KR102120263B1 - Flexible wireless charging apparatus based on artificial intelligence - Google Patents

Abstract

The present invention relates to an artificial intelligence-based flexible wireless charging device, which transmits wireless power to a wireless power receiving module and has wireless flexibility that is variable based on characteristics including pressure, temperature, or image of the wireless power receiving module. It includes a power transmitter and a non-integral power converter connected to the wireless power transmitter through a wired cable. Accordingly, the present invention can be applied to a plurality of user terminals having various sizes and shapes through an artificial intelligence-based sensor.

Description

Flexible wireless charging device based on artificial intelligence {FLEXIBLE WIRELESS CHARGING APPARATUS BASED ON ARTIFICIAL INTELLIGENCE}

The present invention relates to an artificial intelligence-based flexible wireless charging technology capable of multi-charging, and more specifically, an artificial intelligence that has an artificial intelligence-based sensor and can be applied to a plurality of user terminals of various sizes and shapes. It relates to an intelligent-based flexible wireless charging device.

With the development of Deep Learning technology, its application fields are continuously expanding. In the past, it has been mainly implemented in the form of software or digital circuits. Recently, a method of removing the redundancy of the existing Neural Network has been studied, and reports have shown that the recognition rate can be close to the existing method while showing a circuit size and power consumption that is tens of times smaller than the existing method. For this reason, Edge Computing, which implements Deep Learning function in the device itself without access to the cloud or server, is evolving to the commercialization stage. Binary Neural Network (BNN) is a type that has a binary value of +1/-1 or 1/0 in the form of weight or activation, and is being developed as one of the powerful technologies that can minimize power and circuits. In general, BNN (Binary Neural Network) mainly refers to a Neural Network that has a weight and activation value of +1/-1 or 1/0 binary value in the Hidden Layer except the Input Layer and Output Layer. In order to maintain prediction accuracy even in the BNN that binarizes both the weight and activation values, it is common to use the input value of the input layer as the first layer and the output value of the output layer as the last layer without binarization.

Technologies such as BNN are quite useful in themselves, but mainly improved only in the digital domain. However, Sensing of input data is mainly done in analog form. In order to optimize the power and circuit scale from the overall device point of view, it is necessary to effectively connect the part that converts the analog signal of the sensor to digital and BNN.

Since smart phones, the use of various wearable mobile devices such as smart watches and wireless headsets has increased. Recently, various mobile devices have been used in the health-care field, such as continuous blood glucose meter and artificial joint monitoring. In the use of these devices, short battery life of up to 3 to 5 days is limiting wear. Wireless Power Transfer technology is being developed as an alternative. The existing wireless charging device 700 has been commercialized mainly based on an induction method, and in the case of the induction method, the coupling coefficient of the coil of the device Tx for transmitting wireless power and the coil of the device Rx for receiving wireless power It is operated under conditions close to 1. In order for the coupling coefficient to be close to 1, the size and shape of the transmitting and receiving coils must be similar, and alignment is important. Depending on the degree of alignment, fluctuations in efficiency can occur significantly. Therefore, it is essentially difficult to respond to various mobile devices with one wireless charging device 700. Efforts have been made to develop and use various technologies such as resonance, but full-scale commercialization is insufficient.

Korean Patent Publication No. 10-2016-0022035 (2016.02.29) receives a voltage range information from a mobile terminal and generates a voltage pattern information in response to the voltage range information. It may include a power conversion unit 710 for generating the authentication power to have, and a resonator for transmitting the authentication power to the mobile terminal.

Korean Patent Publication No. 10-2011-0068007 (2011.06.22) generates a high frequency from a commercial power supply, amplifies and transmits the amplified high frequency charging matrix, and receives the high frequency derived from the charging matrix to receive the high frequency received. It includes an adapter module composed of a power supply for rectifying DC power and supplying DC power rectified through the charging terminal of the portable terminal to the battery of the portable terminal. The invention as described above has an effect of separating the charger during charging and reducing the inconvenience of making a call by configuring the portable terminal to charge wirelessly.

