KR102244156B1 - Method for processing voice commands using bluetooth mesh network - Google Patents

Abstract

The present invention relates to a voice command processing method using a Bluetooth mesh network, wherein when voice data is input to at least one or more nodes of a Bluetooth mesh network, a control unit of one specific node specified in a predetermined manner among the nodes Dividing the voice data into a plurality of voice data according to a predetermined method; Dividing the divided plurality of voice data into voice data to be distributedly processed for each node of the Bluetooth mesh network, by a controller of the specific node; Allowing the control unit of the specific node to transmit the classified voice data for each node to perform distributed processing, and to reconfigure the distributed voice data for each node by concentrating on one specified central node in a predetermined manner; And recognizing a voice command corresponding to the reconstructed voice data by the controller of the central node and executing the command.

Description

Voice command processing method using Bluetooth mesh network {METHOD FOR PROCESSING VOICE COMMANDS USING BLUETOOTH MESH NETWORK}

The present invention relates to a voice command processing method using a Bluetooth mesh network, and more particularly, in a Bluetooth mesh network in which a plurality of voice command processing devices are connected as nodes, respectively, additionally use resources of external devices connected to the node. The present invention relates to a method for processing a voice command using a Bluetooth mesh network, which enables the voice command input to the node to be processed.

Conventionally, in homes and offices, many remote control devices (eg, remote controllers) for controlling devices (eg home appliances, lights) using infrared or wireless signals have been released, but recently, target devices using voice commands There are a lot of voice recognition control devices (eg AI speakers) to control (eg LED lights, fluorescent lights, boilers, air purifiers, etc.).

At this time, if the control device does not have a voice recognition function, a low-spec CPU can be used because a complex operation is not required. However, in the case of adding a voice recognition function to the control device, a high-spec CPU ( CPU).

Accordingly, there is a problem that the manufacturing cost of the voice recognition control device increases, and as the number of voice commands to be processed increases or the length of the voice commands increases, it is necessary to use a higher specification CPU to shorten the processing time. Therefore, there is a problem that the manufacturing cost is further increased. This is acting as an obstacle to the popularization of the voice recognition control device.

Accordingly, there is a need for a method to further improve the processing performance of voice commands while lowering the manufacturing cost of the voice recognition control device.

The background technology of the present invention is disclosed in Korean Patent Registration No. 10-0984528 (registered on September 24, 2010, a system and method for speech recognition in a distributed speech recognition system).

According to an aspect of the present invention, the present invention is created to solve the above problems, in a Bluetooth mesh network in which a plurality of voice command processing devices are connected as nodes, respectively, the resources of external devices connected to the nodes are Another object of the present invention is to provide a method for processing a voice command using a Bluetooth mesh network, which allows the voice command input to the node to be processed by additional use.

In the method of processing a voice command using a Bluetooth mesh network according to an aspect of the present invention, when voice data is input to at least one or more nodes of the Bluetooth mesh network, the Dividing the voice data into a plurality of voice data according to a predetermined method; Dividing the divided plurality of voice data into voice data to be distributedly processed for each node of the Bluetooth mesh network, by a controller of the specific node; Allowing the control unit of the specific node to transmit the classified voice data for each node to perform distributed processing, and to reconfigure the distributed voice data for each node by concentrating on one specified central node in a predetermined manner; And recognizing, by the controller of the central node, a voice command corresponding to the reconstructed voice data and executing the command.

In the present invention, the specific node compares the amplitude peak value of the voice data with the voice input time by at least one or more nodes simultaneously receiving voice data, and according to the comparison result, the voice input time is faster or the amplitude A node having a larger peak value is specified or characterized in that it is specified in a random manner.

In the present invention, in the step of concentrating and reconstructing the voice data distributed by each node to one specified centralized node in a predetermined manner, the centralized node includes a performance table generated in advance by a control unit of a specific node. Or, by transmitting a'performance question request message' to other nodes in real time, and receiving a'performance reply message' from other nodes, based on the information included in the'performance reply message', It is characterized by finding a node and selecting it as a centralized node.

In the present invention, the centralized node estimates the user's current location based on physical group information of other nodes that have transmitted distributed-processed voice data, and when the user inputs a voice command as the current location of the user is estimated. Even if the physical group information is omitted, the physical group corresponding to the estimated current position of the user is automatically recognized to simplify the voice command.

In the present invention, the nodes of the Bluetooth mesh network are implemented by integrating a voice command processing device and a control target device into one.

In the present invention, in the step of dividing the voice data into a plurality of voice data according to a predetermined method, performing a scaling process of expanding or reducing the input voice data according to a preset reference value; It is characterized by that.

