JP6983405B2 - Cooker - Google Patents

Description

The present invention relates to a cooking device.

In a stove equipped with a burner, recognizing a user's voice command and performing control according to it is effective, for example, when the hand cannot be released during other work or when the hand is dirty and difficult to operate. .. In the sound signal control type stove described in Patent Document 1, a sound signal control unit is attached to the control unit. The sound signal control unit recognizes a memory that stores at least one control sound signal corresponding to a predetermined operation of the heater, a sound signal input unit, and a sound signal recognition unit that recognizes the input sound signal. It is composed of a sound signal collating unit that collates the generated sound signal with the control sound signal in the memory. When a sound signal is input, when the sound signal recognized by the sound signal recognition unit and the stored sound signal match in the collation by the sound signal matching unit, the control unit operates the burner in response to the sound signal. To control. To determine whether or not the sound signal recognized by the sound signal recognition unit matches the sound signal stored in the memory, for example, the matching rate of these two sound signals is calculated and the calculated matching rate is calculated. When the sound signal exceeds the threshold value, it is generally judged that the sound signals match. If the sound signal recognized by the sound signal recognition unit does not match the sound signal stored in the memory, the control unit does not execute the operation corresponding to the sound signal.

Japanese Patent No. 4313495

When the user inputs a fire extinguishing instruction with high urgency by voice, for example, if the user's pronunciation is bad and some of the sounds are difficult to recognize, it matches the sound signal stored in the memory. Since the rate did not exceed the threshold, it was possible that the burner could not be extinguished promptly.

An object of the present invention is to provide a cooking device capable of more reliably recognizing a fire extinguishing instruction by voice.

The heating cooker according to claim 1 is stored in a storage means for storing acoustic model data indicating selectable operation instructions by voice information, voice data of voice instructions input from the outside, and the storage means. A heating cooker provided with a calculation means for calculating a matching rate with the acoustic model data and a determination means for determining that the voice instruction is recognized when the matching rate calculated by the calculation means exceeds a threshold value. In the above, the first threshold value when the determination means recognizes the fire extinguishing instruction, which is the voice instruction for instructing the fire extinguishing of the heating source by voice, is from the second threshold value when recognizing the voice instruction other than the fire extinguishing instruction. Is also low.

According to the heating cooker according to claim 1, in order to prevent erroneous operation due to noise when inputting a voice instruction, the cooking cooker calculates a matching rate with acoustic model data, and the matching rate exceeds a threshold value. It is judged that the voice instruction is recognized only when the voice instruction is recognized. In such a cooking device, the first threshold value for determining the recognition of the fire extinguishing instruction is set lower than the second threshold value for recognizing the voice instruction other than the fire extinguishing instruction. As a result, it is possible to more reliably recognize a fire extinguishing instruction that is more urgent than other voice instructions, so that it is possible to respond without loss in an emergency. The operation instruction is an instruction for operating the cooking cooker, and the voice instruction is a voice instruction corresponding to the operation instruction.

It is a perspective view of a stove 1. It is a block diagram which shows the electric structure of a stove 1. It is a conceptual diagram at the time of voice recognition of a voice instruction "heat power large". It is a conceptual diagram of the threshold table 741. It is a flowchart of voice operation control processing. It is a flowchart which shows the continuation of FIG. It is a flowchart of a voice recognition process. This is an example of a state transition diagram when a voice instruction is input. It is a figure which shows the state which the explanation screen 101 is displayed on the display part 22. It is a figure which shows the state which the explanation screen 102 is displayed on the display part 22.

Hereinafter, embodiments of the present invention will be described. Unless otherwise specified, the structure, flowchart, and the like of the apparatus described below are not intended to be limited thereto, but are merely explanatory examples. The drawings are used to illustrate the technical features that may be employed in the present invention. The following description uses left and right, front and back, and up and down indicated by arrows in the figure.

The structure of the stove 1 will be described with reference to FIG. The stove 1 is a built-in stove. A top plate 3 is installed on the upper surface of the stove 1. In the top plate 3, a right stove burner 4 is provided on the front right side, a left stove burner 6 is provided on the front left side, and a back stove burner 5 is provided on the back side of the center. At the center of each stove burner 4 to 6, a sensor assembly 20 that comes into contact with the bottom of the pot and detects the temperature of the cooking pot is provided. The sensor assembly 20 is provided so as to be able to move in and out in the vertical direction, and houses the thermistor 40 (see FIG. 2). A thermocouple 50 (see FIG. 2) for detecting a misfire and an igniter 60 (see FIG. 2) for igniting are provided in the vicinity of the stoves 4 to 6. At the rear portion of the top plate 3, an exhaust port 7 of a grill storage (not shown) installed in the stove 1 is provided. A rectangular display unit 22 is provided at a substantially central portion on the front side of the upper surface of the top plate 3. The display unit 22 is, for example, a liquid crystal display. In addition to the operation screen of the stove 1 (not shown), the display unit 22 displays explanatory screens 101 to 104 (see FIGS. 8 to 11) for voice operation, which will be described later.

A grill door 8 is provided at substantially the center of the front surface of the stove 1. The grill door 8 is movably supported toward the front side, and opens and closes an opening on the front side of the grill storage (not shown) provided inside the stove 1. A grill burner (not shown) is provided in the grill chamber. A thermocouple 50 and an igniter 60 are also provided in the vicinity of the grill burner. The grill burner may be provided with, for example, a top heat burner and a bottom heat burner. In the area on the right side of the grill door 8, operation buttons 11 to 13 having a circular front view are provided in a horizontal row in order from the right side to the left side. Also in the area on the left side of the grill door 8, the operation buttons 14 having the same circular shape are provided at the same height positions as the operation buttons 11 to 13.