Korean Patent Publication No. 10-2016-0022035 (2016.02.29) Korean Patent Publication No. 10-2011-0068007 (2011.06.22)

One embodiment of the present invention is to provide an artificial intelligence-based flexible wireless charging device that can be applied to a plurality of user terminals of various sizes and shapes by having an AI-based sensor.

An embodiment of the present invention is an artificial intelligence-based flexible wireless charging device applying a technology that organically integrates the multiplication-accumulation function of an ADC and an input layer of a Deep Learning Network (DLN) that converts an analog signal of a sensor into a digital signal. Want to provide.

An embodiment of the present invention is implemented in a thin form by separating the wireless power transmission module and the power conversion module, and the wireless power transmission module can be flexibly implemented to be optimal even when a user terminal receiving wireless power is not in a flat form. It is intended to provide an artificial intelligence-based flexible wireless charging device that can obtain charging efficiency.

Among the embodiments, the artificial intelligence-based flexible wireless charging device transmits wireless power to the wireless power receiving module and has a bendability based on characteristics including pressure, temperature, or image of the wireless power receiving module. 720 and a wireless power transmitter 720 and a non-integrated power converter 710 connected through a wired cable.

The wireless power transmitter 720 includes a coil module 722 for transmitting wireless power to the wireless power receiving module, a sensor module 724 connected to the coil module 722 and detecting characteristics of the wireless power receiving module, and It may be integrally configured to include a signal conversion module 726 connected to the sensor module 724.

The wireless power transmitter 720 may configure the coil module 722 as a plurality of wireless power transmitter coils formed in a specific arrangement regardless of the characteristics of the wireless power receiver module.

The wireless power transmitter 720 may be configured with a plurality of sensors that form the sensor module 724 as a specific arrangement of the coil module 722 and detect characteristics of the wireless power receiver module, respectively.

The signal conversion module 726 may include at least one ADC and an input layer corresponding to the first layer of a binary neural network (BNN) organically integrated with the at least one ADC.

The signal conversion module 726 may control each coil current of the plurality of wireless power transmission coils based on the output result of the BNN to simultaneously charge the plurality of wireless power reception modules.

The signal conversion module 726 includes a plurality of successive-approximation register (SAR) ADCs, and the switch capacitor bank of each ADC is maintained as it is by processing the output values of the plurality of SAR ADCs to the input layer of the BNN. And multiply-accumulate the input layer of the BNN.

The signal conversion module 726 separates the first layer from the first layer of the BNN and the remaining stages after the second layer, and integrates the first layer of the BNN with the sensor module 724 and the ADC module to output the multi-bit form of the ADC. After passing the data through the input layer of the BNN, it can be converted into binary data and transmitted to the second stage.

The disclosed technology can have the following effects. However, since a specific embodiment does not mean that all of the following effects should be included or only the following effects are included, the scope of rights of the disclosed technology should not be understood as being limited thereby.

The artificial intelligence-based flexible wireless charging device according to an embodiment of the present invention is equipped with an artificial intelligence-based sensor and can be applied to a plurality of user terminals of various sizes and shapes.

The artificial intelligence-based flexible wireless charging device according to an embodiment of the present invention integrates the first layer of the sensor module, the ADC, and the BNN into one wireless power transmitter to input the analog sensor module and the BNN with minimal power and circuit complexity. Layers can be incorporated.

The artificial intelligence-based flexible wireless charging device according to an embodiment of the present invention can organically integrate the multiply-accumulate function of the input layer of the ADC and Deep Learning Network (DLN) that converts the analog signal of the sensor into a digital signal. have.

The artificial intelligence-based flexible wireless charging device according to an embodiment of the present invention is implemented in a thin form by separating the wireless power transmission module and the power conversion module, and receiving the wireless power by flexibly implementing the wireless power transmission module. Optimal charging efficiency can be obtained even when the user terminal does not have a flat shape.