In the present invention, when the voice data is divided in a predetermined manner, the control unit of the specific node checks whether the number of divided voice data is less than the preset number, and as a result of the check, the number of divided voice data When the number is less than a predetermined number, the voice command processing apparatus of the specific node does not distribute the voice data to a plurality of nodes, and the voice command processing apparatus of the specific node is implemented to perform voice analysis and processing on its own.

In the present invention, when the voice data is concentrated, the control unit of the central node checks whether command essential information corresponding to the number of divided voice data is included, and as a result of the check, corresponds to the number of voice data. If all the required command information is included, the corresponding voice command is executed, and if the command essential information corresponding to the number of voice data is not included, the voice command containing insufficient command essential information is re-entered through the information output unit. It characterized in that it is implemented to request.

In the present invention, in the step of dividing by the control unit of the specific node into voice data to be distributed and processed for each node of the Bluetooth mesh network, the control unit of the specific node includes: According to the performance of the connected external device, whether or not the voice data is distributed, the distribution priority of the voice data, and the number of voice data to be distributed are adjusted.

In the present invention, in order for the control unit of the specific node to classify the voice data to be distributed and processed for each node of the Bluetooth mesh network, the control unit of the specific node includes: the number of nodes, the performance of each node or an external device connected to the node. It is characterized by referring to a performance table in which information is stored.

In the present invention, in the performance table, the controller of each node of the Bluetooth mesh network transmits a'performance question request message' to another node or an external device connected to the node, and the other node or an external device connected to the node is Generated based on information included in the returned'performance reply message', and the performance reply message includes at least one of CPU clock information, CPU core count information, current CPU usage rate information, and current spare memory information. It is characterized.

In the present invention, the controller of each node of the Bluetooth mesh network, in order to select a node that can process voice data fastest or an external device connected to the node, not only the information of the'performance reply message', but also leaps forward. It is characterized in that the information and distance information are additionally reflected.

In the present invention, the nodes of the Bluetooth mesh network are classified based on a physical group according to a location of a target device to be controlled by the node, a type of the target device, and a number of the target device.

In the present invention, the physical group may be configured by a user using an external device, or a node itself may be configured to automatically set based on distance information by communicating with another node of the Bluetooth mesh network.

In the present invention, in order to automatically set the physical group based on distance information by communicating with another node of the Bluetooth mesh network by itself, a new node in a state in which nodes already registered with a physical group are installed. When is newly installed, the new node requests registration of a physical group to another node, and each physical group node estimates the distance between itself and the new node and transmits this distance information to the central node. Compares the distances between other nodes in the physical group to which it belongs and nodes in the other physical group, searches for which physical group the new node is closest to, and determines which physical group is the closest to the central node according to the search result. It characterized in that it is implemented to determine the closest physical group information to.

According to an aspect of the present invention, in a Bluetooth mesh network in which a plurality of voice command processing devices are connected as nodes, respectively, the voice command input to the node is transmitted by additionally using resources of external devices connected to the node. By making it possible to process, it is possible to improve the functionality and economics of the node.

1 is an exemplary view showing a schematic configuration of a voice command processing apparatus according to an embodiment of the present invention.
2 is a flowchart illustrating a voice command processing method using a Bluetooth mesh network according to an embodiment of the present invention.
3 is an exemplary view showing a schematic form of a Bluetooth mesh network according to an embodiment of the present invention.
FIG. 4 is an exemplary diagram illustrating a method of generating and updating a performance table for storing performance information for each node in FIG. 2.
FIG. 5 is an exemplary diagram illustrating a method of dividing a plurality of voice data from one input voice data in FIG. 2.
FIG. 6 is an exemplary diagram illustrating a process of distributing voice data in FIG. 2, dividing voice data according to the number of nodes to be distributed, and then concentrating the distributed voice data again in order to distribute the voice data. .
FIG. 7 is an exemplary diagram illustrating scaling performed when segmenting voice data for distributed processing in FIG. 5.
FIG. 8 is an exemplary diagram illustrating a method of excluding a designated specific node among nodes of a Bluetooth mesh network from distributed processing in FIG. 2.
9 is an exemplary view shown in FIG. 2 to explain the types and number of information that can be included in a voice command according to a user's current location.
10 is an exemplary view illustrating a method of grouping target devices to be controlled and matching voice commands in FIG. 2.

Hereinafter, an embodiment of a method for processing a voice command using a Bluetooth mesh network according to the present invention will be described with reference to the accompanying drawings.

In this process, the thickness of the lines or the size of components shown in the drawings may be exaggerated for clarity and convenience of description. In addition, terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of users or operators. Therefore, definitions of these terms should be made based on the contents throughout the present specification.

1 is an exemplary view showing a schematic configuration of a voice command processing apparatus according to an embodiment of the present invention.

1, the voice command processing apparatus 100 according to the present embodiment includes a voice input unit 110, a communication unit 120, a control unit 130, a storage unit 140, and an information output unit 150. , And a target device driver 160.