The operation button 11 is the right stove burner 4, the operation button 12 is the back stove burner 5, the operation button 13 is the left stove burner 6, and the operation button 14 is pressed to ignite and extinguish the grill burner. The operation buttons 11 to 14 are almost flush with the front surface of the stove 1 when the corresponding burner is extinguished (see FIG. 1). When pressed for ignition, the control circuit 70 described later executes ignition processing for the corresponding burner, and a known push-push mechanism (not shown) allows the operation buttons 11 to 14 to be pressed from the front surface of the stove 1. It protrudes in a substantially columnar shape toward the front, and the rotation operation is possible in the protruding state. When the operation buttons 11 to 14 are rotated, the control circuit 70 adjusts the gas supply amount to the corresponding burner so that the gas amount corresponds to the rotation operation amount.

Inside the stove 1, four gas amount adjusting mechanisms 30 (see FIG. 2) are provided. The gas amount adjusting mechanism 30 adjusts the gas supply amount to the corresponding stove burners 4 to 6 and the grill burner. The gas amount adjusting mechanism 30 includes a valve mechanism portion 31 and a motor 32. The valve mechanism portion 31 includes various valves such as a needle valve and a solenoid valve that adjust the amount of gas flowing through the gas flow path, for example. The motor 32 drives the valve of the valve mechanism portion 31 to adjust the gas flow rate flowing through the gas flow path. The drive of the motor 32 is controlled by the control circuit 70.

When the operation buttons 11 to 14 are pushed in and projected toward the front, the control circuit 70 energizes the motor 32 of the gas amount adjusting mechanism 30. As a result, the valve mechanism portion 31 is driven and the gas flow path is opened, so that gas is supplied to the corresponding burner. Then, the control circuit 70 rotates the motor 32 in the forward and reverse directions according to the rotation direction and the rotation angle of the operation buttons 11 to 14, and adjusts the opening degree of the valve in the valve mechanism unit 31. As a result, the amount of gas flowing through the gas flow path is adjusted, so that the thermal power of the corresponding burner is adjusted. Further, when the operation buttons 11 to 14 are pushed into the stove 1 and accommodated, the control circuit 70 rotates the motor 32 in the reverse direction. As a result, the valve of the valve mechanism portion 31 is driven in the opposite direction, and the gas flow path is closed, so that the supply of gas to the corresponding burner is cut off.

A power button 25 is provided directly above the operation button 11. A hole 17 penetrating the panel on the front surface of the stove 1 is provided diagonally above the left of the operation button 14, and the microphone 76 is provided so as to face the hole 17 from the inside of the stove 1. The stove 1 of the present embodiment has a voice operation function. The voice operation function is a function that receives a user's voice instruction by the microphone 76, identifies the content of the voice instruction by performing voice recognition, and enables the fire power operation of the burner targeted by the voice instruction. The voice instruction is, for example, an operation instruction for operating the thermal power of the burner, designating the location of the burner for performing the thermal power operation, and the like by voice to the stove 1. The voice operation control process (see FIGS. 5 to 7) for executing the voice operation function will be described later. Further, in this embodiment, for convenience of explanation, the right stove burner 4, the left stove burner 6, the back stove burner 5, and the grill burner may be referred to as a right stove 4, a left stove 6, a back stove 5, and a grill, respectively.

The electrical configuration of the stove 1 will be described with reference to FIG. The stove 1 includes a control circuit 70. The control circuit 70 includes a CPU 71, a ROM 72, a RAM 73, a non-volatile memory 74, and the like. The CPU 71 comprehensively controls various operations of the stove 1. The ROM 72 stores various programs such as a voice operation control program. The voice operation control program executes voice operation control processing (see FIGS. 3 and 4) described later. The RAM 73 temporarily stores various types of information. The non-volatile memory 74 stores various data such as burner thermal power information, acoustic model data, and threshold table 741 (see FIG. 3) in addition to various parameters for temperature control control. The burner thermal power information is information in which the rotation angle of the output shaft of the motor 32 is associated with the burner thermal power. In the present embodiment, for example, the thermal power from small to large is divided into 10 stages of 1 to 10, and the rotation angle of the motor 32 is assigned to each stage. Thereby, the CPU 71 can specify the corresponding burner thermal power based on the rotation angle of the output shaft of the motor 32. The acoustic model data and the threshold table 741 will be described later.

The control circuit 70 includes a power supply circuit 81, a thermista input circuit 82, a thermocouple input circuit 83, an igniter circuit 84, an operation panel 24, a motor control circuit 85, a voice input circuit 86, a voice output circuit 87, a buzzer circuit 88, and display control. The circuit 89, the sensor 19, and the like are electrically connected to each other. When the power button 25 is pressed by the user, the power supply circuit 81 steps down the alternating current (for example, 100V) supplied from the power supply 23 to a direct current (for example, 5V), rectifies it, and supplies electric power to various circuits. Therefore, the power of the stove 1 is turned on. When the power button 25 is pressed again by the user, the power supply circuit 81 cuts off the power supply to various circuits. Therefore, the power of the stove 1 is turned off.

