JP2000029490A - Word dictionary data building method for voice recognition apparatus, voice recognition apparatus, and navigation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To embody the dictionary memory for recognition words necessary for voice input to a voice recognition apparatus for instructing and inputting one kind of command while permitting the substantially the same and plural input forms to this voice input. SOLUTION: To give a command for enlarging a map, a user is able to give the same command even with any pattern of, for example, 'please enlarge the map', 'please make the map larger', 'zoom the map', 'please enlarge' or 'zoom'. Namely, the specification corresponding to the same command of 'enlarge the map' is made possible even if the ways of speaking are mutually different. The user is thus able to more easily use the apparatus. There is nevertheless no need for storing the plural ways of speaking as discretely corresponding word dictionary data and, for example, the building of the common words, such as, for example, 'the map', 'enlarge', 'zoom' or 'please' as the common word dictionary date is possible.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech recognition apparatus which is effective for, for example, enabling a destination setting or the like in a navigation system to be input by speech, and a navigation system having the speech recognition apparatus.

[0002]

2. Description of the Related Art Conventionally, there has been proposed a speech recognition apparatus which compares an inputted speech with a plurality of pattern candidates to be compared which are stored in advance and determines a speech having a high degree of coincidence as a recognition result. It has already been put into practical use, and is used, for example, by a user to input a predetermined command by voice such as an instruction to enlarge or reduce a map displayed in a navigation system, a setting of a route, and the like. In particular, when the driver uses the in-vehicle navigation system, voice input does not involve button operation or screen gaze, so that it is effective because the safety is high even when the vehicle is running.

For example, consider a case where a voice input of "Please enlarge the map" is required in order to instruct and input the above-described command of map enlargement. In this case, it is troublesome for the user that the pronunciation must be divided into predetermined units based on words, such as "map", "wo", "enlarge", "do", "shi". So
It is preferable that input can be performed continuously (collective input). However, in the case where the configuration corresponding to the batch input is used, a situation in which the usability is degraded for the user may be assumed. That is, if the user does not make a specific voice input of “enlarge the map” correctly, it may not be recognized as the input of a desired command.

Therefore, it is conceivable to allow a plurality of voice inputs indicating substantially the same command. For example, in order to input a command to enlarge the map, in addition to the recognition word "Enlarge the map" described above, "Enlarge the map", "Zoom the map", "Map "Enlarge", "enlarge the map",
Many variations are conceivable by paraphrasing or omitting them, such as “zoom”.

However, for one type of command,
If it is necessary to prepare all of these many variations as recognition words, as the types of commands increase, a large amount of dictionary memory for recognition words is required. SUMMARY OF THE INVENTION The present invention solves such a problem, and allows substantially the same plurality of input modes for a voice input to a voice recognition device for instructing and inputting one type of command, while performing the necessary recognition. An object of the present invention is to realize a dictionary memory for words with a smaller capacity.

[0006]

Means for Solving the Problems and Effects of the Invention The word dictionary data construction method for a speech recognition device according to the first aspect of the present invention, which has been made to achieve the above object, is used in the following speech recognition device. . That is, the comparison target pattern corresponding to each word is stored as word dictionary data,
For example, a voice input through a microphone or the like is compared with a pattern to be compared in the word dictionary data, and a result having a high degree of matching is regarded as a recognition result, and a plurality of recognition results having substantially the same content correspond to the same command. This is a speech recognition device that specifies a command to be executed and outputs the specified command to an external device. Then, word dictionary data is constructed according to the following procedures (1) to (3).

(1) First, for each command, a word specifying a command object indicating a command object or a target function, a word "wo" as a target particle, and a word specifying a command operation indicating a command operation content Then, a standard comparison target pattern configured by sequentially connecting verbs added linguistically after the word specifying the command operation is set. (2) Next, a pattern that satisfies at least one or more of the following conditions with respect to a standard comparison target pattern, and a pattern that satisfies any of the alternative conditions having substantially the same contents are Set as an alternative comparison pattern indicating the same content.

