JPH0141912B2 - - Google Patents

Abstract

PURPOSE:To improve the utility of controlling heating by voice input by providing indefinite speaker read only parameter memory which registers words relative to the fundamental operation of a heating device and a specific speaker rewritable parameter memory in voice recognizing means. CONSTITUTION:In order to enable the operation of a heating device with voice, voice recognizing means 14 is provided. This means 14 extracts and outputs by a feature parameter extractor 16 phonemic feature of a voice signal inputted from the microphone 7 of a voice input unit 15, and compares by a parameter comparator 17 it with feature parameter group of voice word registered in advance in undefinite speaker read only parameter memory 18 or specific speaker rewritable parameter memory 19. If the minimum value of the distance from each feature parameter group is lower than the threshold value, it is outputted as a recognized word to a controller 10, thereby controlling the heating, and if it, on the contrary, exceeds the threshold value, a reject signal or the like is outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a heating device using voice recognition means, and the registered parameter memory for voice recognition includes a read-only memory (ROM) for unspecified speakers and a rewritable memory (ROM) for specific speakers. The aim is to improve the usability of the heating device by using two types of RAM.

Conventionally, in heating devices that have digital control units, such as microwave ovens, as the number of cooking menus increases and cooking methods become more complex, the number of key switches on the operation panel also increases, making it difficult to make operational errors such as pressing the wrong button. I'm starting to have to put things away a lot.
Also, when my hands were occupied doing other tasks and I wanted to immediately shut off the microwave, it did not work as expected, which was inconvenient. Another option is to register a combination of heating output and heating time as a home menu, and then call it up whenever you want to cook. It was difficult to increase the number of users, and setting errors were also likely to occur.

The present invention uses voice recognition means to allow the user to easily operate the heating device using only his or her voice, and words related to the basic operation of the heating device are stored in advance in a read-only memory (ROM) as parameters for non-specific speakers. Anyone can use it, and if you want to use it in your own way, such as storing a home menu, you can program the heating procedure and then register the menu name in your own voice. This heating device allows you to call up the home menu as many times as you like using your voice.

An embodiment of the present invention will be described below with reference to FIGS. 1 and 2.

FIG. 1 is a perspective view of the main body of the heating device according to the present invention, and the front of the heating device main body 1 has a door 2 that can be opened and closed.
and an operation panel 3, on which a keyboard group 4, an output display group 5, and a time display section 6 are arranged. Further, a microphone 7 for voice input is connected to the main body 1. The user of the device should press "H", "MH", "M" on keyboard group 4,
Press any of the "ML", "L", or "W" keys to set the heating output, then press "10M" on keyboard group 4,
Press the "1M", "10S", and "1S" keys to set the heating time, and press the "STA" key to start heating. This combination of output and heating time can be set in two or more stages. The "STP" key can temporarily stop heating if it is in progress, and cancel all settings if it is in progress. Operations using these keys can be performed in exactly the same way by voice using the microphone 7, and moreover, the registered words are for unspecified speakers and can be used by anyone in the same way. In addition, if you want to memorize a home menu created by the user, after programming the heating time and output combination, press the "REC" key on the keyboard group 4 and register the menu name aloud from the microphone 7. The user's own voice is registered in the rewritable parameter memory (RAM) for a specific speaker, and heating time and output data are stored in the memory as home menu data.From now on, simply input the menu name from the microphone 7. You can call up programs from the home menu.

FIG. 2 is a block diagram according to the present invention. The heating source 8 is a magnetron in a microwave oven, and is controlled by an output control means 9 to achieve a predetermined output and heating time. The output control means 9 is controlled by a PWR signal issued from a controller 10.

The controller 10 receives a key strobe signal
A key input signal taken in from an input means 11 such as a keyboard is stored in a memory 12 by the KSTB. This data such as heating time or output is stored as home menu data in the memory 12, and at the same time, the menu name is also registered as voice data, and the data stored by inputting the menu name by voice is transferred from the memory 12. During the heating operation, the data is called again into the controller 10 and becomes basic data for timer control or output control. Further, the controller 10 outputs display data to display means 13 such as the output display group 5 and the time display unit 6 to perform predetermined display.

