In yet another embodiment, the transmission unit and the back-end device use a wireless or wired interface as a communication interface.

In another embodiment, the moving trajectory refers to a person obtained by the recognition unit analyzing the feature vector using a recognition algorithm.

The invention proposes a three-dimensional capacitive wearable human-machine interaction method, the steps of which include: detecting the capacitance change of the three-dimensional capacitive wearable human-machine interactive device in response to a specific trajectory drawn by a user, and using feature values to vectorize The method calculates a feature vector; inputs the feature vector to a fuzzy system and performs signal fuzzification processing to obtain a fuzzified signal; performs the normalization processing on the fuzzified signal using an average weight method; and executes the fuzzified signal through a fuzzy rule base Output defuzzification processing to generate an output signal; obtain an error value from the output signal to perform fuzzy rule training; record the trained fuzzy rules; and use the trained fuzzy rules to determine the use of the three-dimensional capacitive formula through the identification algorithm The identity of the user wearing the human-machine interactive device.

In the above method, the specific trajectory refers to a user generated by the user performing an arbitrary movement of his finger.

In the above method, the eigenvalue vectorization method includes Euclidean distance, Minkowski distance, or Markov distance. (Mahalanobis).

In the above method, the blurring process is a function of compressing the input signal of the feature vector to a value of -1 to 1.

In the above method, the normalization process includes using an average weight method, a center of gravity method, a sum center method, or a maximum average method.

In the above method, the output defuzzification processing may adopt a semantic (Mamdani) fuzzy rule or a functional (Sugeno) fuzzy rule.

In the above method, the recognition algorithm records and trains the feature vector, and stores it in a recognition framework training expert knowledge base for subsequent recognition.

The three-dimensional capacitive wearable human-machine interactive device and method of the present invention, wherein the capacitance sensing unit senses changes in the surrounding capacitance and transmits them to the identification unit, thereby achieving the purpose of online real-time feedback identification, and the power unit can supply three-dimensional capacitive wear Human-computer interaction device power, real-time calculation technology is used in the recognition unit to calculate the user's motion trajectory to achieve the recognition effect. During operation, the user can freely set different gesture control trajectories, and use the finger to trace specific trajectories in advance to let the recognition unit record the desired trajectory As the user ’s gestures change, the sensed capacitance is also different, and the movement behavior is used as the basis for identification. The feature points of the movement trajectory are captured to achieve the purpose of allowing the user to easily control the human-machine interaction. Finally, the transmission unit transmits the identification calculation results to other back-end devices to integrate various types of equipment for control purposes.

1‧‧‧Three-dimensional capacitive wearable human-machine interactive device

11‧‧‧Capacitive sensing unit

12‧‧‧Identification unit

13‧‧‧Transmission Unit

14‧‧‧ Power supply unit

15‧‧‧node

C1‧‧‧ Internal capacitor

C2‧‧‧body capacitor

S1 ~ S7‧‧‧step

FIG. 1 is a hardware architecture diagram of a three-dimensional capacitive wearable human-machine interactive device according to the present invention; Figure 2 is a diagram of the detection range of a non-contact capacitive capacitance of the present invention; Figure 3 is a finger-type sensing architecture diagram of a three-dimensional capacitive wearable human-machine interactive device; and Figure 4 is a three-dimensional capacitive wearable human-machine of the present invention An interactive device installation method; and FIG. 5 is a flowchart of the operation framework of the three-dimensional capacitive wearable human-computer interaction method of the present invention.

The technical content of the present invention will be described below with specific embodiments. Those skilled in the art can easily understand the advantages and effects of the present invention from the content disclosed in this specification. However, the present invention can also be implemented or applied in other specific embodiments.

Please refer to FIG. 1, which shows a hardware architecture diagram of a three-dimensional capacitive wearable human-machine interactive device according to the present invention. As shown in the figure, the three-dimensional capacitive wearable human-machine interactive device 1 includes a capacitive sensing unit 11, a recognition unit 12, a transmission unit 13, and a power supply unit 14.