[0008] Synonyms are used for words that specify a command target or a command action. The inflected form is used for verbs added in a language custom after a word specifying a command action.

[0009] The object case particle "o" is omitted. The verb added in the language custom after the word that can specify the command operation is omitted. When the verb added in the language custom after the word that can specify the command operation is “Please”, only “Please” is omitted. (3) If the standard comparison target pattern and the alternative comparison target pattern have the same syllable data, they are based on a tree structure allocated to have the same parent, but the number of paths reaching the vertex is In order to correspond to a directed graph format similar to a tree structure which is not necessarily 1, syllable data constituting the pattern to be compared and identification data indicating the end of a word are assigned to respective vertices in accordance with a preceding sequential scan.

The construction method described in the above (1) to (3) will be further described with reference to specific examples. Here, "map enlargement" is taken as an example of the command. As an example of the standard comparison target pattern set in the above (1), "enlarge the map" can be considered. In other words, the word specifying the command or the command target indicating the target function of the command “map enlargement” is “map”, and the word specifying the command operation indicating the operation content of the command is “magnify”. As a verb added in the language custom after the word “enlargement” that specifies the command action, a compound verb “Please” is given. Therefore, add the case particle "wo" and "enlarge the map"
Is set as a standard comparison target pattern.

In the above (2), an alternative comparison pattern showing substantially the same content as the standard comparison pattern "Enlarge the map" is set. As a condition, “map” is considered as a synonym for the word “map” that specifies the command target. Also,
As a synonym for the word “enlarge” that specifies the command operation, “large”, “zoom”, and the like can be considered.

As a condition, when "please" is added after "expansion", the verb "do" is a continuous form of "do"
Is used, and its utilization form may be the end form of "do" or the imperative form of "shiro". Since the condition "o" is omitted, "enlarge the map" corresponds to "enlarge the map".

The condition is "please" after "enlarge"
Are omitted. Therefore, for example, “enlarge the map” corresponds to “enlarge the map”. As for the condition, only "please" of "please" is omitted. Therefore, for example, "enlarge the map" corresponds to "enlarge the map". Naturally, the case where two or more of the following conditions are satisfied is also applicable. Therefore, for example, when both of the conditions for omitting the omission are satisfied, only “map enlargement” is performed for “enlarge map”.

When the standard comparison target pattern and the alternative comparison target pattern are set as described above, (3)
Constructs word dictionary data. In other words, when having the same syllable data, the tree structure is allocated so as to have the same parent, but the number of paths reaching the vertices is not necessarily one. The syllable data constituting the target pattern and the identification data indicating the end of the word are assigned to each vertex according to the preceding forward scanning.

The "directed graph format similar to a tree structure" basically conforms to a tree structure. However, the number of paths reaching a vertex is not necessarily one but may be plural. After that, they are allowed to "merge", so they will be called like this.

By providing syllable data assigned to each vertex and identification data indicating the end of a word, word data is obtained by connecting the syllable data in the order of parent and child in a tree structure. For example, the standard comparison target pattern is set to the above-mentioned “enlarge the map” and word dictionary data including all the alternative comparison target patterns is constructed as shown in FIG.

By constructing word dictionary data in this way,
In a speech recognition device using this word dictionary data, for example, "enlarge the map", "enlarge the map", "zoom the map", "enlarge the map", "map The wording of words such as "Large" and "Zoom" is different, but it can be specified as corresponding to the same command of "Map enlargement" for voice input with words that indicate substantially the same content, making it easier for users to use. Get better. Nevertheless, there is no need to store these multiple phrases as individually corresponding word dictionary data, such as "map", "enlarge", "large", "zoom" or "please". Since words can be constructed as shared word dictionary data, a memory for word dictionary data can be realized with a smaller capacity.

Under the condition (2) that is permitted to be set as an alternative comparison pattern having substantially the same contents as the standard comparison pattern, the command may be, for example, “map enlargement”. In the case of
As for the command "map" and the command operation "magnify", the use of synonyms was only permitted, and the omission was not a condition. This is because if the object and operation of the command are not known, the command is meaningless.