The thermistor input circuit 82 inputs a detection signal from the thermistor 40 stored in the sensor assembly 20 provided in each burner to the control circuit 70. The thermocouple input circuit 83 inputs the detection value (signal corresponding to the thermoelectromotive force) from the thermocouple 50 to the control circuit 70. The igniter circuit 84 drives the igniter 60 of the corresponding burner based on the control signal from the CPU 71. The operation panel 24 is housed in a lower right portion of the front panel of the stove 1, and is pulled out toward the front side by pushing the portion. The operation panel 24 receives inputs such as timer setting by the user and selection of thermal power control according to the cooking menu, and inputs the input signal to the control circuit 70. The motor control circuit 85 controls the driving of the motors 32 of the four gas amount adjusting mechanisms 30 based on the control signals from the CPU 71.

The voice input circuit 86 converts the analog signal of the voice collected by the microphone 76 into a digital signal and inputs it to the control circuit 70. The audio output circuit 87 converts an audio digital signal into an analog signal based on the control signal from the CPU 71, and outputs the audio to the speaker 77. The buzzer circuit 88 drives the piezoelectric buzzer 78 based on the control signal of the CPU 71. The display control circuit 89 displays various screens on the display unit 22 based on the control signal from the CPU 71. The sensors 19 are provided on the operation buttons 11 to 14 respectively, detect the rotation direction and the rotation angle of the operation buttons 11 to 14, and input the detection signal to the control circuit 70. Based on the detection signal from the sensor 19, the CPU 71 rotates the output shaft of the motor 32 so as to correspond to the rotation direction and the rotation angle of the operation buttons 11 to 14, and drives the valve mechanism unit 31. As a result, the amount of gas flowing through the gas flow path is adjusted, and the thermal power of the corresponding burner is adjusted.

The gas amount adjusting mechanism 30 is provided with a position meter 35. The position meter 35 detects the rotation angle of the output shaft of the motor 32, and inputs the detection signal to the control circuit 70. Since the CPU 71 can recognize the rotation angle of the output shaft of the motor 32 based on the detection signal from the position meter 35, the corresponding burner thermal power can be specified by referring to the burner thermal power information stored in the non-volatile memory 74. The stove 1 may include a communication unit capable of wirelessly or wiredly communicating with an external device. In that case, the stove 1 can receive various programs and data from the external device via the communication unit, and can update the control contents, for example.

The principle of speech recognition of the stove 1 will be briefly described. For voice recognition, a well-known voice recognition technique can be used, for example, matching (matching) between "acoustic model data" stored in the non-volatile memory 74 and voice information (for example, waveform data) of the input voice instruction. It is good to do it at. "Acoustic model data" is a sound waveform created based on average pronunciation data, which is divided into elements called phonemes, and then "a", "i", "u", etc. This is model data for identifying which feature amount is possessed among the vowels and consonants such as "k", "s", and "t". The non-volatile memory 74 stores a plurality of acoustic model data in which operation instructions that can be selected by the user for operating the stove 1 are indicated by voice waveforms. For matching, for example, a theory called "HMM (Hidden Markov Model)" may be used. Matching is usually done in units of 10 to 20 milliseconds, starting from the beginning of the word.

An example of voice recognition will be described with reference to FIG. Speech recognition is executed in the speech recognition process (see FIG. 7) described later. Here, the description will be made on the assumption that a voice instruction of "high thermal power" is input to the microphone 76 from the user. When the voice instruction of "high thermal power" is input, the CPU 71 converts the voice data of the input voice instruction into waveform data. The CPU 71 matches the waveform data of "high thermal power" with the acoustic model data stored in the non-volatile memory 74. When matching with the acoustic model data of "Large thermal power", the acoustic model data of "Large thermal power" is, for example, five phonemes of "ka", "Ryo", "ku", "da", and "i". Separated into data. In the phoneme data, the feature amount corresponding to each sound is defined. On the other hand, the waveform data of the voice instruction input and converted is also separated into five phoneme data based on the time change. The CPU 71 matches the waveform data of the voice instruction with the acoustic model data for each phoneme data, and calculates the matching rate.

FIG. 3 illustrates two examples of the match rate, (1) and (2). In the example of (1), the match rates for each phoneme data of "ka", "ryo", "ku", "da", and "i" are 75%, 85%, 95%, 70%, and 85%. The overall average is 82%. On the other hand, in the example of (2), it is 73%, 82%, 62%, 74%, 75%, and the overall average is 73%. The overall average may be regarded as the overall match rate of the input voice instructions. The CPU 71 reads the threshold value corresponding to the "large thermal power" stored in the non-volatile memory 74, and compares the matching rate of each of (1) and (2) with the threshold value.

When the match rate exceeds the threshold value, the CPU 71 determines that the waveform data of the voice instruction matches the acoustic model data, and determines that the voice instruction has been recognized. When the match rate is equal to or less than the threshold value, the CPU 71 determines that the waveform data of the voice instruction does not match the acoustic model data, and determines that the voice instruction could not be recognized. For example, when the threshold value of "high thermal power" = 80%, it is determined that the example (1) matches, and the example (2) does not match. By using such a method, the CPU 71 can recognize the voice instruction. The voice recognition method may be any other method.

The threshold table 741 will be described with reference to FIG. The threshold table 741 sets a threshold for each voice instruction. The voice instructions that can be input to the stove 1 are, for example, "high firepower", "medium firepower", "small firepower", "fire extinguishing", "start voice operation", "end voice operation", "right stove", "back stove", "left stove", and "left stove". There are 10 types of "grill". The higher the threshold value, the higher the degree of matching with the acoustic model data is required. Therefore, it is difficult to recognize the voice instruction unless it is input to the stove 1 with clear and correct pronunciation. Therefore, it is possible to prevent erroneous recognition of external noise. On the other hand, the lower the threshold value, the higher the degree of matching with the acoustic model data is not required, so that it is possible to make it easier to recognize the voice instruction even if there is some deviation from the acoustic model data.