In the voice command processing apparatus 100 according to the present exemplary embodiment, a plurality of voice command processing apparatuses 100 may be connected as nodes to form a Bluetooth mesh network.

If the voice command processing device 100 is integrated into the target device (eg, LED lamp, fluorescent lamp, boiler, air purifier, etc.) in this embodiment, the target device itself can be connected to the Bluetooth mesh network as a node. Please note.

The voice input unit 110 includes a microphone (MIC), and may receive a sound from outside, for example, a user's voice.

The communication unit 120 connects the communication (or network) between the voice command processing devices according to the present embodiment, or communicates between the voice command processing device according to the present embodiment and an external device (for example, a smartphone, tablet, PC, etc.). Can be connected.

Here, the external device is not the voice command processing device, but includes a processor (eg, CPU) to install an application program, and also processes various data or commands transmitted from the voice command processing device or performs calculations. It means a device that can. However, external devices are not necessarily connected to each node, and external devices may not be connected to all nodes at all.

For example, the communication unit 120 may connect the communication through wireless communication (eg, Bluetooth, WiFi, NFC, etc.) or wired communication.

Hereinafter, the voice command processing device 100 (or a target device in which the voice command processing device is integrated) to which the communication (or network) is connected may be described as a node.

The communication unit 120 divides all of the voice (or voice command) input through the voice input unit 110 or a part of the voice (or voice command), and the voice command processing device (or node) to which the communication is connected, or an external device. It can be transferred to devices (e.g., smartphones, tablets, PCs, etc.). In addition, the communication unit 120 may receive a voice (or voice command) recognition result from a voice command processing device (or node) to which the communication is connected or an external device (eg, a smartphone, tablet, PC, etc.).

The controller 130 controls at least one or more other components of the voice command processing apparatus 100 and performs operations or data processing related to communication.

For example, the controller 130 may control a plurality of hardware or software components connected to the controller 130 by driving an operating system or an application program, and may perform various data processing or operations. In addition, the control unit 130 may include one or more of a central processing unit, an application processor (AP), or a communication processor (CP), and may be implemented in the form of a system on a chip (SoC).

The control unit 130 processes a voice command by recognizing a voice input through the voice input unit 110. Here, since there are various known techniques for recognizing speech, a description of the speech recognizing technology itself is omitted in this embodiment.

The storage unit 140 stores at least one voice command to be recognized and processed by the voice command processing apparatus 100. In addition, in response to the voice command, data related to the device to be controlled (or driven) (ie, the target device) (eg, location, ID, signal output to the target device for control) is stored.

The storage unit 140 may include volatile and/or nonvolatile memory.

The storage unit 140 may include an internal memory or an external memory.

The storage unit 140 may store software and/or programs (eg, applications). Here, the application is a set of programs (or instructions (instructions)) for performing at least one specified function, for example, in a Bluetooth mesh network in which a plurality of voice command processing devices are connected as nodes, respectively, each connected to the node. It may include a voice command distributed processing application, etc., which enables the voice command input to the node to be processed (distributed processing) by additionally using resources of external devices.

The information output unit 150 receives a voice (or voice command) through the voice input unit 110, recognizes the received voice (or voice command), or recognizes the voice (or voice command). In the case of processing, information according to the above steps (eg, information on receiving a voice, information on a voice recognition state, information on a result of driving a target device by processing a voice command) may be output in a sound or display method. Through this, the user can check whether the voice command is being properly processed by recognizing the operation state of the voice command processing apparatus 100.

The target device driving unit 160, according to the control of the control unit 130, to control the target device (eg, LED lamp, fluorescent lamp, boiler, air cleaner, etc.) to be controlled by the voice (or voice command) The target device is controlled by outputting a predetermined control signal (eg, on/off switching signal, dimming signal, level control signal, etc.).

Hereinafter, a method of processing a voice command by the voice command processing apparatus 100 using a Bluetooth mesh network will be described.

For example, the Bluetooth mesh network can exchange data with an external device using Bluetooth Low Energy (LE). The external devices are always connected to the Bluetooth mesh network, and data can be exchanged at any time. Accordingly, the calculation to be performed in each node (ie, the voice command processing device or the target device in which the voice command processing device is integrated) can be transferred to the external device for distributed processing. .

Hereinafter, it is assumed that an application program (application) capable of distributed processing operations for speech recognition by communicating with the node (i.e., a voice command processing device or a target device in which a voice command processing device is integrated) is installed in the external device. Explain.

2 is a flowchart illustrating a method of processing a voice command using a Bluetooth mesh network according to an embodiment of the present invention, and FIG. 3 is an exemplary view showing a schematic form of a Bluetooth mesh network according to an embodiment of the present invention. .

For convenience of description, the following description assumes a situation in which a voice (ie, a voice command) is input to any one node (eg, A) among a plurality of nodes connected to the Bluetooth mesh network shown in FIG. 3.

