JP2003195891A - Electronic apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an electronic apparatus that a user can place in a predetermined operation through speech recognition without specifically indicating the start of the speech recognition. <P>SOLUTION: When a detecting means detects a state wherein 'a thing is put in a box', 'a door is closed', 'the temperature in the box is 0 to 30°C', and 'a timer is not set', a recognition processing control means in a speech recognizing device refers to 'various state tables' in a recognition object word managing means to obtain a state number '30' corresponding to the state. Then recognition word set numbers '10000000', '10000001', and '10000005' corresponding to the state number '30' are acquired from a 'recognition word set use tables by states' and a recognition word, explanatory speech data, display character data, and notification sound data corresponding to the recognition word set number are acquired from a 'table of words included in a recognition word set and its detailed data'. The acquired recognition word is registered in a speech recognizing means, which starts speech recognition according to it. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to an electronic device having a voice recognition function.

[0002]

2. Description of the Related Art Conventionally, the problems to be solved by the invention
Electronic devices using voice recognition technology have been proposed in various fields. These devices require an explicit start instruction such as a specific switch operation by a user, voice generation, or a physical action as a start trigger for voice recognition itself. Therefore, when the user gives a voice instruction to the electronic device to perform a predetermined operation, it is necessary to give an explicit start instruction of the voice recognition and then speak, which is inconvenient. was there.

Therefore, the present invention has been made in view of the above problems.
An object of the present invention is to provide an electronic device that allows a user to perform a predetermined operation by voice recognition without explicitly instructing to start voice recognition.

[0004]

According to the electronic device of claim 1, which has been made to solve the above-mentioned problems, the electronic device is required to operate as an electronic device. It is determined whether to start the voice recognition or whether to end the voice recognition based on the state related to, the voice recognition is performed, and a predetermined operation is performed based on the recognition result. For example, if the electronic device is a microwave oven, whether the door is closed, whether there is an object in the refrigerator, how much temperature is in the refrigerator, etc. Etc.) is detected and voice recognition is started or ended based on such a state related to the electronic device necessary to perform the original operation of the electronic device. For example, in the case of such a microwave oven, voice recognition is started when the door is closed, there is an object in the refrigerator, and the temperature in the refrigerator is 0 ° C to 30 ° C. Therefore, it is not necessary to input a start trigger for explicit voice recognition itself such as a specific switch operation, voice generation, or physical action, and a predetermined action is performed by voice recognition in a normal usage pattern of an electronic device. Can be made.

Particularly, as described in claim 2, it is preferable that the start or the end of the voice recognition is determined based on the change of the state detected for the operation as the electronic device to perform the voice recognition. For example, in the case of a microwave oven, voice recognition is started by detecting that the door is closed. In this way, the user does not need to explicitly input the start point and end point of the voice recognition, and the voice recognition corresponding to the operation is automatically performed only by performing a normal operation of the electronic device. Will be started and ended. It should be noted that such a state change includes a change in electrical or temporal state inside or outside the system of the electronic device, or reaching a certain set state.

By doing so, for example, in order to start the voice recognition, an explicit voice recognition start instruction such as a conventional specific switch operation or instruction voice generation or physical action is given. Detects (monitors) state changes such as mechanical, electrical, and temporal state (situation) changes inside and outside the electronic device system and reaching a predetermined setting without the need for a "start trigger" As a result, the electronic device can spontaneously (automatically) start and end the voice recognition process from the “normal” use operation. Then, a predetermined operation can be performed as the operation of the electronic device based on the result of the voice recognition. Note that the predetermined operation is an operation that realizes the function of the electronic device, for example, not only an operation involving movement caused by driving a motor, an actuator, etc., but also a change in display content on a display device, for example. Including those that do not move. That is, it includes an operation of executing a predetermined process and realizing a predetermined function.

In this way, the next operation or the like should be performed without making the user aware that an extra manual operation or a specific start trigger for instructing the start of voice recognition such as voice utterance or physical action is required. It is possible to provide an electronic device having a voice recognition function capable of inputting a voice.