The capacitance sensing unit 11 is configured to detect a capacitance change around the three-dimensional capacitive wearable human-machine interactive device. In simple terms, the capacitance sensing unit 11 can collect the surrounding capacitance changes in real time, and has a wireless or wired communication interface to interface with the identification unit 12. The composition of the capacitance sensing unit 11 can be a PCB (Printed circuit board) hard circuit Board, soft circuit, and in addition, the capacitive sensing unit 11 can be presented by a single metallic node or finger-type mechanism sensing design.

The identification unit 12 is configured to detect the capacitance detected by the capacitive sensing unit 11. The amount of capacitance change is used to calculate the feature vector of the three-dimensional capacitive wearable human-machine interactive device 1, and the feature vector is analyzed to obtain the identification result of the movement trajectory of the three-dimensional capacitive wearable human-machine interactive device 1. Specifically, the identification unit 12 can detect signals of the capacitance sensing unit 11 to detect the capacitance trajectory in real time, thereby achieving the purpose of gesture recognition. The identification unit 12 has a wireless or wired communication interface that can be connected to the transmission unit 13 .

The transmitting unit 13 is configured to transmit the recognition result to a back-end device (not shown). Specifically, the transmission unit 13 is a result obtained by analyzing the amount of capacitance change, and transmits it to other devices on the back end by wireless or wired means, thereby controlling other devices on the back end.

The power supply unit 14 can provide the power required by the three-dimensional capacitive wearable human-machine interactive device, that is, the power supply unit 14 can provide power to the capacitance sensing unit 11, the identification unit 12, and the transmission unit 13, so that the capacitance sensing unit 11, the identification unit 12. The transmission unit 13 can operate.

See Figure 2 for a diagram of the non-contact capacitive detection range. As shown in the figure, the capacitive sensing unit 11 uses the capacitive effect to sense the positions of the surrounding user's fingers and other finger-type capacitive sensing units, that is, the capacitive sensing unit 11 can sense its relative position relationship with the user's fingers. It can also sense the relative positional relationship with other surrounding finger-type capacitive sensing units.

Specifically, based on the capacitance effect, the internal capacitance C1 and the human body capacitance C2 of the capacitance sensing unit are connected in parallel with each other, so that the purpose of detecting the change in capacitance can be achieved. This method can be used to detect the distance of objects. Solve applications based on different gestures in different situations.

Refer to Figure 3, which shows a 3D capacitive wearable human-machine interactive device Fingertip-type sensing architecture diagram. As shown in the figure, the three-dimensional capacitive wearable human-machine interactive device 1 can be presented in the form of a finger sleeve. The periphery of the three-dimensional capacitive wearable human-machine interactive device 1 includes a plurality of nodes for collecting capacitance in the capacitance sensing unit 11. 15. The nodes 15 have the ability to sense changes in surrounding capacitance from a long distance.

Please refer to FIG. 4, which shows a method for installing a three-dimensional capacitive wearable human-machine interactive device. As shown in the figure, the three-dimensional capacitive wearable human-machine interactive device 1 of the present invention can be placed on a user's finger. When the three-dimensional capacitive wearable human-machine interactive device 1 enables a sensing function, the three-dimensional capacitive wearable human-machine interactive device 1 1 can start to sense the surrounding capacitance change, and collect relevant data in real time for subsequent built-in algorithms to perform identification to obtain the identification results.

The steps of the identification algorithm executed by the identification unit 12 are as follows. First, the identification unit 12 is set in a learning training mode, records the eigenvalue strength of the motion trajectory capacitance to be identified, and establishes a eigenvector value through a eigenvalue vectorization method. The eigenvalue vectorization method can be European Distance (Euclidean), Minkowski distance, or Mahalanobis calculation method. In this embodiment, Euclidean calculation is used as an example. The vectorized expression is as shown in the following formula (1):

Among them, there are c capacitive sensing nodes, which are a ₁ to a _c , a ₁ ( n ) is the capacitance change detected by the capacitive sensing unit a ₁ at time n , and a ₂ ( n ) is the capacitive sensing. The capacitance change detected by the measurement unit a ₂ at the time n , and so on, a _c ( n ) is the capacitance change detected by the capacitance sensing unit a _c at the time n .