In the present embodiment, the threshold value of the voice instruction other than "fire extinguishing" is uniformly set to 80% (corresponding to the "first threshold value" of the present invention), and only the threshold value of "fire extinguishing" is lower than the threshold value of other voice instructions. It is set to 60% (corresponding to the "second threshold" of the present invention). By lowering the threshold value of "fire extinguishing", it is possible to more reliably recognize the voice instruction of "fire extinguishing", which is more urgent than other voice instructions, so that it is possible to respond without loss in an emergency. The threshold value may be the same value for all voice instructions, or may be changed for each voice instruction. Preferably, the threshold of the "fire extinguishing" voice instruction is set to a value lower than the threshold of other voice instructions. The non-volatile memory 74 stores at least 10 types of acoustic model data corresponding to the above 10 types of voice instructions.

The voice operation control process will be described with reference to FIGS. 5 to 7. When the user presses the power button 25 of the stove 1 and power is supplied from the power circuit 81 to the control circuit 70, the CPU 71 reads the voice operation control program from the ROM 72 and executes this process.

As shown in FIG. 5, the CPU 71 performs initial settings (S1). In the initial setting, the CPU 71 initializes the set values required for various processes, and for example, initializes the recognition failure number k to be stored in the RAM 73 to 0, which will be described later. The CPU 71 determines whether or not any of the burners of the right stove 4, the left stove 6, the back stove 5, and the grill is ignited (S2). The CPU 71 returns to S2 and waits until any burner is ignited (S2: NO). When any of the burners is ignited (S2: YES), the CPU 71 identifies the ignited burner based on the detection signal from the thermocouple 50 (S3), and the information of the igniting burner is used as the ignition burner information. Stored in RAM 73. The CPU 71 outputs an ignition guidance voice (S4). For example, when the right stove 4 is ignited, the voice "The right stove has been ignited" is output from the speaker 77. After that, even when another burner is ignited and a plurality of burners are in the ignition state, the CPU 71 identifies the ignited burner based on the detection signal from the thermocouple 50 and stores it in the RAM 73 as ignition burner information. do.

The CPU 71 outputs a voice explaining the method of starting the voice operation from the speaker 77 in order to inform the user of the method of starting the voice operation (S5). For example, the CPU 71 outputs a voice message "To start the voice operation, say" Start voice operation "" from the speaker 77. Upon hearing this, the user speaks to the microphone 76, "Start voice operation." The CPU 71 executes a voice recognition process in order to recognize a voice instruction of "start voice operation" from the user (S6).

The voice recognition process will be described with reference to FIG. 7. The CPU 71 determines whether or not there is a voice input by the microphone 76 (S41). Until there is a voice input (S41: NO), the CPU 71 returns to S41 and waits. When there is a voice input (S41: YES), the CPU 71 converts the input voice analog data into waveform data (S42). The converted waveform data is temporarily stored in the RAM 73. The CPU 71 sequentially reads the acoustic model data stored in the non-volatile memory 74 (S43), matches the waveform data stored in the RAM 73, and calculates the matching rate respectively (S44). The matching method and the calculation method of the match rate are as described above. The calculated match rate is temporarily stored in the RAM 73. The CPU 71 reads the threshold value corresponding to the matched voice instruction from the threshold value table 741 (see FIG. 4) stored in the non-volatile memory 74 (S45).

The CPU 71 determines whether or not the number of recognition failures stored in the RAM 73 is 3 or more (S46). The recognition failure count k is the number of recognition failures for the same voice instruction. When the number of recognition failures k is less than 3 (S46: NO), the CPU 71 determines whether or not there is acoustic model data whose matching rate exceeds the threshold value among the matched acoustic model data (S48). When there is acoustic model data whose matching rate exceeds the threshold value (S48: YES), the CPU 71 determines that the recognition of the voice instruction corresponding to the acoustic model data has been successful (S49). When there are a plurality of acoustic model data whose matching rate exceeds the threshold value, it may be determined that the recognition of the voice instruction corresponding to the acoustic model data having the highest matching rate is successful.

When the CPU 71 succeeds in recognizing the "start of voice operation" from the user (S48: YES, S49), the CPU 71 stores the information of the voice instruction for which the recognition is successful in the RAM 73 as the voice recognition information (S50). The CPU 71 initializes the recognition failure number k stored in the RAM 73 to 0 (S51), and determines whether or not all the burner thermal powers are being changed to 1 (S52). Since not all the burner thermal powers are being changed to 1 (S51: NO), the CPU 71 ends the voice recognition process and shifts to the process of S7 in FIG.

On the other hand, in the CPU 71, the matching rate may not exceed the threshold value due to external noise, the accent of the user, the strength of pronunciation, and the like. In this case (S48: NO), the CPU 71 determines that the recognition of the voice instruction has failed (S54), and for example, outputs a voice "Could not recognize" from the speaker 77 to perform error notification (S55). ). The error notification may be performed by displaying an error on the display unit 22. The CPU 71 adds 1 to the recognition failure count k stored in the RAM 73 (S56), and determines whether or not the recognition failure count k after 1 addition is 4 or more (S57). When the number of recognition failures k is less than 4 (S57: NO), the CPU 71 returns to S41 and receives the voice instruction from the user again.