In addition to the function of deciding the start point or the end point of the voice recognition based on the change in the state detected to perform the operation as the electronic device and performing the voice recognition, as described in claim 3. Then, as in the conventional case, the instruction of the start / end of the voice recognition may be input by inputting the "specific start trigger". In this way, the user directly inputs the voice recognition start instruction to the electronic device,
Voice recognition can be performed.

And, in such an electronic device, the fourth aspect is provided.
As shown in, the set of recognition words such as a voice recognition dictionary used for voice recognition may be switched according to the detected state. For example, the set of vocabulary that can be recognized is switched according to the change in the detected state. By doing so, it is possible to perform appropriate voice recognition corresponding to the state. That is, the vocabulary search range can be limited in accordance with the state, as compared with the case where a general-purpose speech recognition dictionary is used as it is, so that the processing amount of speech recognition can be reduced. Therefore, it becomes possible to use a microcomputer or the like having a relatively low processing speed, and it is possible to reduce the cost of the electronic device. Moreover, since the set of recognition words such as the voice recognition dictionary is changed according to the state, only the dictionary of the vocabulary to be recognized in that state can be used, and the recognition rate can be increased.

Further, for example, according to a fifth aspect,
It is possible to change the set of recognition words used for voice recognition in correspondence with a state detected from some setting state or the like of a plurality of sensors or the electronic device. Then, when the voice recognition start point is determined based on such a state and the voice recognition is started, as shown in claim 6, it is preferable to give a notification indicating the start of the voice recognition. For example, the start of voice recognition is displayed or notified by voice or the like from a display device, a voice output device, or the like. In this way, by voluntarily letting the electronic device inform the user that voice input is possible, the user can be prompted to speak. For example, it is possible for the electronic device to easily interact with the user and to allow the user to select which of a plurality of operations the electronic device has, for example.

Note that such electronic devices include, for example, in-vehicle devices, home electric appliances, various terminals for providing ATM and other various services, manufacturing devices, and the like. For example,
A navigation device may be provided as described in claim 7.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments to which the present invention is applied will be described below with reference to the drawings. Needless to say, the embodiment of the present invention is not limited to the following examples, and various forms can be adopted as long as they are within the technical scope of the present invention. [First Embodiment] Electronic device 100 of the first embodiment shown in FIG.
Is a microwave oven, and includes the voice recognition device 1 and the electronic device 10.
It is provided with various state detectors 2 for detecting various states for performing 0 operation, a voice recognition operation switch 3, a microphone 4, an operation processing section 5, and a storage medium management section 6.

The voice recognition device 1 includes a control device 1a and state detecting means 1 for detecting a state from various state detectors 2.
b, voice input means 1d for A / D converting the voice signal from the microphone 4 and inputting it as a digital voice signal, and outputting a guide voice during voice recognition and interpreting the recognition result to the operation processing unit 5. Control information output means 1e for outputting information necessary for operation, and selecting a recognition target word according to the detected state, managing information such as an operation method for each recognition target word, and the like. The recognition target word management means 1f is provided. These are configured as peripheral circuits of a computer, or implemented as software processing in the same computer.

The control unit 1a is constructed as an ordinary computer, and has a well-known CPU and ROM / R inside.
A memory such as an AM, an I / O, a bus line connecting these components, and the like are provided, and when the CPU executes a program stored in the memory, at least the recognition processing control unit 1aa and the voice recognition unit 1ab are provided. A function is realized, and the state detecting means 1b and the operation input detecting means 1 described above are realized.
c, the voice input means 1d, the control information output means 1e, the recognition target word management means 1f, and other peripheral means are integrally controlled. The recognition processing control means 1aa and the voice recognition means 1ab may be realized by the processing by a single CPU or a plurality of CPs.
You may implement | achieve by the process by U.

The recognition target word management means 1f may manage only the information held in the ROM, RAM, etc. constituting the voice recognition device 1, or by acquiring the information from the storage medium management unit 6 and managing it. You may increase the types of operations.
The storage medium 6b in the storage medium management unit 6 is a well-known RO.
In addition to M and RAM, it may be a recording medium such as a CD, a DVD, an HDD, a memory card, and various DBs existing on the network.