The motion trajectory capacitance eigenvalues are trained using fuzzy expert knowledge, and the fuzzy attribution function can be expressed by the following formula (2):

Among them, exp is an exponential function, v _m is a function vertex of the fuzzy attribution function, d _m is a function width of the fuzzy attribution function, and m is the number of fuzzy rules. In addition, the weighted average method is also used as the normalization of the assignment function, as shown in the following formula (3):

among them, . Finally, the output obtained through fuzzy rules is as follows (4):

Among them, β is a fuzzy rule, and its initial values are all 0. In the training of fuzzy systems, the energy convergence method is used to converge the parameters of the attribution function and the fuzzy rules. The defined energy function is as follows (5):

After the gradient differential convergence, the updated values can be obtained as shown in equations (6) to (8):

Where η > 0. Finally, when the difference between the output inference σ and the given target x _d reaches the expected value, the training mode can be ended.

After completing the above learning and training mode, when the recognition unit 13 of the three-dimensional capacitive wearable human-machine interactive device 1 is set to the recognition mode, it uses the performance index calculation to obtain the recognition conclusion, as shown in the following formula (9):

Among them, x _d is the training target.

Please refer to FIG. 5, which shows a flowchart of the operation structure of the three-dimensional capacitive wearable human-computer interactive device, which mainly describes the program calculation process of the identification unit 12. In step S1, the capacitance sensing value is collected, that is, the collected feedback capacitance strength is used first, and then the process proceeds to step S2, where the capacitance sensing value is vectorized, that is, the identification unit 12 performs vectorization on the capacitance strength, and then, proceeds In step S3, the system enters a fuzzy system.

At this time, if the three-dimensional capacitive wearable human-computer interactive device 1 is set to the training mode (ie step S4), then it will proceed to step S5, that is, the feature vector fuzzy rule training, that is, the recognition algorithm is used to perform the fuzzy rule training. In addition, if the three-dimensional capacitive wearable human-machine interactive device 1 is in the recognition mode (ie step S6), then it will proceed to step S7, that is, the capacitance value vector is input to the fuzzy system and the result is determined. In the trained fuzzy system, the results are judged by the performance indicators, so that To achieve the effect of identifying security.

From the above description, it can be known that the three-dimensional capacitive wearable human-machine interaction method disclosed in the present invention is characterized by calculating the capacitance vector values detected by a plurality of groups of devices and calculating the capacitance vector values, and the vectors at each time point The value is input to the fuzzy system for identification. Due to the user's special movement habits, the depth of the trajectory movement and the time and speed change when the same trajectory is drawn. Therefore, the feature vector can be used to identify the identification algorithm. Its identity.

Taking a finger-type capacitive sensing device as an example, the above method includes the following steps: (a) Setting the finger-type capacitive sensing device in the recognition mode, and the user draws a specific trajectory as the basis of the identification record; (b) The finger-capacitive capacitance sensing device obtains the respective capacitance values of the capacitance sensing units thereon, and uses the eigenvalue vectorization method to calculate the eigenvector values; (C) inputs the eigenvector values into the fuzzy system for signal blurring (D) Normalize the signals that have been fuzzified using the average weighting method; (E) Defuzzify the fuzzed signals through the fuzzy rule base; (F) Output the signals Obtain error values to perform fuzzy rule training; (G) record fuzzy rules after training is completed; and (H) enable the recognition mode of the finger-capacitive sensing device, and obtain whether the correct gesture is obtained based on the performance indicators to achieve the recognition effect.

