JP2009042298A - Device controller by means of voice - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device controller by means of voice capable of correctly controlling other kinds of devices with relatively small number of vocabularies. <P>SOLUTION: In a device operation vocabulary storage part 12a, a device operation vocabulary which gives an instruction for every action of each device, first link information and a control content are registered. In a sensory vocabulary storage part 12b, a sensory vocabulary formed by verbalizing sensory information received on a sensory organ of a user and second link information are registered. The first link information and the second link information include medium information which correlates the sensory vocabulary with the device operation vocabulary, and the first link information includes additional condition information (location information, time information) for specifying an device 3 which becomes a control object from the device operation vocabulary of the same content as the medium information. A voice recognition processing part 10 extracts the device operation vocabulary which is correlated with the sensory vocabulary according with a vocabulary extracted from a voice input to a microphone 2 and the medium information, and specifies one device operation vocabulary by using the additional condition information. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a device control apparatus that uses voice to instruct the operation of a device.

  Conventionally, various control devices for instructing the operation of an illumination load or an air conditioner by voice have been proposed (see, for example, Patent Document 1 and Patent Document 2).

  The voice recognition type illumination switch described in Patent Document 1 solves the problem that the wall switch that is manually operated is difficult to find the wall switch when the room is dark, and the wall surface around the wall switch is soiled with hand dirt. For this purpose, “on”, “lighting”, “all on”, “partial on”, “off”, “off” are illustrated as vocabulary given by voice when controlling the operation of the lighting load.

On the other hand, the air conditioning control system described in Patent Document 2 recognizes emotions that the user feels about the indoor environment, and creates the indoor environment based on the recognized emotions. And the degree of emotion from the vocabulary that expresses the degree (“more up / up / down / lower”, “very hot / hot / cold / very cold”, etc.) A technique for instructing the air conditioner to adjust the air conditioner is described. Patent Document 2 also describes the use of emphasized words to determine the degree of emotion.
JP 2002-289371 A JP 2003-185221 A

  Since the technique described in Patent Document 1 basically gives instructions to a device that performs two types of operations, on and off, by accepting a vocabulary corresponding to on and off. Often, instructions relating to the degree of control (for example, the number of lamps lit) are distinguished by the vocabulary “all” and “part”. In addition, since the technique described in Patent Document 2 gives an instruction to the air conditioner device, it is necessary to instruct the direction and degree of adjustment. Therefore, the type of emotion is associated with the direction of adjustment, and the degree of emotion is adjusted. Corresponds to the degree.

  By the way, the technology described in Patent Literature 1 and Patent Literature 2 is basically a technology for controlling a device using a voice input from a microphone associated with the device for a single device. There is no consideration for the possibility of erroneous control in the presence of other devices. In general, there are various devices such as lighting loads, air conditioners, bath water heaters, etc., and there may be multiple devices of the same type in the home, even if they are limited to the devices used by users at home. If the operation of the device is instructed by voice, the user has to remember the vocabulary that distinguishes each device individually, and the number of vocabularies that the user must memorize for controlling the device increases. .

  If the number of vocabulary stored by the user is a few words, it can be used without a doubt, but as the number of vocabulary increases, confusion occurs in the storage of vocabulary that can be used for device control, and devices that do not want to be used There arises a problem that the device is controlled, the operation of the device becomes an unintended operation, or the device does not operate at all. Also, in a room with a plurality of devices, if a common vocabulary is used to control each device, there may be confusion that devices other than the target device operate.

  The present invention has been made in view of the above-described reasons, and its purpose is to increase the types of vocabulary that a user must memorize for device control while allowing control of a plurality of devices. An object of the present invention is to provide a voice-based device control apparatus that can suppress and control other types of devices with a relatively small number of words without error.