For reference, in this embodiment, the Bluetooth mesh network is described as an example, and in other embodiments, it may be just a wireless mesh network.

Referring to FIG. 2, when voice data (ie, voice command) is input (S101), the control unit 130 of the voice command processing apparatus 100 according to the present embodiment sets the voice data in a predetermined manner (eg: It is divided into phoneme units, phoneme units, etc.) (S102) (see Fig. 5).

For example, as shown in FIG. 5, one continuous input voice data is divided into a plurality of voice data according to a predetermined method.

FIG. 5 is an exemplary diagram illustrating a method of dividing a plurality of voice data from one input voice data in FIG. 2, and FIG. 7 is a division of voice data for distributed processing in FIG. It is an exemplary diagram shown to explain the scaling performed during the time.

Referring to FIG. 7, since voice data (i.e., voice command) input by a user can be spoken relatively slowly or quickly depending on the user, the length of the voice data (i.e., voice command) Time) also makes a difference.

Accordingly, in order to accurately analyze or process voice data (i.e., voice command), the control unit 130 performs a scaling process of matching the input voice data (i.e., input voice) with a preset comparison value (i.e., a reference value). (In other words, the process of expanding or reducing the input voice according to the reference value) is performed.

In this case, a scale factor for the scaling process may be set differently each time voice data (ie, a voice command) is input (ie, according to a user who inputs voice).

Referring back to FIG. 2, when one voice data (i.e., voice command) is divided in a predetermined manner as described above, the controller 130 determines the number of voice data divided in the predetermined manner as a predetermined value ( It is checked if it is less than the number) (S103).

For example, when the number of divided voice data is less than or equal to a predetermined value (number), it means that voice analysis and processing can be processed without using many resources (ie, CPC resources). In other words, it means that even if the distributed processing is not performed, the voice data can be processed by the input node (eg, A) itself.

As a result of the check (S103), if the number of the voice data divided in the preset manner is less than or equal to a preset value (number) (YES in S103), the controller 130 ) Performs speech analysis and processing in itself (S104).

At this time, the control unit 130 checks whether essential command information corresponding to the number of divided voice data is included (S108).

For example, the command essential information includes the type of the target device to be controlled (or the ID of the target device or the number designated for the target device) and a control command (eg, on, off, on, off, etc.). And the number of voice data is the number of voice data divided in a predetermined manner according to the specific degree of voice data (i.e., voice command) (i.e., the number of information included in the voice command) (i.e., included in the voice command). It means the number of information). For example, the user may use "Turn off the fluorescent lamp", "Turn off fluorescent lamp B", and "Turn off all fluorescent lamps in the corridor" (see Fig. 9(b)), and the number of information included in the voice data (ie, voice command) is different. I can.

9 is an exemplary diagram shown to explain the type and number of information that can be included in a voice command according to the user's current location in FIG. 2, and when the user's current location is a library, the library is equipped with a stand (D ) And the fluorescent lamp (E), the user can input a voice command including two pieces of information (eg, fluorescent lamp, turn off), such as "turn off the fluorescent lamp". However, if the user's current location is a corridor, there are two fluorescent lamps (F, G) and a stand (H) in the corridor, so the user has three pieces of information (eg, fluorescent lamp, F, turn off), such as "Turn off fluorescent lamp F". You must enter the included voice command. In addition, if the user's current location is a corridor and wants to control the fluorescent lamp (C) in my room, the user has a voice that includes four pieces of information (eg, my room, fluorescent lamp, c, turn off), such as “Turn off the fluorescent lamp C in my room”. You have to enter the command.

Therefore, when the user inputs a voice command including two pieces of information (eg, audio, off), such as "turn off audio", in order to control a target device (eg, audio) with only one user, The control unit 130 determines that the input voice command includes all of the essential command information.

However, even when controlling a plurality of target devices (eg, fluorescent lamps), when the user inputs a voice command including only two pieces of information (eg, fluorescent lights, turn off), such as “turn off the fluorescent lamp”, the controller 130 ) Indicates that the voice command does not contain command essential information.

When all of the essential command information corresponding to the number of voice data is included as described above (YES in S108), the controller 130 executes the voice command (S110). That is, the target device control signal is output.

However, if the command essential information corresponding to the number of voice data is not included (No in S108), the control unit 130 voices through the information output unit 150 (that is, a voice command including insufficient command essential information). ) A re-input may be requested (S109).

On the other hand, as a result of the check (S103), if the number of divided voice data is greater than a predetermined value (number) (No in S103), the control unit 130 is Voice data to be processed through the device) is identified (S105).

For example, when the number of nodes (eg, A) itself and other nodes (or external devices) is 6, the voice data is divided into 6 (see FIGS. 5 and 6).