1 shows a conceptual diagram of a general Deep Learning Network (DLN).
FIG. 2(a) shows the structure of a switched-capacitor DAC of a single-ended ADC.
Shows. 2(b) is a conceptual diagram illustrating that the connection between VREF and GND in (a) is made through the SAR Register.
FIG. 3 is a conceptual diagram illustrating the concept of fusion of the switched layer capacitor DAC of the SAR ADC and the input layer operation of the DNL of FIG. 1.
4(a) shows the structure of a switched capacitor DAC of a differential structure SAR ADC. FIG. 4(b) is a conceptual diagram representing VREF and GND in FIG. (a) as SAR Register Control.
FIG. 5(a) is a conceptual diagram in which the concept of a switched capacitor capacitor DAC and a switched-capacitor neuron for wide vector suspension is fused for a differential structure. 5(b) is a conceptual diagram of implementation of analog wide summing including a switched capacitor structure for bias and offset and common mode adjustment.
FIG. 6(a) is a conceptual diagram of a sensor signal processing system when the ADC and the BNN are separately operated according to the conventional method, and FIG. 6(b) is an integrated BNN of the ADC and the BNN according to an embodiment of the present invention. It is a conceptual diagram of a wireless power transmitter that is separated from the rest of the layers and processes sensor signals.
7 is a configuration diagram of an artificial intelligence-based flexible wireless charging device according to an embodiment of the present invention.
FIG. 8(a) shows the structure of a dual resonance coil of the Qi standard, and FIG. 8(b) detects the Qi standard for detecting the position of the wireless power receiving coil when the wireless power transmitting coil is movable. This is an example of a coil.
9 is a coil module integrated with a sensor module for detecting characteristics of a wireless power receiving module according to an embodiment of the present invention, a flexible coil array having a wireless power transmitting coil array for power transmission (Flexible Coil Array) It is an example.
10 is a block diagram of an artificial intelligence-based wireless charging system capable of multi-charging according to an embodiment of the present invention.
11 is a flowchart illustrating a charging process of an artificial intelligence-based flexible wireless charging device 700 according to an embodiment of the present invention.

Since the description of the present invention is merely an example for structural or functional description, the scope of the present invention should not be interpreted as being limited by the examples described in the text. That is, since the embodiments can be variously changed and have various forms, it should be understood that the scope of the present invention includes equivalents capable of realizing technical ideas. In addition, the purpose or effect presented in the present invention does not mean that a specific embodiment should include all of them or only such an effect, and the scope of the present invention should not be understood as being limited thereby.

Meanwhile, the meaning of terms described in the present application should be understood as follows.

Terms such as "first" and "second" are for distinguishing one component from other components, and the scope of rights should not be limited by these terms. For example, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component.

When a component is said to be "connected" to another component, it may be understood that other components may exist in the middle, although they may be directly connected to the other component. On the other hand, when a component is said to be "directly connected" to another component, it should be understood that no other component exists in the middle. On the other hand, other expressions describing the relationship between the components, that is, "between" and "immediately between" or "neighboring to" and "directly neighboring to" should be interpreted similarly.

Singular expressions are to be understood as including plural expressions unless the context clearly indicates otherwise, and terms such as “comprises” or “have” are used features, numbers, steps, actions, components, parts or the like. It is to be understood that a combination is intended to indicate the existence, and does not preclude the existence or addition possibility of one or more other features or numbers, steps, operations, components, parts or combinations thereof.

In each step, the identification code (for example, a, b, c, etc.) is used for convenience of explanation. The identification code does not describe the order of each step, and each step clearly identifies a specific order in context. Unless stated, it may occur in a different order than specified. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

All terms used herein have the same meaning as generally understood by a person skilled in the art to which the present invention pertains, unless otherwise defined. The terms defined in the commonly used dictionary should be interpreted as being consistent with the meanings in the context of the related art, and cannot be interpreted as having ideal or excessively formal meanings unless explicitly defined in the present application.