  According to the first aspect of the present invention, a voice input device to which a user's voice is input, a plurality of devices to be controlled, and a vocabulary extracted from the voice input to the voice input device match a vocabulary registered in advance. Then, the operation instruction device controls the device by a control signal according to the control content associated with the vocabulary, and the operation instruction device relates the device operation vocabulary that gives an instruction for each operation of each device to the control content. Registered device operation vocabulary storage unit, sensory vocabulary storage unit registered sensory vocabulary obtained by verbalizing sensory information received by the user's sensory organs, and a vocabulary extracted from speech input to the voice input device is a device operation vocabulary A speech recognition processing unit that outputs the control content related to the vocabulary when it matches one of the vocabulary and the sensory vocabulary, and the control content given from the speech recognition processing unit depends on the device to be controlled A device driving unit for converting the signal into a control signal, the first link information related to the device operation vocabulary is registered in the device operation vocabulary storage unit, and the sensory vocabulary storage unit is related to the sensory vocabulary Second link information is registered, and the first link information and the second link information include medium information that associates the sensory vocabulary with the device operation vocabulary, and includes the first link information and the second link information. At least the first link information includes additional condition information serving as a binding condition for specifying a device to be controlled from among device operation vocabularies having the same medium information, and the speech recognition processing unit is extracted from the speech When a sensory vocabulary that matches the vocabulary is extracted from the sensory vocabulary storage unit, the device operation vocabulary related to the sensory vocabulary is extracted from the media information. When there are multiple device operation vocabularies, the voice is input to the voice input device. Wherein the conditionals to adopt device operation vocabulary becomes prescribed relationship with the additional condition information when the.

  According to a second aspect of the present invention, in the first aspect of the invention, the medium information includes a physical quantity controlled by the device and a change direction of the physical quantity.

  According to a third aspect of the present invention, in the first or second aspect of the present invention, the voice input device is installed at a plurality of locations, and the operation instruction device has a location storage unit that registers a location where the voice input device is installed. The additional condition information is location information indicating a location where the device is used, and the speech recognition processing unit inputs the vocabulary when the vocabulary extracted from speech matches the registered sense vocabulary When the location of the voice input device is extracted from the location storage unit as the incidental condition, and the extracted location has a prescribed relationship with location information that is additional condition information of the device operation vocabulary associated with the sense vocabulary by the medium information The device operation vocabulary is adopted.

  According to a fourth aspect of the present invention, in any one of the first to third aspects of the present invention, the operation instruction device includes a clock unit for measuring a current time, and the additional condition information indicates a time when the device is used. The time recognition information, and the speech recognition processing unit detects the time when the vocabulary extracted from the speech matches the registered sense vocabulary when the vocabulary is input to the speech input device by the clock unit as the incidental condition. The device operation vocabulary is adopted when the detected time has a prescribed relationship with time information which is additional condition information of the device operation vocabulary associated with the sense vocabulary by the medium information.

  According to the configuration of the first aspect of the invention, the sensory vocabulary can be used separately from the device operation vocabulary associated with each device for each operation, so that one sensory vocabulary is associated with different operations of a plurality of devices. It becomes possible. In addition, sensory vocabulary is a vocabulary in which sensory information received by the user's sensory organs is verbalized, and is a vocabulary that does not need to be memorized. Therefore, it is more user friendly than storing various vocabulary by device and by operation. It is possible to prevent the device from not operating or a different device from operating due to erroneous storage or misuse of the device operation vocabulary. Since sensory vocabulary may be related to multiple device operation vocabularies, it may not be possible to specify a single device operation vocabulary using only sensory vocabulary, but link information for associating sensory vocabulary with device operation vocabulary may be used. Since not only the medium information for association but also the additional condition information is registered, a plurality of device operation vocabularies can be narrowed down by the additional condition information and specified as one device operation vocabulary. In other words, there is an advantage that many devices can be controlled by voice while using a small number of sensory vocabularies.

  According to the configuration of the second aspect of the invention, since the medium information included in the link information in order to relate the sense vocabulary to the device operation vocabulary includes the physical quantity controlled by the equipment and the change direction of the physical quantity. By looking at the contents of the medium information, it is possible to know for what purpose the vocabulary (device operation vocabulary and sensory vocabulary) having the medium information as link information controls the device. In other words, since the medium information has a meaning, compared to a case where the medium information is simply a symbol, it is possible to reduce the possibility of erroneous work in the work relating the device operation vocabulary and the sensory vocabulary.

  According to the configuration of the invention of claim 3, when the voice input device is installed in a plurality of places and a vocabulary instructing control of the device is input to the voice input device, the place where the vocabulary is input is the device When there is a prescribed relationship with the location information indicating the location where the device is used, it is possible to recognize that it is a control instruction for the device used at that location, and when using sensory vocabulary, the device is specified or changed. There is no need to specify a physical quantity, and it is possible to control a device intended by the user with a simple sense vocabulary.

  According to the configuration of the invention of claim 4, even when there are a plurality of devices having the same medium information at the same place, the time when the voice is input is related to the time information indicating the time when the device is used and the specified relationship. It is possible to recognize that this is a control instruction for the device used at that time. Therefore, it is possible to change which device and how to control it according to the time, and only one sensor vocabulary can be used to identify one device operation vocabulary for each time zone from multiple device operation vocabularies. It becomes possible to do.