In this case, the number of divided voice data does not necessarily have to match the number of nodes, and the number of voice data distributed for each node does not have to be the same. That is, the control unit 130 determines whether the voice data is distributed, the distribution priority of the voice data, and the voice data to be distributed according to the performance of each node (or an external device connected to the node) (that is, the performance of the CPC resource). You can adjust the number.

For example, as shown in FIG. 8, distributed processing may be performed by excluding some nodes (eg, D) from distributed processing among nodes of a Bluetooth mesh network.

To this end, the controller 130 may refer to a performance table in which performance information for each node (or an external device connected to the node) is stored.

FIG. 4 is an exemplary diagram illustrating a method of generating and updating a performance table for storing performance information for each node in FIG. 2. As shown in FIG. 4, each node of a Bluetooth mesh network has a predetermined period. Depending on whether each node is connected to an external device and to select a node (or an external device connected to the node) that can process voice data fastest, a'performance question request message' is sent to another node (or connected to a node). External device), and another node (or an external device connected to the node) returns a'performance reply message'.

In this case, the performance reply message includes information on a CPU clock, information on the number of CPU cores, information on a current CPU usage rate, and information on a current spare memory.

Accordingly, each node reviews the received'performance reply message' and selects a node (or an external device connected to the node) that can perform the fastest voice data processing.

At this time, the information to be reviewed in order to select a node (or an external device connected to the node) that can process the voice data fastest is only the information of the'performance reply message' of each received node (or an external device connected to the node). Rather, it includes leap information (the number of nodes passing through to the final node to which the external device is connected) and distance information (physical distance to the node to which the external device is connected).

For example, even if the performance of the external device is the best, if the transmission/reception time of data is delayed due to a leap through multiple nodes, or the quality of transmitted data is deteriorated due to a long distance, this increases the overall processing time and causes malfunction. Becomes. Accordingly, in order to prevent such an increase in processing time and malfunction, the corresponding node (e.g., A) calculates a score by synthesizing'performance reply message','jumping information', and'distance information' (i.e., prioritizing Calculation), specifies the node with the highest score (i.e., the highest priority) (or an external device connected to the node), and then requests the node (or an external device connected to the node) for distributed processing of voice data.

3, in order to transmit the voice data of node A to the tablet (external device connected to node D), it goes through A → B → D. At this time, since the distance between node B and node D is 30m, transmission/reception time delay and data quality degradation Is somewhat expected.

In addition, in order to transmit the voice data of node A to a PC (an external device connected to node F), it goes through A → C → F. At this time, since the distance between node C and node F is 50 m, more quality degradation is expected than the tablet.

On the other hand, in order to transmit the voice data of node A to the smartphone (external device connected to node E), it goes through A -> E. At this time, the distance between node A and node E is 3m, so the expected time delay and quality deterioration are different from external devices. Much lower than

Accordingly, the node A may select an external device (smartphone) connected to the node E as the node (or external device) having the highest priority for distributed processing. However, note that at least one node (or external device) selected for the distributed processing may be selected according to priority.

Referring back to FIG. 2, the control unit 130 transmits the divided voice data for each node (or to a selected specific node) for distributed processing (S106).

For example, referring to FIG. 3, assuming that node A selects only node E, node A sends voice data and a data processing request message to node E, and node E transmits the message to an external device (smartphone) connected to it. It delivers the received content (voice data, data processing request message). Accordingly, the external device (smartphone) connected to the node E performs a noise removal operation and a speech processing operation of the received voice data, and generates result data (eg, speech processing data composed of text) after the processing is completed. The external device (smartphone) connected to the node E transmits the result data (eg, voice processing data composed of text) to the node A through the node E. Accordingly, Node A processes the received result data (eg, voice processing data composed of text) and executes the corresponding command.

In addition, the control unit 130 receives (or concentrates) the distributed-processed voice data for each node and synthesizes (or reconstructs) it (S107).

FIG. 6 is an exemplary view illustrating a process of distributing voice data in FIG. 2 by dividing voice data corresponding to the number of nodes to be distributed, performing distributed processing, and then concentrating the distributed-processed voice data. to be.

After concentrating and synthesizing (or reconstructing) the distributed voice data as described above, the control unit 130 checks whether the required command information corresponding to the number of divided voice data is included (S108).

And if the command essential information corresponding to the number of voice data is included (YES in S108), the controller 130 executes the voice command (S110), and if the command essential information corresponding to the number of voice data is included. When is not included (No in S108), the controller 130 may request a re-input of a voice (ie, a voice command including insufficient command essential information) through the information output unit 150 (S109).

As described above, when the voice command processing apparatus 100 according to the present embodiment can recognize and process voice without using a large number of resources (ie, CPC resources) based on the number of divided voice data, the voice command processing apparatus 100 The input node (eg, A) itself processes voice commands, and when a large amount of resources needs to be used, the voice commands can be processed through distributed processing.