Since the user who received the error notification knows that the voice instruction was not recognized, the same voice instruction is repeatedly input until the recognition of the voice instruction is successful. When the recognition of "start voice operation" is successful for the second time (S48: YES, S49), the CPU 71 stores the voice recognition information in the RAM 73 (S50) and initializes the recognition failure count k to 0 (S51). .. Since not all the burner thermal powers are being changed to 1 (S52: NO), the CPU 71 ends the voice recognition process and shifts to the process of S7 in FIG.

On the other hand, when the recognition of "start voice operation" fails three times in a row, it is unlikely that the voice is noise because the voice is repeatedly input. That is, it is highly possible that the cause of the recognition failure is the accent of the user's pronunciation, the strength of the pronunciation, the speed, and the like. In this case, even if the user intends to input the voice instruction with the correct pronunciation, the matching rate of matching does not exceed the threshold value, so that the voice instruction is not recognized forever. The user may be stressed if the voice instructions are repeatedly input and not recognized.

Therefore, in order to reduce the stress of the user, the CPU 71 lowers the threshold value used for determining the matching rate by 10% from the current value (S47). Since the threshold value of "start voice operation" is 80%, it is changed to 70%. The changed threshold value is stored in the RAM 73. As a result, the voice instruction of "start voice operation" becomes easier to be recognized as compared with the normal time. When the match rate exceeds the threshold value as a result of lowering the threshold value (S48: YES), the CPU 71 can succeed in recognizing "start of voice operation" (S49). As a result, the stove 1 can reduce the stress of the user due to the fact that the stove 1 is not recognized even if the voice instruction is repeatedly input. The CPU 71 stores the information of "start of voice operation" in the RAM 73 as voice recognition information (S50), and initializes the recognition failure count k to 0 (S51). Since not all the burner thermal powers are being changed to 1 (S52: NO), the CPU 71 ends the voice recognition process and shifts to the process of S7 in FIG.

On the other hand, even if the recognition failure number k is 3 and the threshold value is lowered (S46: YES, S47), if the match rate does not exceed the threshold value (S48: NO), the recognition of the voice instruction fails (S54), so that the CPU 71. Notifies an error (S55) and adds 1 to the number of recognition failures k (S56). In this case, the number of recognition failures k = 4, and the voice instruction is not recognized even if the threshold value is lowered. Therefore, it is highly possible that the user is inputting with the wrong voice instruction. Therefore, the CPU 71 re-notifies the inputable voice instruction in order to re-recognize the correct voice instruction to the user (S58). As an example of re-notification, for example, it is preferable to output the voice "To start the voice operation, please speak" Start voice operation "" from the speaker 77 as in the guide voice of S5. As a result, the user can reconfirm the correct voice instruction and can try the input again with the correct voice instruction.

Here, in the stove 1, even if the burner is ignited, the voice instruction is not recognized. Therefore, in order to temporarily weaken the thermal power, the CPU 71 identifies the burner currently being ignited and the respective burner thermal powers. Then, after temporarily storing it in the RAM 73 as the original thermal power information (S59), all the burner thermal powers currently being ignited are changed to 1 (S60). The burner thermal power may be specified by obtaining the rotation angle of the motor 32 of the gas amount adjusting mechanism 30 corresponding to the burning burner and referring to the burner thermal power information stored in the non-volatile memory 74. As a result, the thermal power is weakened for the time being until the voice instruction is recognized, so that the safety of the stove 1 can be improved. Since the user was able to reconfirm the correct voice instruction, the CPU 71 initializes the recognition failure count k to 0 (S61), returns to S41, and receives the voice instruction from the user again.

Then, as a result of matching the voice instructions re-input by the user, if there is acoustic model data whose matching rate exceeds the threshold value (S48: YES), the CPU 71 determines that the recognition of the voice instructions is successful. (S49). The CPU 71 stores the information of "start of voice operation" in the RAM 73 as voice recognition information (S50), and initializes the recognition failure count k to 0 (S51). Then, in the process of S59, all the burner thermal power is being changed to 1 (S52: YES), so that the CPU 71 returns the burner thermal power to the original thermal power based on the original thermal power information stored in the RAM 73 (S53). The voice recognition process is terminated, and the process proceeds to the process of S7 in FIG.

Returning to FIG. 5, when the voice recognition process (S6) is completed, the CPU 71 starts the voice operation according to the voice recognition information stored in the RAM 73, and at the same time, for example, gives a voice operation start guidance such as "Start the voice operation." , Output from the speaker 77 (S7). As a result, the user can recognize that the voice operation has been started on the stove 1. The CPU 71 determines the currently acceptable thermal power operation based on the burner currently being ignited (S8). In this embodiment, the thermal power operations that can be selected for each of the right stove 4, the back stove 5, the left stove 6, and the grill are 4 of "high thermal power", "medium thermal power", "small thermal power", and "fire extinguishing". Although it is a type, it may be changed for each burner.

Here, the state A shown in FIG. 8 is assumed. In the state A, the left stove 6, the back stove 5, and the grill are ignited. Therefore, the CPU 71 may determine "high thermal power", "medium thermal power", "small thermal power", and "fire extinguishing" as currently acceptable thermal power operations. The CPU 71 stores the determined information on the currently acceptable thermal power operation in the RAM 73. Based on the information on the currently acceptable thermal power operation stored in the RAM 73, the CPU 71 creates a guide voice for informing the user of the options of the items currently acceptable on the stove 1, and outputs the guide voice from the speaker 77 (S9). ). As a guide voice, for example, a voice saying "Please select one desired operation from high firepower, medium firepower, low firepower, fire extinguishing, and end voice operation and give a voice instruction." Is output. It should be noted that the items that can be accepted at present include "end of voice operation" in addition to the thermal power operation that can be accepted at present, which is convenient when switching from voice operation to manual operation in the middle.