  The embodiment described below exemplifies a case where the operation of a plurality of types of devices in the home is instructed by voice. Therefore, in this example, as shown in FIG. 1, an operation instruction for recognizing the content of the sound input from the microphone 2 as a sound input device and converting the instruction into an instruction related to the operation of the device 3 and instructing the operation of the device 3. A device 1 is provided. The operation instruction apparatus 1 is mainly composed of a computer.

  The device 3 is assumed to be a lighting load, an air conditioner, a floor heating device, a water heater, an electric blind, and the like. The operation instruction device 1 is instructed to turn on and off for the lighting load, on and off to the air conditioner, and to run and stop for the water heater. Adds cold water or starts reheating, gives instructions on the operation and stop of the floor heating device and bathroom dryer, and gives instructions for raising and lowering the electric blinds. That is, in this example, only the operation instruction equivalent to ON / OFF or the ON / OFF operation is given to the device 3, and the instruction of the operation for adjusting the lighting load such as the dimming control of the lighting load or the temperature adjustment of the air conditioner is in principle Not as a target.

  A microphone 2 incorporated in the device 3 can be used, but it is also arranged in various places in the house separately from the device 3. In this example, the location where the microphone 2 is disposed is not particularly considered. However, in the example described later, the information on the location where the microphone 2 is disposed is used together to narrow down the devices 3 to be controlled. Use.

  The microphone 2 is input to the operation instruction device 1 via the home network, and the operation instruction device 1 is connected to each device 3 via the home network. Although there are no particular restrictions on the transmission medium, transmission protocol, and topology of the home network, the illustrated example shows an example in which star connection is adopted as the topology. In addition, if the bus connection is adopted, wiring construction becomes easy. Here, since the microphone 2 and the device 3 are connected to the operation instruction device 1 via the home network, it is not necessary to provide the operation instruction device 1 for each device 3, and a single operation instruction device 1 is used to configure a plurality of devices. 3 can be instructed to operate, and can be instructed to control the device 3 by installing the microphone 2 at any place in the house.

  The operation instruction device 1 includes an audio signal input unit 11 to which an audio signal output from the microphone 2 is input. The voice signal input unit 11 removes noise other than human voice so that the vocabulary instructing the control of the device 3 can be easily extracted. Various techniques of this kind are known. As the simplest technique for separating noise other than human speech, there are known techniques of separation based on the difference in frequency components between speech and noise and separation based on the level of speech and noise. That is, by using a bandpass filter, frequency components other than the voice band are removed, and only the voice signal obtained during a period in which the signal level (average signal level) of the voice band is equal to or higher than a predetermined reference value is extracted. . Furthermore, the audio signal input unit 11 has a function of performing analog-digital conversion and a function of extracting frequency components using FFT (Fast Fourier Transform) or the like.

  The frequency component obtained by the voice signal input unit 11 is input to the voice recognition processing unit 10. The speech recognition processing unit 10 extracts phonemes by analyzing frequency components, and converts the phoneme sequence into text data. Further, the speech recognition processing unit 10 divides the text data into morphemes (word minimum units), and uses the morphemes to extract words such as words and collocations. In this example, complicated morphemes such as ending changes are not handled, and the vocabulary is extracted only when it matches the vocabulary given in the form of a reserved word. Therefore, semantic analysis is not performed, and if it is a collocation, the whole is treated as one vocabulary. However, it may be converted into a reserved word form by performing semantic analysis in speech recognition and recognizing the presence or absence of a particle or a ending change.

  In the semantic analysis, the speech recognition processing unit 10 collates with a registered vocabulary registered in the data storage unit 12 as a vocabulary related to the control of the device 3, and extracts only the registered vocabulary. The registered vocabulary registered in the data storage unit 12 includes a device operation vocabulary that indicates the type of the device 3 to be controlled and a state corresponding to the on / off state of the device 3, and a person related to the operation of the device 3. There is a sensory vocabulary that verbalizes sensory information received by the sensory organs. The device operation vocabulary and the sensory vocabulary are stored in different areas in the data storage unit 12. That is, the data storage unit 12 is provided with a device operation vocabulary storage unit 12a for storing device operation vocabulary and a sensory vocabulary storage unit 12b for storing sensory vocabulary.

  Here, only the sensory information whose contents change depending on the operation of the device 3 has a meaning, and the sensory vocabulary is a vocabulary having sensory information that can be associated with brightness, temperature, and the like. In other words, among the vocabulary expressing the senses perceived through visual and tactile senses (such as warm and cold points), the vocabulary associated with the control of the device 3 is treated as the sensory vocabulary.