However, the present invention is not limited thereto, and voices (ie, voice commands) may be processed through unconditionally distributed processing regardless of the number of divisions of voice data.

Meanwhile, FIG. 10 is an exemplary diagram illustrating a method of grouping a target device to be controlled and a matching method of a voice command in FIG. 2, and a configuration of a voice command to be matched to a target device to be controlled with reference to FIGS. 9 and 10 Explain about.

As described above, all nodes of the Bluetooth mesh network according to the present embodiment basically have a physical group of the target device to be controlled by the node (e.g., a location where the target device is installed, my room, study, corridor, etc.) and the target device. It is classified based on the type of (eg, stand, fluorescent light, etc.) and the number of the target device (eg, A to H).

For reference, the physical group can be set by the user when purchasing the target device using an external device (e.g., smartphone, tablet, PC, etc.), and the target device itself communicates with other target devices to provide a physical group based on distance information. Can also be set automatically (see Fig. 10).

In addition, when the voice command processing device 100 is integrated into a target device (eg, an LED lamp, a fluorescent lamp, a boiler, an air cleaner, etc.), it should be noted that the target device itself may become a node of the Bluetooth mesh network. In this case, the target device number (eg, A to H) and the node number (eg, A to H) are the same number.

9A and 10, a target device (or node) to be controlled by each node of a single Bluetooth mesh network is a target of the same location (physical group) and the same type (e.g., stand, fluorescent lamp, etc.) There may be multiple devices. Therefore, for individual control of the target device (or node), the user must be familiar with the form of the voice command matched with the target device (or node) as shown in (b) of FIG. 9.

For example, the form of the voice command is the call sign of the node (or target device) (eg, Blue Night), the function of the target device group (eg, light, boiler, air cleaner, audio, etc.), and the physical group of the target device (eg. : Location of my room, study, hallway, etc.), the type of target device (eg, stand, fluorescent lamp), and the number of each target device by type (eg: A ~ H), and control commands (eg: on, off, on). , Off, etc.).

For example, assuming that "Fluorescent Light B" installed in'My Room' is turned off, you can input voice (ie, voice command) in the form below.

Voice Command: "Blue Night" + "My Room" + "Fluorescent Light" + "B" + "Off"

However, it is difficult for a user to know the numbers (or values of individual nodes) of all target devices for individual control of the target devices, and a method of simplifying voice commands is needed to solve this difficulty.

Hereinafter, a method of recognizing a voice command in each node of a Bluetooth mesh network will be described.

For example, when a user inputs a voice (i.e., voice command) (e.g., turn off the fluorescent light B in my room, blue night) at an arbitrary location (e.g., my room), each node receiving the voice (i.e., voice command) at the same time (Example: A, B, C, F, G) compares the amplitude peak value of the corresponding speech data with the speech input time, and the result of this comparison (e.g., a node with a faster speech input time or a larger amplitude peak value) ) To specify any single node for speech processing. In this case, the single node may be specified in a random manner.

The specified single node (e.g., node A) divides the voice data and transmits it to other nodes for distributed processing, and also selects a central node (e.g., node G) to concentrate (or reconstruct) the distributed voice data. Select.

At this time, the single node (e.g., node A) may refer to a performance table (see FIG. 4) previously generated for the selection of the centralized node (e.g., node G), or a'performance question request message' in real time. By transmitting to other nodes and receiving a'performance reply message' from other nodes, it is possible to find a node with the best performance (eg, node G) and select it as a central node.

The centralized node (eg, node G) selected as described above concentrates the distributed voice data from each node, that is, reconstructs the voice data, and executes the command.

At this time, the centralized node (e.g., node G) estimates the user's current location (physical group) by the physical group of the other nodes (i.e., nodes that simultaneously received the user's voice) that transmitted the distributed voice data. do. For example, when the physical groups of other nodes that have transmitted the distributed-processed voice data are equal to or greater than a predetermined ratio (e.g., 80%), the centralized node (e.g., node G) assigns the physical group to the user's current location (physical group). ).

As the user's current location (physical group) is estimated as above, when the user inputs a voice (i.e., voice command), it is possible to omit the physical group (e.g., location of my room, study, corridor, etc.). It is possible to automatically recognize the physical group corresponding to the current location), thus simplifying the voice command.

For example, if the user inputs a voice (ie, voice command) saying "Blue night, turn off the fluorescent lamp" without including information on the physical group, the centralized node (eg node G) estimates the user's current location ( Example: My room), "Turn off the fluorescent lights in my room", and also recognizes all fluorescent lights (B, C) in "My room" without distinction. That is, when only the type of the target device is specified and the target device number is not entered, the centralized node (eg, node G) recognizes the command by integrating all target devices of the corresponding type.