At the same time, the CPU 71 displays the explanation screen 101 (see FIG. 9) on the display unit 22 (S10). The same content as the output guide voice is displayed on the explanation screen 101. As a result, the user can recognize the currently acceptable item by both voice and display, so that the word of the voice instruction can be clearly recognized. Especially during cooking, there are noises of various factors such as the sound of oil when frying ingredients in a frying pan, the sound of a ventilation fan, the sound when cooking utensils hit each other, etc. , You may miss it, or even if you hear it, you may forget it later. In the present embodiment, the guide voice is output from the speaker 77 and the explanation screen 101 is displayed on the display unit 22. Therefore, even if the explanation output by the guide voice is overlooked, it is displayed on the display unit 22. It can be confirmed on the explanation screen 101. Further, since the words of the items currently accepted can be clearly displayed on the explanation screen 101, the user can speak with accurate words to the stove 1. Therefore, the possibility of erroneous recognition of voice instructions can be reduced. In addition, it is possible to prevent the user from inputting a voice instruction with the wrong word.

The user who hears the guide voice selects one operation to be performed on the burning burner from the recognized items and speaks to the stove 1. The CPU 71 executes the same voice recognition process as in FIG. 6 (S11). In the voice recognition process, similarly to the above, the CPU 71 receives the voice instruction of the thermal power operation input from the user, calculates the match rate by matching with the acoustic model data, and is based on the calculated match rate and the threshold value. , Judge the success or failure of voice instruction recognition.

For example, in the state A shown in FIG. 8, it is assumed that the burner thermal power of the left stove 6 is increased. In this embodiment, the burner thermal power can be switched in 10 steps. For example, "Large thermal power" is a voice instruction to change the burner thermal power to 9, and "Medium thermal power" changes the burner thermal power to 5. It is a voice instruction to change, and "small firepower" is a voice instruction to change the burner firepower to 2.

As shown in FIGS. 3 and 7, when the CPU 71 receives a voice instruction of "high thermal power" from the user (S41), the CPU 71 converts the voice data of "high thermal power" into waveform data (S42), and the converted waveform. Divide the data into phoneme data. The CPU 71 sequentially reads the acoustic model data stored in the non-volatile memory 74 (S43), matches the waveform data stored in the RAM 73, and calculates the matching rate (S44). The CPU 71 reads the threshold value from the threshold value table 741 (S45) and compares the match rate with the threshold value (S46: NO, S48). As shown in FIG. 3, in the matching with the acoustic model data of "high thermal power", when the matching rate of each phoneme data is (2), the overall average is 73%, which is a threshold value of 80%. Lower than. Therefore, since the recognition of the voice instruction is unsuccessful (S48: NO, S51), the CPU 71 re-accepts the voice instruction from the user and tries to recognize the voice instruction again (S41 to S45).

If the voice recognition fails three times in a row (S46: YES), the voice is unlikely to be noise. Therefore, in order to make it easier to recognize the voice instruction from the user, the CPU 71 performs " The threshold value of "high firepower" is lowered by 10% from the current value (S47). The threshold for "Large Firepower" is changed from 80% to 70%. The changed threshold value is temporarily stored in the RAM 73. As a result of the threshold being lowered to 70%, the match rate in (2) exceeds the threshold (S48: YES), so that the CPU 71 determines that the recognition of "high thermal power" has been successful (S49). As a result, the stove 1 can reduce the stress of the user due to the fact that the stove 1 is not recognized even if the voice instruction is repeatedly input. The CPU 71 stores the information of "high thermal power" that has been successfully recognized in the RAM 73 as voice recognition information, and shifts to the process of S12 in FIG.

Returning to FIG. 5, the CPU 71 determines whether or not the voice instruction that has been successfully recognized is "end of voice operation" based on the voice recognition information stored in the RAM 73 (S12). When the voice instruction is "voice operation end" (S12: YES), the CPU 71 ends the voice operation and outputs a voice "voice operation is finished" from the speaker 77 (S13). Return to the process of S5. As a result, the user can recognize that the voice operation is completed.

Since the voice instruction that has been successfully recognized is "high thermal power" (S12: NO), as shown in FIG. 6, the CPU 71 determines whether or not there is only one burner currently being ignited (S20). In the state A shown in FIG. 8, since three burners (left stove 6, back stove 5, grill) are igniting (S20: NO), which of the three igniting burners is the received thermal power operation. It is unknown whether it is targeted at the burner of the place. Therefore, the CPU 71 needs to specify the location of the burner, which is the target of the thermal power operation, by voice instruction. Therefore, the CPU 71 identifies the burner currently being ignited and each burner thermal power, and temporarily stores them in the RAM 73 as the original thermal power information (S21).

In the state A, the firepower of the left stove 6 is 3, the firepower of the back stove 5 is 3, and the firepower of the grill is 8. Then, the CPU 71 drives the motor 32 of the gas amount adjusting mechanism 30 of each of the left stove 6, the back stove 5, and the grill that are being ignited, and drives the valve mechanism portion 31, so that the state B is the same as in the state B. The left stove 6, the back stove 5, and the burner firepower of the grill, which are in the process of being ignited, are all changed to 1 once, and the firepower is switched to the minimum (S22). As a result, the firepower is weakened for the time being until the location of the target burner on which the "high firepower" is executed is specified, so that it is possible to sufficiently cope with cases where it is necessary to immediately reduce the firepower.