  Examples of device operation vocabulary are “Show Me On” “Show Me Off” “Air Conditioner On” “Air Conditioner Off” “Blind Up” “Blind Down”. For example, “Atsui”, “Samui”, “Tsumai”, “Akaloui”, and “Cry” are examples of sensory vocabulary. The reason why each vocabulary is indicated by katakana is that the phoneme sequence extracted as text data is used as the vocabulary.

  If the text data alone has other conditions such as context, it may be possible to identify the meaning in the semantic analysis. For example, whether "hot" means "hot" or "hot" Cannot be determined without being bound by other conditions. Therefore, the katakana notation is used here in the meaning of the text data as the phoneme sequence before the meaning is determined. In other words, a plurality of vocabularies may be extracted as candidates for text data that is a sequence of phonemes. In this case, it is necessary to specify which vocabulary is based on other binding conditions. Furthermore, even if the vocabulary is specified from the candidates, if there are a plurality of devices 3 of the same type in the home, it may not be possible to specify which device 3 the instruction is directed to. In this example, as will be described later, the vocabulary is specified by giving a constraint condition to the vocabulary candidates, and the device 3 can be specified.

  By the way, as shown in Table 1, control vocabulary data that is a triple of (device operation vocabulary, control contents, link information) is registered in the device operation vocabulary storage unit 12a, and the sensory vocabulary storage unit 12b stores As shown in Table 2, sensory vocabulary data that is a pair of (sensory vocabulary, link information) is registered. “Control content” is an instruction related to the operation of the device 3. For example, instructions such as “lighting on”, “lighting off”, “air conditioner operation”, “air conditioner stop”, “blind up”, “blind down” are included in the control content. Become.

  The operation control vocabulary “show-me-on” and “show-me-off” are related to the control details for lighting load “lighting on” and “lighting off”, respectively, and the operation control vocabulary “blind up” and “blind down” are motorized respectively. The control contents for the blind are related to “blind rise” and “blind fall”. The control contents related to the device operation vocabulary “air conditioner on” and “air conditioner off” are for the air conditioner unit and change according to the season. In the summer, they are related to “air conditioner operation” and “air conditioner stop” respectively. If so, they are related to “air conditioner stop” and “air conditioner operation”, respectively. Since the judgment of the season varies depending on the region, the region can be set in addition to the calendar function (a function for managing the date) (not shown, which is configured with a dip switch or the like in the clock unit having the calendar function described later) It is desirable to switch between the cooling operation and the heating operation of the air conditioner according to the set region. In addition, you may exclude heating operation about the area which does not heat with an air-conditioner apparatus in winter.

  The speech recognition processing unit 10 extracts words and collocations using morphemes of text data converted from phonemes, and compares the words and collocations with control vocabulary data registered in the device operation vocabulary storage unit 12a. When the device operation vocabulary that matches the word / collocation is registered in the device operation vocabulary storage unit 12a, the control content of the control vocabulary data including the device operation vocabulary is acquired. That is, the device operation vocabulary is converted into control content. The acquired control content is converted into a control signal that enables the operation of the device 3 through the equipment driving unit 13, and the control signal is given to the device 3, whereby the device 3 performs an operation according to the control content. Conversion from a control content to a control signal is performed using a conversion data storage unit 12c (provided in the data storage unit 12) in which signal conversion data that is a pair of (control content, control signal) is registered.

  Incidentally, “link information” is stored in each of the device operation vocabulary storage unit 12a and the sensory vocabulary storage unit 12b, and at least a part of the link information is medium information indicating the purpose of the operation of the device 3, It is used to relate each device operation vocabulary in the device operation vocabulary storage unit 12a with each sensory vocabulary in the sensory vocabulary storage unit 12b.

  For example, when the device 3 is a lighting load, the device operation vocabulary has two types of “show-me-on” and “show-me-off”, and the operation contents are “lighting on” and “lighting off” as described above. The purpose of “illumination lighting” is to “increase the illuminance”, and the purpose of “illumination off” is “to decrease the illuminance”, so these purposes are used for the medium information.

  Further, as described above, when the device 3 is an air conditioner, if the season is summer, the device operation vocabulary “air conditioner on” is intended to be “reducing the room temperature”. Used for. Since “air conditioner off” is substantially “increasing the room temperature” in the summer, this is used for the medium information.

  When the device 3 is an electric blind, the purpose of “blind up” and “blind down” in the device operation vocabulary is “increase illuminance” and “decrease illuminance”, and the medium information is the case of the illumination load. It will be the same.