The various state detectors 2 are state detectors provided to realize the functions of the electronic device 100, and each of them is made up of known sensors and electrical / mechanical contacts. FIG. 1 shows an electronic device 100 in which a voice recognition device 1 is incorporated into a microwave oven, and various state detectors 2 include an in-compartment sensor 2a for detecting whether or not there is an object in the microwave oven, and an electronic device. A door opening / closing sensor 2b for detecting whether or not the range door is open, an inside temperature sensor 2c for detecting the inside temperature of the microwave oven, a timer 2d for setting the cooking time and the like are provided.

The voice recognition operation switch 3 is a switch for explicitly instructing the voice recognition device 1 to start or end the voice recognition process, and the voice recognition device 1 is input from this switch. It also operates according to the switch information. The motion processing unit 5 includes a control unit (motion instruction unit) 5a, a speaker 5b, a display device 5c, and the like, and outputs voice such as a guide and information to the display unit according to the voice recognition result output from the voice recognition device 1. Display. In the case of a microwave oven, like the "warm" operation, the actual electronic device 1
The device operation as "00" is performed. Although the control unit 5a in the operation processing unit 5 shown in FIG. 1 is shown as a separate part from the voice recognition device 1, the control device 1a performs both processes (control) of the voice recognition process and the operation process. May be.

Recognition target word management means 1 in the voice recognition device 1.
In f, the contents of the guide output corresponding to the state detected by the state detector 2 (voice data for explaining the recognition word, character data for displaying the recognition word on the screen, notification sound data for notifying the recognition), The information about the processing method of voice recognition such as the content of the recognition word set (dictionary) which is a set of recognition words, the parameter indicating the type of operation, and the like is stored. A voice recognition process is performed based on the information stored in the word management unit 1f.

As shown in FIG. 2, the recognition target word managing means 1f of the present embodiment gives the states of the sensors of the various state detectors 2 detected by the state detecting means 1b and the states given for each combination of the states. Speaker 5 corresponding to the number and its state
“Various state table” that stores and holds the content of the guide voice to be output from “b” and the like, and the number (recognized word set number) of the recognition word set (that is, dictionary) that corresponds to the state number of the various state tables Recognition word set (dictionary) usage table for each state ", a recognition word corresponding to a recognition word set number, explanation voice data used for notification from the speaker 5b, display character data and notification used for notification from the display device 5c. It is provided with a “table of words included in a recognition word set (dictionary) and detailed data thereof” that stores and holds sound data, and the control device 1a refers to these tables to determine the start point and end point of voice recognition. Then, the content of the notification is determined and notified as necessary, and voice recognition is performed.

For example, the recognition processing control means 1aa in the voice recognition device 1 uses the state detection means 1b to "put an object in the refrigerator""close the door""inside the refrigerator temperature of 0 ° C to 30".
When it is detected that the temperature is "° C" and "the timer is not set", the "various state table" of the recognition target word managing means 1f is referred to and the state number "30" corresponding to this state is obtained.
Then, from the "recognition word set usage table for each state", the recognition word set "10000000""10000001""10000005" corresponding to the state number "30" is obtained, and "the words included in the recognition word set and their details" are obtained. From the table of data, this recognition word set "10000000""1000000
The recognition word corresponding to "1" and "10000005", the explanation voice data, the display character data, and the notification sound data are acquired. Then, the recognition words (word set group, dictionary) are registered in the voice recognition means 1ab.

The recognition processing control means 1aa outputs the obtained guide output information to the motion processing section 5 using the control information output means 1e. For example, (guidance voice) "Is it warm? Is it decompression? Please speak" (Signal) "Peep" is performed from the speaker 5b, and the voice input waiting state is entered.

At this time, as described above, the word set (dictionary) group to be recognized is set in the voice recognition means 1ab, and the recognized words are limited to "warm and warm" and "kai and kaiishi". It The voice recognition means 1ab is a microphone 4
From the above, the voice uttered by the user is captured and recognized via the voice input means 1d.

The recognized result is, for example, "warm and warm".
If so, the recognition processing control means 1aa acquires the explanation voice data, the display character data, and the notification sound data corresponding to the recognition word “warm and warm” from the recognition target word management means 1f, and performs the operation. It is determined whether or not all the information is available, and an instruction is output to the operation processing unit 5 via the control information output unit 1e, and the speaker 5b (guide voice) "Warm, isn't it?" (Guide voice) "time is specified. Please output a voice message such as “Peep” and wait for voice input.