Subsequently, the CPU 71 extracts a place where the previously accepted thermal power operation can be accepted based on the burner currently being ignited (S23). The places where the thermal power operation can be accepted are the back stove 5, the left stove 6, and the grill that are igniting. Therefore, in order to have the user specify the location of the thermal power operation by voice instruction, the CPU 71 creates a guide voice for notifying the choice of items that can be accepted by the stove 1 and outputs the guide voice from the speaker 77 (S24). As the guide voice, for example, the voice "Please select one from the left stove, the back stove, the grill, the fire extinguishing, and the end of the voice operation and give a voice instruction." Is output. At the same time, the CPU 71 displays the explanation screen 102 (see FIG. 10) on the display unit 22 (S25). The same content as the output guide voice is displayed on the explanation screen 102.

The user selects the location of the burner that performs the thermal power operation specified by the voice instruction from the locations notified by voice and display, and speaks to the stove 1. The user speaks to the stove 1 "left stove". The CPU 71 executes the voice recognition process (S26). The voice recognition process is the same as in FIG. In the voice recognition process, in the same manner as above, the voice instruction of the place where the thermal power operation is performed is received from the user, the matching rate is calculated by matching with the acoustic model data, and the calculated matching rate and the threshold value are used. Based on this, the success or failure of the recognition of the voice instruction is determined. Then, when the recognition of the "left stove" received from the user is successful (S48: YES, S49), the CPU 71 stores the information of the "left stove" that has been successfully recognized in the RAM 73 as voice recognition information (S50). ). The CPU 71 ends the voice recognition process (S51, S52: NO).

Returning to FIG. 6, the CPU 71 determines whether or not the recognized voice instruction is “end of voice operation” (S27). When the voice instruction is "voice operation end" (S27: YES), the CPU 71 proceeds to S5 in FIG. 5, ends the voice operation, and emits the voice "voice operation is finished" to the speaker 77. Output from (S34). The CPU 71 returns to the process of S5. As a result, the user can recognize that the voice operation is completed.

Since the voice instruction that has been successfully recognized is the "left stove" (S27: NO), the CPU 71 determines the left stove 6 as the target burner (S28). The target burner is the target burner for performing the thermal power operation received earlier. The CPU 71 creates an execution operation voice for changing the left stove 6 to a large thermal power, and outputs the execution operation voice from the speaker 77 (S29). The CPU 71 outputs, for example, an execution operation voice such as "change the left stove to a large thermal power." As a result, the user can recognize that the burner firepower of the left stove 6 is significantly changed.

After outputting the execution operation voice, the CPU 71 executes "high thermal power", which is a thermal power operation that has been successfully recognized, on the left stove 6 which is the target burner (S30). As a result, the burner firepower of the left stove 6 is changed from the firepower 1 to 9, as in the state C. Further, at the same time or slightly later than this, the CPU 71 returns the burner thermal power other than the target burner to the original thermal power by referring to the original thermal power information stored in the RAM 73 (S31). The burners other than the target burner are the back stove 5 and the grill. As a result, the burner firepower of the back stove 5 returns from 1 to 3 which is the original firepower, and the burner firepower of the grill returns from 1 to 8 which is the original firepower as in the state C. In this way, even if a plurality of burners are in the ignition state, only a specific burner can be operated correctly and safely by the voice instruction of the user. Further, in the present embodiment, since the execution operation voice is output before the thermal power operation is executed, the user can recognize in advance that the thermal power will be switched from now on, so that it is possible to prevent the thermal power from being switched unexpectedly.

Next, the CPU 71 determines whether or not there is a burning burner (S32). If there is an igniting burner (S32: YES), the CPU 71 returns to S8 in FIG. 5 and repeats the above process for the currently igniting burner. On the other hand, when all the burners of the stove 1 are extinguished (S32: NO), the CPU 71 ends the voice operation and outputs the voice of the end of the voice operation from the speaker 77 (S33). The CPU 71 returns to S2 in FIG. 5, monitors that the burner is ignited again, and repeats the above process.

In the process of S20 in FIG. 6, when there is only one burning burner (S20: YES), the target burner is one burner currently being ignited, so that the place where the thermal power operation is executed is determined. There is no need to specify. Therefore, if only the left stove 6 is ignited, the CPU 71 creates an execution operation voice having the content of switching the left stove 6 to a large thermal power, and outputs the execution operation voice from the speaker 77 (S34). Then, the CPU 71 executes a thermal power operation, which is a thermal power operation, on the left stove 6 which is being ignited (S35). The CPU 71 advances the process to S32 and repeats the above process.

As described above, the CPU 71 of the stove 1 of the present embodiment converts the voice instruction input from the outside into waveform data. The non-volatile memory 74 of the stove 1 stores acoustic model data showing selectable operation instructions as voice waveforms. The CPU 71 calculates the matching rate between the waveform data obtained by converting the voice instruction and the acoustic model data stored in the non-volatile memory 74. When the match rate exceeds the threshold value, it is determined that the voice instruction is recognized. As a result, the stove 1 can prevent the stove 1 from erroneously recognizing the noise input from the outside and operating. In such a stove 1, the threshold table 741 is stored in the non-volatile memory 74. In the threshold table 741, a threshold is set for each of a plurality of voice instructions. Among the plurality of voice instructions, a fire extinguishing instruction for instructing the fire extinguishing of the burner by voice is included. The threshold value for recognizing the fire extinguishing instruction is set lower than the threshold value for recognizing the voice instruction other than the fire extinguishing instruction. As a result, the CPU 71 can more reliably recognize the fire extinguishing instruction, which is more urgent than the other voice instructions, and can respond without loss in an emergency.