  The medium information is the purpose of the operation of the device 3 as described above. In other words, the medium information changes when the device 3 is operated (that is, controlled by the device 3). ) And the direction of change (increase / decrease) of the physical quantity. Meaning can be given to the medium information by using the items related to the operation of the device 3 for the medium information.

  Media information can use simple symbols (for example, A, B, C, etc.), and even if symbols are used, sensory vocabulary can be related to device operation vocabulary. Therefore, the possibility of erroneous work at the time of registering data in the device operation vocabulary storage unit 12a and the sensory vocabulary storage unit 12b increases. On the other hand, by using meaningful media information, if media information irrelevant to the device operation vocabulary or sensory vocabulary is registered, it can be found immediately and the possibility of erroneous data registration Can be reduced.

  As is clear from the above example, the device operation vocabulary is associated with the control content on a one-to-one basis, but the sensory vocabulary is not necessarily one-to-one with respect to the device operation vocabulary, and a plurality of devices for one sensory vocabulary Allows to associate operation vocabulary. The link information adds a binding condition to at least the device operation vocabulary of the device operation vocabulary and the sensory vocabulary.

  In order to associate the sensory vocabulary data registered in the sensory vocabulary storage unit 12b with the control vocabulary data registered in the device operation vocabulary storage unit 12a, medium information having the same contents is assigned to the control vocabulary data and the sensory vocabulary data. Give it. In the speech recognition processing unit 10, when the word / collocation word is checked against the control vocabulary data registered in the device operation vocabulary storage unit 12a, the device operation vocabulary that matches the word / collocation word is registered in the device operation vocabulary storage unit 12a. If not, the word / collocation word is collated with the sensory vocabulary in the sensory vocabulary storage unit 12b, and if a matching sensory vocabulary is found, the media information of the sensory vocabulary data is replaced with the media information in the device operation vocabulary storage unit 12a. Collate.

  The medium information of the sensory vocabulary data registered in the sensory vocabulary storage unit 12b is guaranteed to match any of the control vocabulary data registered in the device operation vocabulary storage unit 12a. Therefore, the device operation vocabulary related to the sensory vocabulary is always found. The operation after the device operation vocabulary is found is the same as the case where the word / collocation extracted from the text data matches the device operation vocabulary, and the control content related to the device operation vocabulary is controlled by the equipment driving unit 13. The control signal is given to the target device 3 and the device 3 is controlled.

  An example of the operation will be described. To operate the air conditioner device, the voice of the device operation vocabulary “air conditioner on” may be input to the microphone 2, and “air conditioner on” is the device operation vocabulary in the device operation vocabulary storage unit 12a. In the summer season, the control content “air conditioner operation” is extracted. That is, the air conditioner is turned on and the cooling operation is performed.

  When the sensory vocabulary “Atsui” is input to the microphone 2 instead of the device operation vocabulary “Air-conditioner on” under the same conditions, the device operation vocabulary that matches “Atsui” is extracted from the device operation vocabulary storage unit 12a. Therefore, the speech recognition processing unit 10 collates “hot” with the sensory vocabulary in the sensory vocabulary storage unit 12b. Here, there is a sensory vocabulary that matches “Atsui”, and the medium information of the sensory vocabulary data is “lower the room temperature”, so the speech recognition processing unit 10 has the medium information “lower the room temperature”. Since the control vocabulary data is retrieved from the device operation vocabulary storage unit 12a and the control vocabulary data of the device operation vocabulary “air conditioner ON” is extracted, the air conditioner apparatus is thereafter operated in the same manner as when the device operation vocabulary is input from the microphone 2. The cooling operation is performed.

  By the way, the device operation vocabulary is a combination of the name of the device 3 to be controlled (air conditioner, shomei, etc.) and the operation of the device 3 (on / off, etc.), so the target device is specified. can do. On the other hand, since the device 3 is not specified in the sensory vocabulary, for example, when the sensory vocabulary is “hot”, the purpose is “reducing the room temperature” or “reducing the hot water temperature” in the bathtub. Cannot be distinguished. In other words, it is impossible to distinguish whether “hot” in the sense vocabulary is “hot” or “hot”.

  Therefore, the link information has not only medium information for associating the sensory vocabulary described above with the device operation vocabulary but also additional condition information as a constraint condition. The additional condition information is information for distinguishing the device operation vocabulary having the same content as the medium information. By specifying the additional condition information, one of the device operation vocabulary having the same content as the medium information is specified and adopted. Can do.