As another example, when the user inputs a voice (i.e., voice command) saying "Blue night, turn off the lights", the centralized node (e.g., Node G) estimates the user's current location (e.g., my room) , Recognizes as "Turn off the lights in my room", and recognizes all lights (ie, lights) in "my room" without distinction. That is, when only the function of the target device is specified and the physical group, type, and number of the target device are not entered, the centralized node (eg, node G) recognizes the command by integrating all target devices of the corresponding function.

As described above, the present embodiment estimates the user's current location (physical group), and when only specific upper information is input in the form of a voice command and lower information is omitted, the target device is selected only with the specific higher information in which lower information is integrated. It has the effect of simplifying the voice command by recognizing it in an integrated manner.

In addition, as described above, in the present embodiment, a node (eg, node A) instructing distributed processing of voice data is different from a node (eg, node G) that executes commands by concentrating (or reconstructing) the distributed-processed voice data. That is, since this embodiment does not have a separate server (that is, since there is no need to change the server settings), it is possible to improve scalability, and it is active when a target device (ie, a target device acting as a node) is added or deleted. It has the effect of being able to respond to it.

Meanwhile, as already described above, the physical group of the target device can be set by the user when purchasing the target device using an external device (eg, smartphone, tablet, PC, etc.), and the target device (or node) itself It is possible to establish a physical group by communicating with the target device (or other node).

For example, in the case of a general home, as shown in FIG. 10, as the number of rooms is small, there is no difficulty in setting the physical group of the target device by the user. However, when the physical space is divided into a large number, such as an entire apartment or a large factory, the user In order to directly set the physical group of the target device, there is a problem that the complexity and difficulty of the work increase. Therefore, it is necessary for the target device to set its own physical group.

Hereinafter, when a target device is newly installed, a method of setting a physical group by communicating with another target device (or another node) by itself will be described with reference to FIG. 10.

Referring to FIG. 10, when a new target device (i.e., a new node) is newly installed while target devices to which a physical group is registered are installed, the new target device (i.e., a new node) is physically connected to another node. Request group registration. Accordingly, the node of each physical group estimates the distance between itself and the new target device (ie, the new node) and transmits this distance information to a specific node (eg, a centralized node).

Here, for the specific node (eg, centralized node, node a), a node (eg, node A) closest to the new target device (ie, new node) is selected.

For reference, in the case of a Bluetooth mesh network, it is possible to estimate the distance between each node using a Heartbeat message or an ibeacon.

Accordingly, the centralized node (e.g., node A) compares the distances between other nodes in the physical group (e.g., my room) to which it belongs and nodes belonging to another physical group (e.g., corridor), and compares the distance between the new target device (i.e. , New node) Search which physical group (N) is closest to.

For example, a new node (N) can communicate with all nodes (A, B, C) whose physical group is set to'My Room' and is close to within a certain distance, and nodes (F, G, C) whose physical group is set to'Corridor' H) is searched for that it is farther apart than a certain distance or that communication is impossible. Accordingly, the centralized node (eg, node A) determines that the new node N is the closest to the'my room' physical group and stores the corresponding content.

And the centralized node (e.g., node A) transmits physical group information (e.g., my room) of the new node (N) to the user's external device (e.g., a smartphone), so that the user Whether to register a physical group is determined, and when the user decides to register, the user's external device (eg, a smartphone) updates the physical group information (eg, my room) of the new node (N). That is, the user only needs to check the physical group of the new node N that the central node (eg, node A) has searched for and decide whether to register it as the physical group.

Accordingly, the user has an effect of increasing convenience in registering a physical group of a new target device (new node, N) and shortening a physical group registration time of the new node (N).

As described above, the present invention has been described with reference to the embodiments shown in the drawings, but this is only an example, and various modifications and other equivalent embodiments are possible from those of ordinary skill in the field to which the technology pertains. I will understand the point. Therefore, the technical protection scope of the present invention should be determined by the following claims.

100: voice command processing device 110: voice input unit
120: communication unit 130: control unit
140: storage unit 150: information output unit
160: target device driving unit

Claims (15)