As can be seen from the above, the three-dimensional capacitive wearable human-machine interactive device and method thereof according to the present invention sense different capacitance values according to different gestures of a user. The identification algorithm uses both time and capacitance sensing values to perform record recognition and then control. End-connected devices, because of the built-in identification function, the identification unit does not need to send data to the cloud background to achieve user identification and improve system security of. Furthermore, in order to make it easy for users to use the device described in the present invention, the present invention adopts a finger sleeve design for the capacitive sensing device, and can place different fingers to make gesture control instructions according to the user's needs, improving the traditional wearable gesture recognition device , The user needs to stabilize the sensing device mechanism at a specific position and the cost of the required calculation processor is expensive.

In addition, the present invention can improve the traditional three-dimensional gesture sensing method. The capacitive measurement method has a different principle than the inductive sensing method. The capacitive effect method is used to achieve the measurement purpose. Based on the huge human body capacitive effect, the resolution of the sensing signal can also be high. Because of the inductive type, the capacitive effect for different gestures and the user can be significantly different, so that the recognition gesture signal has unique characteristics. Furthermore, the capacitive measurement method is also superior to the myo-inductive measurement method in biological gesture sensing technology. Its EMG signal is a non-linear time-varying problem, and the capacitive measurement method is a linear non-time-varying problem. Therefore, the identification unit does not Expensive computing cores are required, which can reduce the cost of wearable devices.

Another advantage of the present invention is that the above-mentioned behavior can be operated when the network is offline. Since the capacitive signal is a linear and time-invariant problem, it does not require any network service function to achieve local recognition. The present invention proposes to use the human capacitive effect in conjunction with the wear sensing method, which is convenient for users to install and operate. In addition, three-dimensional capacitive wear The installation position of the human-machine interactive device can be set according to the user's needs. It does not need to be installed at the specified position as in the conventional technology. In addition, the user can use the gesture drawing trajectory to perform feature recognition comparison, so it can achieve an intuitive The identification method, in addition, the use of a microprocessor for real-time calculation does not require an Internet connection, which can reduce costs and increase the benefits of commercial applications.

Compared with the conventional technology, the traditional technology is mainly aimed at the capacitive two-dimensional recognition technology or the use of high-priced sensing elements to achieve the discrimination function. However, in the development of more and more diverse information technology and various terminal devices for different gesture recognition sensing methods Under development, if the user's gesture movement space is limited, and the restriction device must be placed in a designated fixed position, it will cause a lot of inconvenience in use. In contrast, in order to achieve the performance of controlling through gestures, and considering the use of the device and the ability to control a large number of back-end devices, characteristics such as flexibility, easy expansion, and easy installation are more important. Therefore, the three-dimensional capacitive type proposed by the present invention Compared with conventional technology, wearing human-machine interactive devices has the following advantages: (1) the present invention can be used in various fingers according to different control needs of users, and can be flexibly used in various situations; (2) the present invention is Adopting capacitive sensing design, the analytical signal is a linear non-time-varying problem, and the identification effect can be achieved without using a high-order arithmetic processor; (3) The present invention adopts a long-distance identification method in the capacitive sensing unit without contact To the capacitive touchpad, compared with the conventional technology using the contact method is more limited; (4) the invention can set different styles of gesture changes by the user, the training recognition model can be different; (5) this The identification algorithm described in the invention records and recognizes both the time unit and the capacitance value, so that gesture movement or stillness can be achieved as a training model; (6) the invention Better than the inductance measurement method, the present invention improves the resolution of the sensing signal so that the subsequent identification unit can perform differential identification by using the human body as a huge capacitance characteristic and the capacitive effect generated by the sensing board; (7) The present invention is excellent In the gesture electromyography method, that is, the signal collection function can be achieved without touching the human skin.

The above detailed description is for a possible embodiment of the present invention. It should be noted that this embodiment is not intended to limit the patent scope of the present invention, and equivalent implementations or changes that do not depart from the technical spirit of the present invention should be included in the patent scope of this case.