  In other words, by giving additional condition information, it becomes possible to distinguish whether the instruction content is for room temperature or hot water temperature, and it is possible to distinguish between “hot” and “hot” simply by giving a sense vocabulary. It becomes possible to distinguish.

  As such additional condition information, at least one of location information indicating a location where the device 3 is used and time information indicating a time (time zone) when the device is used is used. Accordingly, at least one of the location and time when sound is input to the microphone 2 is used as an incidental condition, and it is determined whether the incidental condition has a prescribed relationship with the additional condition information, thereby corresponding to the sensory vocabulary. A device operation vocabulary is specified.

  First, a case where the additional condition information is location information will be described. In order to specify the location of the microphone 2 to which the vocabulary is input, a location storage unit 12d in which the installation location of each microphone 2 is registered is provided in the data storage unit 12, and the microphone 2 to which the device operation vocabulary or the sensory vocabulary is input is provided. The location of the microphone 2 having the highest sound pressure level is searched from the location storage unit 12d.

  Also, location information is included in the link information of the device operation vocabulary storage unit 12a. For example, the equipment operation vocabulary related to the air conditioner includes the location information of the indoor microphone 2 in which the air conditioner is installed in the link information, and the equipment operation vocabulary related to the water heater that supplies hot water to the bath includes the bathroom microphone. 2 location information is included in the link information. The speech recognition processing unit 10 collates the location extracted from the location storage unit 12d as an incidental condition with the location information of the link information registered in the device operation vocabulary storage unit 12a, and the location as the incidental condition is defined as the location information. The device 3 for which the operation is instructed is specified based on whether or not the relationship is satisfied (normally, the range of the location may be determined in the location information). It should be noted that in a plurality of devices 3, sets of device operation vocabulary and location information do not overlap.

  Now, it is assumed that two pieces of medium information “lower the room temperature” and “lower the hot water temperature” are registered in the sense vocabulary storage unit 12b for “hot”. Further, in the device operation vocabulary storage unit 12a, “living room” and “bedroom” are associated as location information with respect to the medium information “decrease room temperature”, and the location with respect to the media information “decrease hot water temperature”. It is assumed that “bathroom” is related as information.

  For example, when the voice of the sensory vocabulary “Atsui” is input to the microphone 2 installed in the bathroom, the sensory vocabulary “Atsui” is extracted from the sensory vocabulary storage unit 12b, and the sensory vocabulary “Atsui” is further extracted. The location “bathroom” is extracted from the location storage unit 12d based on the input position of the microphone 2. As described above, there are two types of medium information for the sensory vocabulary “Atsui”, “lower the room temperature” and “lower the hot water temperature”, and the device operation vocabulary storage unit 12a matches these medium information. There are two or more types of device operation vocabulary to be performed. These device operation vocabularies are called candidate vocabularies.

  If there is one type of candidate vocabulary, the candidate vocabulary may be handled as a device operation vocabulary. However, if there are a plurality of candidate vocabularies, the location extracted from the location storage unit 12d and the device operation vocabulary storage unit 12a Check stored location information. As described above, a single device operation vocabulary can be determined from the candidate vocabulary by specifying the location information for the candidate vocabulary under the condition that the pair of the device operation vocabulary and the location information does not overlap. .

  In the above example, there are two types of device operation vocabulary (candidate vocabulary) extracted by collating media information for the sensory vocabulary “Atsui”: “Air-conditioner on” and “Ray switch nuis”. Since the device operation vocabulary whose location information is “Bathroom” is “Ray Switch Menu”, “Ray Switch Menu” is extracted as a single device operation vocabulary. Therefore, the control content is specified as “hot water heater, cold water input”.

  There are several types of water heaters that operate differently. However, in a water heater that can be reheated, when the hot water temperature is raised, it is reheated for a certain period of time, and when the hot water temperature is lowered, a certain amount of cold water is used. Is generally used. In addition, when the cold water is supplied, if the hot water is being supplied or being reheated, the operation is stopped before the cold water is supplied. In this example, it is assumed that the media information “Increase hot water temperature” is related to the device operation vocabulary “Oidaki Kaishi”, and the media information “Reduce hot water temperature” is related to the device operation vocabulary “Ray switch screen”. . Further, in the water heater, since a series of operations when the hot water temperature is raised or lowered is defined in advance by the controller, for the water heater, an instruction given from the equipment driving unit 13 to the device 3 corresponds to the medium information. And it is sufficient.