The artificial intelligence-based flexible wireless charging device 700 (hereinafter, the wireless charging device 700) includes an analog to digital converter (ADC) and a deep learning network (DLN) that converts an analog signal of a sensor into a digital form. ), which integrates the multiply and accumulate part of the first layer, the input layer, and comprehensively optimizes the sensing and deep learning parts to deliver and process analog signals to the binary neural network (BNN) with minimal energy and circuit scale. Make it possible. At this time, the wireless charging device 700 separates the first layer and the rest of the BNN, but the first layer can be integrated with the sensor and ADC parts. In this case, the output data in the form of multi-bits of the ADC may be directly input to the input layer, and the weight of the input layer may be in the form of binary. The output data of the multi-bit form of the ADC is converted to binary data after being processed by the input layer of the BNN, and then transmitted to the next stage. As a result, the wireless charging device 700 can perform digital conversion of the analog signal, binary conversion, and processing of the input layer of the BNN in a minimum energy and circuit form.

In addition, the wireless charging device 700 may optimize the performance of the rest of the BNN after the second layer in terms of pure digital performance. In one embodiment, the wireless charging device 700 is not affected by the level switching problem and power disconnection problem between the analog circuit portion and the digital portion or the digital circuit and interface portion of the ADC.

In general, in the case of deep learning networks, the complexity of the first layer is relatively low compared to the rest of the layers. Therefore, when the first part of the ADC and the BNN are integrated and the other parts are separated, the data traffic between them is digital, and the amount of data is relatively small. In one embodiment, the wireless charging device 700 may apply a magnetic sensing coil, an image sensor, a temperature sensor or a pressure sensor array corresponding to a wireless power transmission/reception coil. When a function for recognizing such sensor data is added, safety and efficiency of the wireless charging device 700 can be improved, and a device for charging multiple wireless power receiving modules with one wireless charging device 700 can be implemented. have. Here, the wireless power receiving module may correspond to a charging unit of a user terminal including a mobile phone, a tablet PC, and the like. The wireless charging device 700 may integrate the sensor array, the ADC portion, and the first layer of the BNN into one sensor and signal conversion module 726 as described above for these sensors.

The wireless charging device 700 may include a single wireless power transmitter 720 that separates the wireless power transmitter coil and the power converter and combines the wireless power transmitter coil and the aforementioned sensor and signal conversion module 726 integrally. . In one embodiment, the wireless power transmitter 720 may have flexibility. More specifically, the wireless charging device 700 is provided with a wireless power transmitting coil and a wireless power receiving module of the wireless power transmitting unit 720 even when the wireless power receiving module is not in a flat shape through the wireless power transmitting unit 720 having flexibility. It is possible to maintain optimal alignment between the wireless power receiving coils. When the wireless power transmitter 720 is implemented flexibly, it is light and flexible, so it is easy to drive. In one embodiment, the processing capability of the sensing and sensor data of the wireless charging device 700 may be used to automatically perform a charging operation by grasping the type and location of the wireless power receiving module. If the type of the wireless power receiving module (Rx) is not various, it may be in the form of charging several wireless power receiving modules sequentially with one wireless power transmitting coil (Tx-Coil). Ultimately, the way to charge various Rxes at the same time is the resonance method. To maximize the advantages of the method, the Tx-Coil itself may also be implemented in an array form. When the wireless charging device 700 applies the Rx recognition result using the sensor array and the BNN integrated therein to each coil current control of the Tx-coil array in real time, multiple wireless power receiving modules (user terminals) are used. At the same time, it can be charged under optimal conditions. As a result, accurate alignment and one-to-many charging problems between various types of user terminals and the wireless charging device 700 can be solved.

First, a method of integrating the analog sensor array of the artificial intelligence-based flexible wireless charging device 700 and the input layer corresponding to the first layer of the BNN will be described in FIGS. 1 to 6.