Claims (15)

A three-dimensional capacitive wearable human-machine interactive device includes: a power unit for providing power required by the three-dimensional capacitive wearable human-machine interaction device; and a capacitive sensing unit for detecting the three-dimensional capacitive wearer. The amount of capacitance change caused by a finger or another capacitive sensing unit around the interactive device according to a specific trajectory drawn by the user; the identification unit is used to calculate the capacitance change based on the capacitance change detected by the capacitive sensing unit. Calculate the feature vector of the three-dimensional capacitive wearable human-machine interactive device, and analyze the feature vector to obtain the identification result of the movement trajectory of the three-dimensional capacitive wearable human-machine interactive device; and a transmission unit for The recognition result is sent to the back-end device. The three-dimensional capacitive wearable human-machine interactive device according to item 1 of the scope of patent application, wherein the capacitive sensing unit is formed by one of a printed circuit board (PCB) hard circuit board or a soft circuit. The three-dimensional capacitive wearable human-machine interactive device described in item 1 of the scope of the patent application, wherein the capacitive sensing unit is presented by a single metallic node or finger-type mechanism sensing design. The three-dimensional capacitive wearable human-machine interactive device according to item 1 of the scope of patent application, wherein between the capacitive sensing unit and the identification unit, between the identification unit and the transmission unit, or between the transmission unit and the back end Devices use a wireless or wired interface as a communication interface. The three-dimensional capacitive wearable human-machine interactive device according to item 1 of the scope of the patent application, wherein the identification unit is a microprocessor or a processor equipped with an operating system as a computing unit. According to the three-dimensional capacitive wearable human-machine interactive device described in item 1 of the scope of the patent application, a wireless or wired interface is used as the communication interface between the identification unit and the transmission unit. The three-dimensional capacitive wearable human-machine interactive device described in item 1 of the scope of patent application, wherein the transmission unit and the back-end device use a wireless or wired interface as a communication interface. According to the three-dimensional capacitive wearable human-machine interactive device described in item 1 of the scope of the patent application, the movement trajectory refers to the feature unit obtained by analyzing the feature vector with a recognition algorithm. A three-dimensional capacitive wearable human-machine interaction method, comprising the steps of detecting a change in capacitance generated by a finger or another capacitive sensing unit around the three-dimensional capacitive wearable human-machine interactive device in response to a specific trajectory drawn by a user, Use the eigenvalue vectorization method to calculate the eigenvector; input the eigenvector to the fuzzy system and perform signal fuzzification processing to obtain the fuzzified signal; use the average weight method to perform the normalization processing on the fuzzified signal; The signal is subjected to an output defuzzification process by the fuzzy rule base to generate an output signal; an error value is obtained from the output signal to perform fuzzy rule training; a trained fuzzy rule is recorded; and an identified algorithm is used according to the trained fuzzy rule through the identification algorithm Determine the identity of the user using the three-dimensional capacitive wearable human-machine interactive device. According to the three-dimensional capacitive wearable man-machine interaction method described in item 9 of the scope of the patent application, the specific trajectory refers to a user generated by the user's arbitrary movement of his finger. The three-dimensional capacitive wearable human-computer interaction method according to item 9 of the scope of the patent application, wherein the eigenvalue vectorization method includes Euclidean, Minkowski, or Mahalanobis. The three-dimensional capacitive wearable human-computer interaction method as described in item 9 of the scope of patent application, wherein the fuzzification process is a function of compressing the input signal of the feature vector to a value of -1 to 1. The three-dimensional capacitive wearable human-computer interaction method according to item 9 of the scope of patent application, wherein the normalization process includes adopting an average weight method, a center of gravity method, a sum center method, or a maximum average method. For example, the three-dimensional capacitive wearable human-computer interaction method described in item 9 of the scope of patent application, wherein the output defuzzification processing may adopt a semantic (Mamdani) fuzzy rule or a functional (Sugeno) fuzzy rule. The three-dimensional capacitive wearable human-computer interaction method as described in item 9 of the scope of the patent application, wherein the recognition algorithm records and trains the feature vector and stores it in a knowledge base for recognition architecture training experts required for subsequent recognition.