  The above-described operation is also applied when the same type of device 3 is installed in a plurality of different rooms. For example, if air conditioners are installed in the “living room” and “bedroom”, the device operation vocabulary will be used not only when the sensory vocabulary “Atsui” is used but also when the device operation vocabulary “air conditioner on” is used. It is not possible to specify which air-conditioning device is instructed only by the use of this alone. In this example, since the location of the microphone 2 to which the device operation vocabulary (or sensory vocabulary) is input is used, it is possible to specify “living room” or “bedroom” by the location information.

  FIG. 2 shows the operation of the speech recognition processing unit 10 when the place information is used as the additional condition information. First, when speech is input from the microphone 2 and a vocabulary is extracted, it is checked against the device operation vocabulary registered in the device operation vocabulary storage unit 12a (S1). Here, if they do not match, the sensory vocabulary registered in the sensory vocabulary storage unit 12b is collated (S2). If the sensory vocabulary also does not match, it is not a vocabulary related to the control of the device 3, and the process ends. If there is something that matches the sensory vocabulary, the media information related to the same sensory vocabulary is extracted in order (S3, S4, S7). Further, every time the medium information is extracted, the device operation vocabulary storage unit 12a is collated (S5).

  Since the device operation vocabulary is associated with location information as additional condition information, the location obtained by collation with the location storage unit 12d is collated with the device operation vocabulary storage unit 12a, and the collation result has a specified relationship. The device operation vocabulary to become is extracted (S6). That is, the device operation vocabulary is specified from the candidate vocabulary. If the device operation vocabulary having the relationship specified by the collation result is found, it is not necessary to continue the collation work any more, and the process of extracting the medium information from the sense vocabulary storage unit 12b is ended.

  When the vocabulary input in step S1 matches the device operation vocabulary or the device operation vocabulary is specified in step S6, the equipment drive unit 13 controls the corresponding device 3 according to the control content of the obtained device operation vocabulary. A signal is output and the device 3 is controlled according to the control content (S8).

  Next, a case will be described in which time information indicating the time when the device operation vocabulary or sensory vocabulary is input to the microphone 2 is used as the additional condition information instead of the location of the microphone 2 where the device operation vocabulary or sensory vocabulary is input. The time information specifies the target device operation vocabulary from the candidate vocabulary based on, for example, the minimum unit of minutes, the time zone of the day, the day of the week, the season of the year, and the like. In order to manage time information in this way, instead of the location storage unit 12d shown in FIG. 1, a clock unit 14 that measures the current time is provided in the operation instruction device 1 as shown in FIG. 3. The clock unit 14 is used for specifying control contents in the voice recognition processing unit 10. In the following, an example of distinguishing the device operation vocabulary according to the time zone of the day will be shown.

  Now, it is assumed that there is an illumination load and a blind as the device 3 capable of controlling the illuminance. In the device operation vocabulary storage unit 12a, as shown in Table 3, the time information “16: 00-7: 00” is related to the device operation vocabulary “Show Mayon”, and the time information is stored in the device operation vocabulary “Blind Up”. It is assumed that “7: 00-16: 00” is related. In other words, the illuminance is adjusted by the blind during a time zone with a large amount of external light, and the illuminance is adjusted by an illumination load during a time zone with a small amount of external light. The sensory vocabulary storage unit 12b has the contents shown in Table 2.

  Here, if the sensory vocabulary “Cry” is input at “9:45”, the medium information “increase illuminance” related to the sensory vocabulary “Cry” is collated with the device operation vocabulary storage unit 12a, and two types of information are stored. The device operation vocabulary “show me on” and “blind up” are extracted as candidate vocabularies. Since the input time of the sense vocabulary “Cry” is “9:45”, the time information “7: 00-16: 00” is adopted, and “blind up” is specified as the device operation vocabulary. As described above, the time when the sensory vocabulary is input to the microphone 2 is used as an incidental condition, and this time is related to the time information associated with the device operation vocabulary (here, the time is within the range of the time information). It is possible to specify the device operation vocabulary by determining whether or not the above relationship is satisfied.

  FIG. 4 shows the operation of the speech recognition processing unit 10 when the time information is used as the additional condition information. First, when speech is input from the microphone 2 and a vocabulary is extracted, it is checked against the device operation vocabulary registered in the device operation vocabulary storage unit 12a (S1). Here, if they do not match, the sensory vocabulary registered in the sensory vocabulary storage unit 12b is collated (S2). If the sensory vocabulary also does not match, it is not a vocabulary related to the control of the device 3, and the process ends. If there is something that matches the sensory vocabulary, the time when the sensory vocabulary is input to the microphone 2 is acquired (S3). Further, the medium information in the link information of the sense vocabulary is collated with the device operation vocabulary storage unit 12a, and the acquired time in the device operation vocabulary having the same medium information indicates the relationship between the time information of the link information and the regulation. Search for a satisfying one (S4). If no search condition is found, it is not a vocabulary related to the control of the device 3, and the process of searching for a device operation vocabulary corresponding to the sense vocabulary is terminated.