When the user inputs "1 minute and 30 seconds" by voice, in the same manner as above, (guide voice) "1 minute and 30 seconds warm up" is guided, and the control section in the operation processing section 5 is guided. 5a is instructed to start warming. The operation processing unit 5 controls a circuit that realizes a function as a microwave oven (not shown) to perform a warming operation.

In a microwave oven, generally, an object (for example, frozen food) is put in the refrigerator, and after the door is closed, "warming", "time designation", etc. are manually input. In the prior art, in this case, the user presses some operation buttons and manually inputs the necessary conditions, or even when inputting by voice recognition, the user presses the voice recognition operation switch 3 to start voice recognition. Although it was necessary to give an explicit instruction, when the present invention is applied, since voice recognition is automatically started after the door is closed, all necessary instructions can be input by voice, and manual operation is not possible. It becomes completely unnecessary.

Moreover, a word set (dictionary) to be voice-recognized according to the states detected by the various state detectors 2
Since the group can be determined and the recognition words can be limited, appropriate recognition can be performed according to the situation, the recognition rate is improved, and the processing is greatly reduced compared to the case of using a general-purpose dictionary. can do. Further, when the word is recognized, the corresponding guide voice and the operation as the electronic device 100 (for example, warming operation) can be obtained.

As a well-known technique, there is a voice recognition means for recognizing continuous utterances of a plurality of words by providing clauses and grammars in a plurality of dictionary sets and associating them with each other. When such a voice recognition means is used. Is "1 minute 3" at the first voice input.
It is also possible to reduce the number of voice inputs by performing a plurality of instructions at the same time, such as "Because it was 0 seconds, so start".

In addition, for example, after the voice recognition is started, when it is detected from the various state detectors 2 that the user touches any of the operation buttons of the microwave oven, it is determined that the voice operation is not desired, The voice recognition may be ended. Further, when the voice recognition device 1 detects that the voice recognition operation switch 3 has been pressed, the voice recognition may be started as in the conventional case.

In the configuration of the electronic device 100 shown in FIG. 1, the state detecting means 1b realizes the function as the detecting means in the claims, and the control device 1a realizes the function as the judging means. The device 1a and the control information output means 1e realize the function as a recognition means. Also,
The voice recognition operation switch 3, the operation input detection means 1d, and the control device 1a realize the function as the trigger detection means, and the control means 1a, the control information output means 1e, and the operation processing unit 5 realize the function as the notification means. There is. [Second Embodiment] Next, an embodiment in which the present invention is applied to a navigation system will be described. The electronic device 100 of the second embodiment is a navigation system, and in the configuration of the electronic device 100 of the first embodiment shown in FIG. 1, the operation processing unit 5 performs navigation operation / processing, and various state detectors. Reference numeral 2 is a touch panel integrally formed with the display device 5c, a sensor such as a vehicle speed sensor for realizing the function of the navigation system, and the like.

The recognition target word management means 1f is shown in FIG.
The table of FIG. 3 is used instead of the table of FIG. That is, as shown in FIG. 3, the states (screen state, guidance state, running state, display state of the address list) of the sensors of the various state detectors 2 detected by the state detecting unit 1b and the states are given for each combination. “Various state table” that stores and holds the state number and the content of the guide voice that should be output from the speaker 5b or the like corresponding to the state, recognition words of all recognizable words, and corresponding explanation voice data -It is provided with "a table of all words and their detailed data" that stores and holds display character data, notification sound data, and a state number to be applied. Then, the control device 1a refers to these tables to determine a start point, an end point, and the like of voice recognition, determines the notification content as necessary, and notifies the voice recognition. That is,
The control device 1a identifies the state number corresponding to the state detected by the state detection means 1b using the "various state table", and the state number is "Apply" in "Table of all words and their detailed data". The recognition word existing in the "state number to perform" is extracted and registered in the voice recognition means 1ab. In this way, the recognized word set can be limited.