1 is a conceptual diagram of a general deep learning network (DLN).

In FIG. 1, the first layer is referred to as an input layer. The output a _li of the i-th neuron of the first layer is the weight matrix of the first layer and the product of the output of the l-1th layer, which is the previous layer, adds Bias, b _l,i and activates the nonlinearity. It is obtained by applying the function f _act . If this is expressed as an equation, it is in the form of Equation 1 below.

[Equation 1]

Here, when X _j is a j-th input, the i-th output a _i of the input layer is expressed in the form of Equation 2 below.

[Equation 2]

The operation of the weight matrix and the output values of the previous layer can be referred to as a multiply-and-add operation. Since this operation is done in digital form, the input signal needs to be represented in digital form. However, in general, since the input signal is obtained in analog form, it must be converted to digital through ADC. In addition, even in the case of a binary neural network (BNN), in order to maintain its accuracy, the input value of the input layer is not changed to a binary form, but maintains the original ADC resolution. When the output of the input layer is output in binary form, the calculation of the rest of the layers other than the input layer can be performed using XOR Gate. Here, BNN (Binary Neural Network) mainly means a Neural Network in which the weight and activation values in the Hidden Layer have a binary value of +1/-1 or 1/0.

2(a) is a diagram showing the structure of a switched-to-capacitor DAC (Digital to Analog converter) of a single-ended SAR ADC, and FIG. 2(b) shows the connection between VREF and GND in FIG. 2(a) through the SAR Register. It is a conceptual diagram expressing what is done.

The wireless charging device 700 organically integrates the ADC that converts the analog signal of the sensor into a digital form, and the multiply and accumulate part of the first layer of the deep learning network to BNN the analog signals with minimum energy and circuit scale. To be delivered and processed. Due to its advantages such as low power, the most popular SAR (Successive Approximation Register) ADC is a switched capacitor structured DAC. In one embodiment, the wireless charging device 700 integrates the multiplication-add operation for the input layer of Equation 2 by using the switched structure of the SAR-ADC as it is when the weight matrix is in the form of a binary form. It can be done.

Figure 3 shows a conceptual diagram of the fused concept of Switched Capacitor DAC and Switched-Capacitor Neuron for Wide Vector Summation of SAR ADC. Here, when the operation shown in Equation 1 or Equation 2 is implemented as a Switched-Capacitor, the term Switched-Capacitor Neuron for Wide Vector Summation may be used.

In Figure 3, the wireless charging device 700 shows the structure of the SAR-ADC to achieve the above object. The output value of each SAR Register in binary form is multiplied with the weight value through XOR Gate. In addition, VTOP, a common node of the capacitor DAC, can be connected to Node-Sum, which is a common node of Multiply-ADD operation, through a separate switch.

Figure 4 shows the structure of the SAR-ADC differential structure. 4(a) shows the structure of a switched capacitor DAC of a differential structure SAR ADC. FIG. 4(b) is a conceptual diagram representing VREF (reference voltage) and GND (ground) in FIG.

5 shows an embodiment for a case having a differential structure. FIG. 5(a) is a conceptual diagram in which the switched capacitor of the SAR ADC is applied to the switched-capacitor wide suspension structure for the differential structure. In FIG. 5, the capacitor bank for weight portion is implemented using the capacitors used in the switched capacitor DAC of each SAR ADC of FIG. 4 as it is. As in Signal-Ended case, XOR Gate is added to process multiplication operation between SAR Register output value and binary weight value, and a separate switch is added to apply to the Switched-Capacitor Wide Summation structure. FIG. 5(b) is a conceptual diagram of analog wide summation implementation including a parallel connection of the structure of FIG. 4 and a switched capacitor structure for adjusting bias and offset and common mode. The structure of FIG. 4 is applied to the capacitor bank for weight portion of FIG. 5(b) in parallel.