The speech recognition means 14 includes a speech input section 15 including a microphone 7, a feature parameter extraction section 16, a parameter comparison section 17, a read-only parameter memory 18 for unspecified speakers, and a rewritable parameter memory 19 for specific speakers. ing. The audio input unit 15 amplifies the audio signal input from the microphone 7, and the feature parameter extraction unit 16 extracts phonetic features included in the audio signal and converts it into digital data, such as a filter bank,
Frequency spectrum analysis using analog multiplexers and AD converters can be easily realized. The parameter comparison unit 17 compares a group of voice word feature parameters registered in advance in the read-only parameter memory 18 for unspecified speakers or the rewritable parameter memory 19 for specific speakers and the unknown speech input from the feature parameter extraction unit 16. Compare the feature parameters of and find the distance to each feature parameter group. If the minimum value of this distance is less than or equal to the threshold, the word corresponding to that parameter is output to the controller 10 as a recognized word, and the minimum value of the distance is If the value exceeds the threshold, it is assumed that no recognized word exists and no output is output, or a reject signal is output.

Read-only parameter memory 1 for non-specific speakers
Parameters for words related to basic operations such as "H" and "MH" in keyboard group 4 are registered in 8, and these parameters are for use by any speaker, meaning they can be recognized even if the voice is generated by anyone. Parameters with general versatility are selected. Parameters extracted from the speaker's own voice at the time of use by the feature parameter extraction unit 16 can be registered in the rewritable parameter memory 19 for a specific speaker.
You can re-register this as many times as you like.

The controller 10 also has a built-in memory 12.
This can be realized using a chip microcomputer, and the parameter comparing section 17 can also be easily realized using a microcomputer.

As described above in detail, according to the present invention,
The usability of the heating device is further improved by using voice recognition means, and the words related to the basic operations of the heating device are pre-registered in the read-only parameter memory for non-specific speakers, so anyone can easily operate the heating device without the need for registration. can be operated by voice. Furthermore, the rewritable parameter memory for a specific speaker can be re-registered with the user's own voice as many times as desired, allowing for unique usage such as the home menu. In particular, by having a speaker-specific rewritable parameter memory that can register word parameters related to the home menu, the heating program can be called up by the menu name itself.
There is no error in calling a program which is easy to make when calling a home menu program using symbols such as numbers. In addition, when a program is called by key operation, there is a restriction on the area that the key occupies within the panel, so there is no restriction on the number of home menus.In the case of this application, there is no such restriction and the number of home menus is limited. can be provided in large numbers depending on the size of the memory portion.

[Brief explanation of drawings]

FIG. 1 is a perspective view of the main body of the heating device according to the present invention;
FIG. 2 is a block diagram according to the present invention. 1...Main body, 7...Microphone, 8...Heating source, 10...Controller, 11...Input means,
12...Memory, 13...Display means, 14...Speech recognition means, 15...Speech input section, 16...Feature parameter extraction section, 17...Parameter comparison section,
18...Read-only parameter memory for unspecified speakers, 19...Writable parameter memory for specific speakers.

Claims (1)

[Claims] 1. A heating source for heating an object to be heated, a controller for controlling the heating source, an input means for inputting instructions regarding heating, and a memory capable of storing home menu data including the instructions and outputting it to the controller. , a display means for displaying the content of the instruction regarding heating, an audio input section including a microphone, a feature parameter extraction section for extracting features of the audio signal input to the audio input section;
A read-only parameter memory for non-specific speakers in which parameters of words that call up data related to basic operations are registered, a rewritable parameter memory for specific speakers that can register parameters of words that call up home menu data, and the feature parameter extraction unit. and a signal from the read-only parameter memory for unspecified speakers or the rewritable parameter memory for specific speakers, and outputs an operation signal to the controller by comparing the signals. Device.