In the above description, the stove 1 is an example of the "cooking device" of the present invention. The right stove 4, the back stove 5, the left stove 6, and the grill are examples of the "heating source" of the present invention. In the voice recognition process of FIG. 7, the CPU 71 that executes the process of S42 is an example of the "conversion means" of the present invention. The non-volatile memory 74 is an example of the "storage means" of the present invention. The CPU 71 that executes the process of S44 is an example of the "calculation means" of the present invention. The CPU 71 that executes the processes of S45 and S48 is an example of the "determination means" of the present invention.

The present invention is not limited to the above embodiment, and various modifications can be made. In this embodiment, the built-in type stove 1 is illustrated, but it may be a table stove. The number of stoves may be one or more, and the grill may be omitted. Further, a cooking device other than the stove may be used, and for example, a cooking device having only a grill may be used. Further, the cooker is not limited to the gas cooker, and may be an electromagnetic cooker.

In the voice recognition process of the above embodiment, the criterion for determining the number of recognition failures k when lowering the threshold value is set to 3, but the criterion is not limited to this and can be freely changed.

In the voice recognition process of the above embodiment, when the number of recognition failures k is 4 or more (S57: YES), the CPU 71 weakens all the burner thermal power during ignition (S60). That is, the criterion for the number of recognition failures k when weakening the thermal power is set to 4 times, which is larger than the criterion for the number of recognition failures k when lowering the threshold value (3 times), but the number of recognition failures when lowering the threshold value. It may be 3 times, which is the same as the criterion of k. Further, the processing of S59 and S60 may be omitted.

In the voice recognition process of the above embodiment, when the number of recognition failures k is 4 or more (S57: YES), the CPU 71 re-notifies the voice instruction that can be input (S58). That is, the criterion for the number of recognition failures k when re-notifying is set to 4 times, which is larger than the criterion for the number of recognition failures k when lowering the threshold value (3 times), but the number of recognition failures when lowering the threshold value. It may be 3 times, which is the same as the criterion of k. Further, the process of S58 may be omitted.

In the voice recognition process of the above embodiment, when the recognition failure number k becomes 4 times or more (S57: YES), the CPU 71 re-notifies the inputtable voice instruction (S58), and then sets the recognition failure number k to 0. (S61), but it does not have to be initialized. Further, when the number of recognition failures k is further increased, for example, the voice operation may be forcibly terminated or the burner may be extinguished.

In S49 of the voice recognition process of the above embodiment, when the voice instruction is received and the recognition is completed, the voice for notifying that the voice has been received, for example, a chime sound of "pawn", is output from the speaker 77. It is good to set it to. As a result, it is possible to quickly recognize whether or not the voice instruction spoken by the user is recognized by the stove 1. If the chime does not sound even after speaking, it means that the voice recognition has failed, and the user can try to speak again.

In S60 of the voice recognition process and S22 of the voice operation control process of the above embodiment, the thermal power of the burning burner other than the target burner is uniformly weakened to the thermal power 1, but it is weaker than the current thermal power. It doesn't have to be weakened to the same firepower.

In the above embodiment, the opening degree of the valve of the valve mechanism portion 31 is adjusted by driving the motor 32 of the gas amount adjusting mechanism 30 corresponding to the burner, and the thermal power of the burner is adjusted. The structure may be any other than this. For example, the gas supply pipe may be provided with a plurality of bypass paths and a solenoid valve, and the solenoid valve may be opened and closed to adjust the amount of gas flowing through the burner.

In the above embodiment, the CPU 71 converts the voice instruction collected by the microphone 76 provided in the stove 1 into waveform data, and matches the waveform data with the acoustic model data stored in the non-volatile memory 74. So, for example, voice recognition is performed by receiving voice instructions with an external input device, converting the voice instructions into waveform data, and inputting them to the stove wirelessly or by wire in a digitized state. good.

For example, the stove is connected to a LAN or the Internet, and the voice instruction data received by an external input device and converted into waveform data is digitized and input to the stove via the LAN or the Internet. You may. Preferably, the stove and the external input device should be in a state of one-to-one communication. For example, a stove and an external input device can be communicated wirelessly, for example, by Bluetooth (registered trademark), and the voice instruction data received by the external input device and converted into waveform data is digitized and wireless. It is better to input to the stove with. Examples of the external input device include a mobile terminal, a smartphone, an input device such as a headphone equipped with a microphone, and the like.

1 stove 70 control circuit 71 CPU
74 Non-volatile memory 741 Threshold table

Claims (1)

Calculates the matching rate between the storage means for storing acoustic model data in which selectable operation instructions are indicated by voice information, the voice data of voice instructions input from the outside, and the sound model data stored in the storage means. In a heating cooker provided with a calculation means for determining that the voice instruction is recognized when the matching rate calculated by the calculation means exceeds a threshold value.
The first threshold value when the determination means recognizes the fire extinguishing instruction which is the voice instruction for instructing the fire extinguishing of the heating source by voice is lower than the second threshold value when recognizing the voice instruction other than the fire extinguishing instruction. A heating cooker that features that.

Images