FIG. 6(a) is a conceptual diagram of a sensor signal processing system when the ADC and the BNN are separately operated according to the conventional method, and FIG. 6(b) is an integrated BNN of the ADC and the BNN according to an embodiment of the present invention. It is a conceptual diagram of a wireless power transmitter 720 that is separated from the rest of the layers and processes sensor signals. The wireless power transmitter 720 includes a coil module 722 for transmitting wireless power to the wireless power receiving module, a sensor module 724 connected to the coil module 722 and detecting characteristics of the wireless power receiving module, and A signal conversion module 726 that is connected to the sensor module 724 to convert an analog signal of a sensor into a digital signal may be integrated.

The sensor signal processing system when the ADC and the BNN operate separately according to the conventional method has a structure as shown in FIG. 6(a). In FIG. 6, the wireless charging device 700 of the present invention separates the first layer and the rest of the BNN, but can integrate the first layer with the sensor module 724 and at least one ADC connected to the sensor module 724. . In one embodiment, the output data of the multi-bit form of the ADC of the wireless charging device 700 is converted to binary data and then transmitted to the next stage. As a result, the wireless charging device 700 can perform digital conversion of the analog signal, binary conversion, and processing of the first layer of the BNN in a minimum energy and circuit form.

In addition, the rest of the second layer of BNN can be optimized for pure digital performance. The wireless charging device 700 is not affected by the level conversion problem and the power disconnection problem between the analog circuit portion and the digital portion or the digital circuit and interface portion of the ADC. In general, in the case of deep learning networks, the complexity of the first layer is relatively low compared to the rest of the layers. Therefore, when the ADC and BNN first layers are combined and the rest are separated, the data traffic between them is digital, and the amount of data is relatively small.

The artificial intelligence-based flexible wireless charging device 700 to which the method of integrating the input layer of the BNN and the sensor module 724 composed of a plurality of sensor arrays described above is applied. It will be described with reference to FIG. 11.

7 is a configuration diagram of an artificial intelligence-based flexible wireless charging device 700 according to an embodiment of the present invention.

Referring to FIG. 7, the artificial intelligence-based flexible wireless charging device 700 (hereinafter, the wireless charging device 700) transmits wireless power to the wireless power receiving module and the pressure, temperature, or image of the wireless power receiving module A wireless power transmitter 720 having a bendability based on a characteristic including a wireless power transmitter 720 and a non-integral power converter 710 connected to the wireless power transmitter 720 through a wired cable. That is, the wireless charging device 700 is formed by separating the wireless power transmission coil (Tx Coil) and the power conversion unit 710 (Tx Main Module) having the functions of DC-AC power conversion and the Tx main body. do.

The wireless power transmitter 720 includes a coil module 722 for transmitting wireless power to the wireless power receiving module, a sensor module 724 for detecting characteristics of the wireless power receiving module, and a signal connected to the sensor module 724 The conversion module 726 may be integrated. In one embodiment, the wireless power transmitter 720 may configure the coil module 722 into a plurality of wireless power transmitter coils (tx coils) formed in a specific arrangement that is not limited to the characteristics of the wireless power receiver module. In addition, the wireless power transmitter 720 includes a plurality of sensors that respectively detect the characteristics of the wireless power receiving module in the form of a sensor array formed as a specific arrangement of the coil module 722. It can be configured as. For example, the wireless power transmitter 720 may implement a plurality of sensors in an array form according to the arrangement of the plurality of wireless power transmitter coils. When a basic sensing capability or sensor data processing and analysis capability is provided to the wireless charging device 700, safety and convenience may be greatly improved. To this end, a magnetic sensing coil, an image sensor, a temperature sensor, or a pressure sensor array may be applied. Through this, the wireless charging device 700 may charge various types of wireless power receiving modules Rx having various sizes or shapes, that is, user terminals.

FIG. 8(a) shows the structure of a dual resonance coil of the Qi standard, and FIG. 8(b) detects the Qi standard for detecting the position of the wireless power receiving coil when the wireless power transmitting coil is movable. This is an example of a coil.