Further, for example, the recognition processing control means 1aa in the voice recognition device 1 causes the state detection means 1b to display "the address list in Aichi prefecture or below is being displayed because the destination is set by the address", When "Entered" is detected, the control information output means 1e instructs the motion processing section (guide voice) to set "Destination. Please speak the address of Aichi Prefecture from the municipality name" (Sign) A voice output such as “P” is output, and the voice input standby state is set. The recognition word set is limited as follows.

That is, in the storage medium 6b managed by the storage medium management unit 6, the recognition words of addresses nationwide are shown in FIG.
It is held in a hierarchical structure as shown in (a), the storage medium 6b is read by the data input device 6a, and the recognition processing control means 1aa obtains a recognition word set from the recognition words of the address based on this hierarchical structure, Register to the voice recognition means 1ab. For example, when the display device 5c displays a screen for setting destinations in Aichi prefecture or lower as shown in FIG. 4B, the recognition words of the recognition words in the hierarchy in FIG. Is registered in the voice recognition means 1ab. For example,
"Aichi Prefecture" is managed as ID = 00023, and in the voice recognition means 1ab, by registering the recognition word of the address and the ID of "Aichi Prefecture" in the voice recognition means, only the recognition words below Aichi Prefecture are recognized. Speech recognition is performed by a known method as a recognition target.

At this time, the recognition processing control means 1aa refers to the recognition target word management means 1f and limits the words to be recognized to addresses in Aichi prefecture and below. When the user gives a voice instruction here, the voice is taken in via the voice input means 1d,
The voice recognition means 1ab recognizes.

For example, if the recognition result is "Naka-ku, Nagoya", the recognition processing control means 1aa.
Acquires the information necessary for the motion from the recognition target word management unit 1f and outputs an instruction from the control information output unit 1e to the motion processing unit 5a (guide voice) "Naka Ward, Nagoya City, Aichi Prefecture. The guide is informed, and the control unit 5a in the motion processing unit 5
Controls to calculate the route to the destination with Naka-ku, Nagoya city, Aichi prefecture as the destination.

By thus limiting the recognition word set based on the hierarchical structure, the processing load of voice recognition can be reduced and the recognition rate can be increased.
In a navigation system, it is generally possible to set a destination from an address list or the like and provide guidance. However, since it is dangerous for the user to operate the list by looking at the screen while traveling, there is a control that restricts the operation to prevent complicated manual operation when the start of traveling is detected. . In the conventional technology, in such a case, the user had to stop the vehicle, operate the list again to display the desired display, and then start traveling after operating the destination. In the case of the configuration of the present embodiment, for example, the system detects that an address list is displayed and the vehicle has started running, automatically starts address voice recognition, and takes over the operation state of manual operation. And so on.

While monitoring the normal operation of the navigation system, the processing of the voice operation mode is
By having the function of spontaneously starting, so to speak, it is possible to completely eliminate unnecessary operations such as a user's voice recognition start instruction and provide an appropriate user interface according to the situation. [Others] The present invention can take various aspects other than the above-described embodiments. For example, the following may be done. (1) When the user cannot operate the electronic device while operating the electronic device due to some state change that occurs in the electronic device, the state change is used as a trigger to start voice recognition. By doing so, the voice operation can be performed by taking over the state of the manual operation up to that point.

For example, a car navigation system,
In the equipment used under vehicle conditions, the running state is detected by capturing the output from the speed detection device,
There is a system (electronic device) that makes it impossible for the driver to manually operate. However, according to the system to which the present invention is applied, even when the user starts traveling while operating the electronic device, the state during the operation can be taken over and the operation can be continued by the voice recognition, which is convenient for use. Can be improved. (2) When the electronic device is in some state and it is determined that the user will perform some operation from that state, the voice recognition is performed even if the user does not operate the specific trigger for starting the voice recognition. By starting it, voice operation becomes possible.

For example, there is an electronic device such as a telephone having a hands-free device, which manually selects whether to use an ordinary handset or to make a hands-free call when an incoming call is received. At this time, with the system to which the present invention is applied, it is possible to improve the usability in that the voice recognition is started by using an incoming call as a trigger and the processing can be started without requiring any switch operation.