  When the vocabulary input in step S1 matches the device operation vocabulary or the device operation vocabulary that satisfies the condition is specified in step S5, the equipment driving unit 13 sends the corresponding device 3 to the corresponding device 3 according to the control content of the obtained device operation vocabulary. On the other hand, a control signal is output, and the device 3 is controlled according to the control content (S6).

  In the above-described example, the time zone of one day is used as the constraint condition, but it is also possible to give the constraint condition in units of one week or season by giving the clock unit 14 a calendar function. For example, the device 3 is controlled by distinguishing between weekdays and Saturdays and Sundays, or the time zone using the lighting load and the time zone using the blind are changed according to the season as the device 3 for adjusting the illuminance. Is possible.

  Of course, location information and time information may be used together as additional condition information. For example, when there are a plurality of rooms having lighting loads and blinds as the device 3, it is possible to use location information to identify the rooms and use time information to select the lighting loads and blinds. It is.

It is a block diagram which shows embodiment. It is operation | movement explanatory drawing same as the above. It is a block diagram which shows other embodiment. It is operation | movement explanatory drawing same as the above.

Explanation of symbols

1 Operation instruction device 2 Microphone (voice input device)
3 Device 10 Voice recognition processing unit 12a Device operation vocabulary storage unit 12b Sensory vocabulary storage unit 12c Conversion data storage unit 12d Location storage unit 13 Equipment drive unit 14 Clock unit

Claims (4)

  If the vocabulary extracted from the speech input device to which the user's speech is input, the devices to be controlled, and the speech input to the speech input device match the vocabulary registered in advance, it is related to the vocabulary. A device operation vocabulary storage unit that registers a device operation vocabulary that gives an instruction for each operation of the device in association with the control content. A sensory vocabulary storage unit in which sensory vocabulary obtained by verbalizing sensory information received by the user's sensory organs is registered, and a vocabulary extracted from speech input to the speech input device is one of the device operation vocabulary and sensory vocabulary. A speech recognition processing unit that outputs the control content related to the vocabulary when they match, and the control content given from the speech recognition processing unit is converted into a control signal corresponding to the device to be controlled Second link information related to the sensory vocabulary is stored in the sensory vocabulary storage unit, and the sensory vocabulary storage unit stores first link information related to the device operation vocabulary. Are registered, and the first link information and the second link information include medium information relating the sensory vocabulary to the device operation vocabulary, and at least the first link information among the first link information and the second link information. The information includes additional condition information serving as a binding condition for identifying a device to be controlled from among device operation vocabularies having the same content as the medium information, and the speech recognition processing unit senses a sense that matches the vocabulary extracted from the speech. When the vocabulary is extracted from the sensory vocabulary storage unit, the device operation vocabulary associated with the sensory vocabulary is extracted from the media information, and when there are multiple device operation vocabularies, the incidental conditions when speech is input to the voice input device Device control apparatus according to sound, characterized by employing the additional condition information and device operation vocabulary a relation defined.   2. The device control apparatus using voice according to claim 1, wherein the medium information includes a physical quantity controlled by the device and a change direction of the physical quantity.   The voice input device is installed at a plurality of locations, the operation instruction device includes a location storage unit that registers a location where the voice input device is installed, and the additional condition information is location information indicating a location where the device is used. The voice recognition processing unit extracts the location of the voice input device that has input the vocabulary as a supplementary condition from the location storage unit when the vocabulary extracted from the speech matches the registered sense vocabulary, and extracts 2. The device operation vocabulary is adopted when the place is a prescribed relationship with the location information that is additional condition information of the device operation vocabulary associated with the sense vocabulary by the medium information. 2. A device control apparatus using voice according to item 2.   The operation instruction device includes a clock unit that counts the current time, the additional condition information is time information indicating a time when the device is used, and the speech recognition processing unit registers a vocabulary extracted from speech. The time when the vocabulary is input to the voice input device when it matches the sense vocabulary is detected by the clock unit as the incidental condition, and the additional condition of the device operation vocabulary in which the detected time is related to the sense vocabulary by the medium information 4. The device control apparatus using voice according to claim 1, wherein the device operation vocabulary is adopted when the information has a prescribed relationship with time information.

Images