Here, the Qi standard may refer to an interface standard developed by the Wireless Power Consortium and wirelessly charged with electric induction within 4 cm. In the existing Qi standard, the concept of locating the wireless power receiving module using a magnetic sensing coil is defined. 8(a), the wireless power receiving coil (Rx Coil) has a dual resonance structure having two resonance frequencies. A magnetic sensing coil as shown in FIG. 8( b) may be applied to the wireless charging device 700. For sensing the wireless power receiving coil of the wireless power receiving module, the wireless power receiving coil has a second resonance frequency at 1 MHz frequency. In one embodiment, in the wireless charging device 700, the sensor module 724 corresponding to various sensor arrays, such as an image sensor, a temperature sensor, or a pressure sensor, is not described in the Qi standard document, and the location of the wireless power receiving module and Can be used for sorting.

9 is a flexible coil array having a coil module 722 integrated with a sensor module 724 for detecting characteristics of a wireless power receiving module according to an embodiment of the present invention, and a wireless power transmitting coil array for power transmission. It is an example of (Flexible Coil Array).

The signal conversion module 726 may include at least one analog to digital converter (ADC) and an input layer corresponding to the first layer of a binary neural network (BNN) organically integrated with the at least one ADC. In one embodiment, the wireless charging device 700 integrates the first layer of the sensor array, ADC, and BNN into one sensor module 724 and a signal conversion module 726 for these sensor arrays. The wireless power transmission coil itself may also be implemented in an array form. In FIG. 9, the sensor module 724 that detects the characteristics of the wireless power receiving module may be a magnetic sensing coil array, or may include a temperature or pressure sensor. More specifically, the wireless charging device 700 includes a plurality of wireless power transmission coil module 722 and the above-mentioned sensor module 724 and the signal conversion module 726 integrally coupled to the wireless power transmitter 720 ) It can be implemented to be flexible (Flexible). As a result, the wireless charging device 700 can maintain an optimal alignment between the wireless power receiving coil and the wireless power transmitting coil of the wireless charging device 700 even when the wireless power receiving module is not in a flat shape. More specifically, the wireless charging device 700 may be formed to be flexible and light in weight and can be implemented in a thin form by connecting the power conversion unit 710 and the wireless power transmission unit 720 in a non-integral form. The signal conversion module may charge each of the plurality of wireless power receiving modules simultaneously by controlling each coil current of the plurality of wireless power transmitting coils based on the output result of the BNN.

10 is a block diagram of an artificial intelligence-based flexible wireless charging device 700 according to an embodiment of the present invention. In FIG. 10, the wireless charging device 700 may include a coil module 722 (Flexible Coil Array) and a signal conversion module 726 that fuses the ADC and the first layer of the BNN. At this time, the wireless power transmission coil (Tx-Coil), the sensor array (Sensor Array), and the signal (data) conversion module 726 may be physically separated without being embedded in the power conversion unit 710 as shown in FIG. 7. Here, the power conversion unit may include a DC POWER, DC-AC Converter and a transmission control unit. In one embodiment, the power conversion unit and the wireless power transmission unit may be connected through a cable. More specifically, the wireless charging device 700 is composed of a wireless power transmission coil that is implemented flexibly and a signal conversion module 726 that fuses the first layer of the ADC and the BNN, so that the wireless power transmission coil, sensor array, and signal conversion The module 726 may be physically separated from the power conversion unit as shown in FIG. 7. As a result, the wireless charging device 700 may have a basic recognition capability to grasp the type and location of the wireless power receiving module and perform an automatic charging operation to automatically perform a charging operation without user intervention.

11 is a flowchart illustrating an artificial intelligence-based wireless charging process capable of multi-charging according to an embodiment of the present invention.