Further, for example, in the case of an electronic device equipped with an e-mail receiving function and a text document reading device, the system to which the present invention is applied detects the arrival of an e-mail and starts voice recognition. Then, by instructing to read the contents of the received mail, the usability can be improved. (3) In the case of (2) above, the user may want to perform a manual operation, so that the voice recognition may be started after a certain period of time has passed since the system reached a prescribed state. Good.

For example, in an electronic device such as a telephone, a game, a car navigation system, etc., which has an "operation menu screen" in addition to a normal screen display, it is difficult to leave the operation menu screen as it is for a long time. It is possible to improve the usability by starting voice recognition, for example, 10 seconds after the screen state is changed by the operation of. (4) In the example of (3) above, when there are a plurality of operation menu screens and condition setting screens, when voice recognition is started, a guide (notification) is provided for each operation menu screen to be set. However, by limiting the vocabulary that can be recognized and then entering the voice recognition state, it is possible to eliminate the unpleasant feeling of the user due to the recognition of unrelated words.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an electronic device according to an embodiment.

FIG. 2 is an explanatory diagram showing a table configuration of a recognition target word management means in the electronic device of the first embodiment.

FIG. 3 is an explanatory diagram showing a table configuration of a recognition target word management means in the electronic device of the second embodiment.

FIG. 4 is an explanatory diagram for associating display contents on a display device with a recognizable vocabulary in the electronic device of the second embodiment.

[Explanation of symbols]

100 ... Electronic device 1 ... Voice recognition device 1a ... Control device 1aa ... Recognition processing control means 1ab ... Voice recognition means 1b ... State detection means 1c ... Operation input detection means 1d ... Voice input means 1e ... Control information output means 1f ... Recognition target Word management means 2 ... Status detector 2a ... In-compartment sensor 2b ... Door opening / closing sensor 2c ... In-compartment temperature sensor 3 ... Voice recognition operation switch 4 ... Microphone 5 ... Motion processing unit 5a ... Control unit 5b ... Speaker 5c ... Display device 6 ... Storage medium management unit 6a ... Data input device 6b ... Storage medium

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) G08G 1/0969 G10L 3/00 571K 5H180 G10L 13/00 571V 15/00 551N 15/22 551Q 571H R F term (Reference) 2F029 AA02 AB13 AC13 AC18 3L086 CA10 DA24 3L087 AA01 BA05 DA24 5D015 HH13 KK01 KK04 LL07 LL10 5D045 AB15 AB17 5H180 AA01 BB12 BB13 EE02

Claims (7)

[Claims] 1. An electronic device that performs a predetermined operation, and a detection unit that detects a state related to the electronic device that is necessary to perform the operation as the electronic device, and a state detected by the detection unit. Determination means for determining whether or not to start the voice recognition based on at least a part of, and to determine the start or end of the voice recognition based on the determination result of the determination means An electronic device comprising: a recognition unit that performs voice recognition and that performs the predetermined operation based on the recognition result. 2. The electronic device according to claim 1, wherein the determination means determines whether to start voice recognition or to end voice recognition based on a change in a state detected by the detection means. An electronic device characterized by making a judgment. 3. The electronic device according to claim 1, wherein the recognition unit includes a trigger detection unit that detects a trigger for starting or ending voice recognition from a user, and the recognition unit detects the trigger. An electronic device which starts or ends voice recognition based on a trigger. 4. The electronic device according to claim 1, wherein the recognition unit is a recognition word used for voice recognition in association with a state related to the electronic device detected by the detection unit. Electronic device characterized by changing the set. 5. The electronic device according to claim 1, wherein the detection unit detects a plurality of states, and the recognition unit detects a recognition word corresponding to each combination of the plurality of states. By storing a set or extracting a vocabulary from all the stored recognition words according to the state and creating a set of recognition words, recognition corresponding to the combination of the current states detected by the detection means is performed. An electronic device characterized by performing speech recognition using a set of words. 6. The electronic device according to claim 1, further comprising: an informing unit, wherein the informing unit starts the voice recognition when the determining unit determines to start the voice recognition. An electronic device characterized by making a notification indicating that. 7. The electronic device according to claim 1, wherein the electronic device is a navigation device.