When voice data is input to at least one of the nodes of the Bluetooth mesh network, the control unit of one specific node specified in a predetermined manner among the nodes divides the voice data into a plurality of voice data according to a predetermined method. The step of doing;
Dividing the divided plurality of voice data into voice data to be distributed and processed for each node of the Bluetooth mesh network, by a controller of the specific node;
Allowing the control unit of the specific node to transmit the classified voice data for each node to perform distributed processing, and to reconfigure the distributed voice data for each node by concentrating on one specified central node in a predetermined manner; And
And executing, by the control unit of the centralized node, a voice command corresponding to the reconstructed voice data.
The method of claim 1, wherein the specific node,
At least one or more nodes receiving voice data simultaneously compare the amplitude peak value of the voice data and the voice input time, and a node having a faster voice input time or a larger amplitude peak value is specified according to the comparison result,
Voice command processing method using a Bluetooth mesh network, characterized in that it is specified in a random manner.
The method of claim 1, wherein in the step of concentrating and reconstructing the voice data distributedly-processed for each node to one specified centralized node in a predetermined manner,
The centralized node,
The control unit of a specific node,
Based on information included in the'Performance Reply Message' by referring to a previously created performance table or by transmitting a'Performance Question Request Message' to other nodes in real time, and receiving a'Performance Reply Message' from other nodes Thus, a voice command processing method using a Bluetooth mesh network, characterized in that the node with the best performance is found and selected as the centralized node.
The method of claim 1, wherein the centralized node,
Estimates the user's current location based on the physical group information of other nodes that have transmitted the distributed voice data,
As the current location of the user is estimated, even if the physical group information is omitted when the user inputs a voice command, the physical group corresponding to the estimated current location of the user is automatically recognized to simplify the voice command. Voice command processing method using a Bluetooth mesh network.
The method of claim 1, wherein the nodes of the Bluetooth mesh network,
A voice command processing method using a Bluetooth mesh network, characterized in that the voice command processing device and the control target device are integrated into one.
The method of claim 1, wherein in the step of dividing the voice data into a plurality of voice data according to a predetermined method,
And performing a scaling process of expanding or reducing the input voice data according to a preset reference value.
The method of claim 1,
When the voice data is divided in a predetermined manner,
The control unit of the specific node,
Checking whether the number of divided voice data is less than a predetermined number,
As a result of the above check,
When the number of divided voice data is less than or equal to a predetermined number, the voice data is not distributed to a plurality of nodes, and the voice command processing apparatus of the specific node performs voice analysis and processing on its own. Voice command processing method using a Bluetooth mesh network.
The method of claim 1, wherein when the voice data is concentrated,
The control unit of the centralized node,
It is checked whether command essential information corresponding to the number of divided voice data is included, and if all of the command essential information corresponding to the number of voice data is included as a result of the check, a corresponding voice command is executed,
Voice command processing using a Bluetooth mesh network, characterized in that it is implemented to request re-input of a voice command including insufficient command essential information through an information output unit when the required command information corresponding to the number of voice data is not included. Way.
The method of claim 1, wherein in the step of dividing by the control unit of the specific node into voice data to be distributed and processed for each node of the Bluetooth mesh network,
The control unit of the specific node,
The number of nodes, and
Depending on the performance of each node or the external device connected to each node,
A voice command processing method using a Bluetooth mesh network, comprising adjusting whether voice data is distributed, a distribution priority of voice data, and the number of voice data to be distributed.
The method of claim 9, wherein in order for the controller of the specific node to classify voice data to be distributed and processed for each node of the Bluetooth mesh network,
The control unit of the specific node,
The voice command processing method using a Bluetooth mesh network, characterized in that referring to a performance table in which the number of nodes and performance information of each node or an external device connected to the node are stored.
The method of claim 10, wherein the performance table,
The control unit of each node of the Bluetooth mesh network,
A'performance question request message' is transmitted to another node or an external device connected to the node, and is generated based on the information contained in the'performance reply message' that the other node or an external device connected to the node replies,
The performance reply message,
A voice command processing method using a Bluetooth mesh network, comprising at least one of CPU clock information, CPU core number information, current CPU usage rate information, and current spare memory information.
The method of claim 11,
The control unit of each node of the Bluetooth mesh network,
In order to select the node that can process the voice data the fastest or the external device connected to the node,
In addition to the information of the'Performance Reply Message',
Voice command processing method using a Bluetooth mesh network, characterized in that additionally reflects the leap information and distance information.
The method of claim 1, wherein the node of the Bluetooth mesh network,
A method of processing a voice command using a Bluetooth mesh network, characterized in that the node itself is classified based on a physical group according to a location of a target device to be controlled, a type of the target device, and a number of the target device.
The method of claim 13, wherein the physical group,
The user sets it using an external device, or
A method for processing voice commands using a Bluetooth mesh network, characterized in that the node itself communicates with other nodes of the Bluetooth mesh network and can be automatically set based on distance information.
The method of claim 14,
In order to automatically set the physical group based on distance information by communicating with other nodes of the Bluetooth mesh network by the node itself,
When a new node is newly installed while the nodes that have already registered a physical group are installed,
The new node requests registration of a physical group to another node,
Each physical group node estimates the distance between itself and the new node and transmits this distance information to the centralized node,
The centralized node compares the distances between other nodes in the physical group to which it belongs and nodes in the other physical group, and searches which physical group the new node is closest to,
The voice command processing method using a Bluetooth mesh network, characterized in that the centralized node is implemented to determine information on the physical group closest to the new node according to the search result.

Images