In FIG. 11, the wireless charging device may perform estimation of the wireless power reception module Rx by the BNN by operating the sensor array and the signal conversion module. Thereafter, the frequency characteristics of the wireless power receiving module may be grasped by the estimation, or an ID (Identification) signal may be detected and accurate alignment with the wireless power receiving module may be performed. The wireless charging device may repeat the above process until the correct alignment is achieved by the operation of the sensor and the signal conversion module when the above-mentioned correct alignment is not achieved. Once the correct alignment is achieved, wireless charging can be performed and terminated if there is no abnormality.

In one embodiment, when there are multiple wireless power receiving modules, the wireless charging device 700 may sequentially charge multiple wireless power receiving modules with one wireless power transmitting coil. In another embodiment, the wireless charging device 700 may ultimately implement the wireless power transmission coil itself in an array form through a resonance method capable of simultaneously charging various wireless power reception modules. That is, the wireless charging device 700 includes a sensor module 724 including the above-mentioned sensor array and a coil module 722 including a plurality of wireless power transmitting coils to recognize the result of the wireless power receiving module using the integrated BNN. By applying to each coil current control in real time, it is possible to charge a plurality of wireless power receiving modules having various types under optimal conditions. As a result, the wireless charging device 700 according to the present invention cannot accurately charge a plurality of user terminals with one wireless charging device 700 and an exact alignment problem with various wireless power receiving modules included in various user terminals. Existing problems can be solved.

Although described above with reference to preferred embodiments of the present invention, those skilled in the art variously modify and change the present invention without departing from the spirit and scope of the present invention as set forth in the claims below. You can understand that you can.

700: artificial intelligence-based flexible wireless charging device
710: power conversion unit
720: wireless power transmitter 722: coil module
724: sensor module 726: signal conversion module

Claims (8)

A wireless power transmitting unit having a flexibility to transmit wireless power to the wireless power receiving module and varying based on characteristics including pressure, temperature, or image of the wireless power receiving module; And
It includes a non-integrated power conversion unit spaced apart and connected to the wireless power transmission unit,
The wireless power transmitter
A coil module that transmits wireless power to the wireless power receiving module;
A sensor module connected to the coil module and detecting a characteristic of the wireless power receiving module; And
It is connected to the sensor module is configured to include an integrated signal conversion module for converting the analog signal of the sensor to a digital signal,
The signal conversion module
At least one ADC (Analog to Digital Converter); And
An artificial intelligence-based flexible wireless charging device comprising an input layer corresponding to the first layer of a binary neural network (BNN) organically integrated with the at least one ADC.
delete According to claim 1, The wireless power transmitter
An artificial intelligence-based flexible wireless charging device, characterized in that the coil module is composed of a plurality of wireless power transmission coils formed in a specific arrangement regardless of the characteristics of the wireless power receiving module.
According to claim 3, The wireless power transmitter
The artificial intelligence-based flexible radio, characterized in that the sensor module is composed of a plurality of sensors each detecting the characteristics of the wireless power receiving module in the form of a sensor array formed as a specific arrangement of the coil module. Charging device.
delete According to claim 1, wherein the signal conversion module
Based on the output result of the BNN, the artificial intelligence-based flexible wireless charging device, characterized in that charging the plurality of wireless power receiving modules at the same time by controlling each coil current of the plurality of wireless power transmitting coils.
According to claim 1, wherein the signal conversion module
A plurality of successive-approximation register (SAR) ADCs are included, and the output values of the plurality of SAR ADCs are transferred to the input layer of the BNN for processing, and the switched capacitor bank of each ADC is used as it is. An artificial intelligence-based flexible wireless charging device characterized by performing a multiply-accumulate function of the input layer of the BNN.
The method of claim 6, wherein the signal conversion module
After separating the first layer of the BNN and the remaining stages after the second layer, and integrating the first layer of the BNN with the sensor module and the ADC module, the binary data after passing the multi-bit output data of the ADC through the input layer of the BNN Converted to the second stage, characterized in that the AI-based flexible wireless charging device.

Images