However, the command object may be omitted if it can be specified by any method. That is, in a situation where a map is specified as a target of a command, if only "enlargement" is input, it can be understood that it is map enlargement. Therefore, as set forth in claim 2, at least one or more of the six conditions including the condition in addition to the above conditions, that is, the condition that "the command object is omitted" is satisfied. The pattern may be set as an alternative pattern to be compared that shows substantially the same contents as the standard pattern to be compared.

This holds in a situation where, for example, in a mode in which some operation is performed on the map at the output destination of the command, the command target itself can be specified and may be omitted. . Further, the invention according to claim 3 provides a voice input means for inputting voice, and a voice input via the voice input means, with a plurality of comparison target pattern candidates previously stored in the dictionary means. Recognition means for recognizing a result having a high degree of coincidence as a recognition result, and specifying means for specifying a plurality of recognition results having substantially the same content as those corresponding to the same command based on the recognition result by the recognition means. Command output means for outputting the command specified by the specifying means,
Wherein the word dictionary data stored in the dictionary means is constructed by the word dictionary data construction method according to claim 1.

According to the speech recognition apparatus, when a user inputs a speech through the speech input unit, the recognition unit compares the input speech with a plurality of comparison target pattern candidates stored in the dictionary unit in advance. The result of the comparison is regarded as a recognition result, and the specifying means specifies a plurality of recognition results showing substantially the same content as corresponding to the same command, and the command output means specifies the specified command. Output. The word dictionary data stored in the dictionary means is constructed by the above-described word dictionary data construction method. The effect of this case is the same as the effect described above when the invention is considered as the invention of the word dictionary data construction method. Therefore, although not repeated in detail here, it is necessary to realize the dictionary memory with a smaller capacity. This is advantageous in miniaturization of the speech recognition device.

Also, as described above, if the command target can be specified by any method, the command target omitted may be set as an alternative comparison target pattern having substantially the same contents. As a voice recognition device corresponding thereto, a configuration as described in claim 3 can be considered.
In other words, there is provided command target input means for inputting information about a command target that can be omitted from a command output destination, and word dictionary data to be stored in the dictionary means is constructed by the word dictionary construction method according to claim 2. A speech recognition apparatus characterized in that it is a speech recognition apparatus.

In the case where the voice recognition device according to the third aspect is used for a navigation system, it is conceivable that the voice recognition device is configured as shown in the fifth aspect. In other words, the voice recognition device according to claim 3 and a navigation device are provided, and the voice input unit of the voice recognition device is a predetermined voice that needs to be specified at least when the navigation device performs a navigation process. A command used by a user to input an instruction of a navigation processing-related command by voice, wherein the command output unit is configured to output the specified command to the navigation device. It is a navigation system.

In this case, the "predetermined navigation processing-related command" includes a command for instructing enlargement, reduction or movement of the displayed map, setting of a route, setting of a destination, and the like. In this case, for example, if the command is "map reduction", variations such as "please reduce the map" to "please reduce the map" can be considered, and if the command is "map movement" For example, a variation such as “map move” for “please move the map” is conceivable. If the command is “route setting”, variations such as “route setting” to “please set the route” are conceivable.

On the other hand, when the voice recognition device according to the fourth aspect is used for a navigation system, it is conceivable to configure as shown in the sixth aspect. In other words, the apparatus includes the voice recognition device according to claim 4 and a navigation device, and the navigation device outputs information on an omissible command target to the voice recognition device based on the currently executed navigation process. Command output means, the voice input means of the voice recognition device,
At least the navigation device is used for a user to input an instruction of a predetermined navigation process related command that needs to be specified in performing the navigation process by voice, and the command output unit is A navigation system configured to output the specified command to the navigation device.

The output destination of the command is a navigation device, and the navigation device performs a navigation process by displaying a map. For example, the mode is such that the map can be enlarged, reduced, or moved down. If so, "map" is output to the speech recognition device as an optional command object. In this case, even if there is a voice input of "enlargement" without specifying the "map" as the command target, it is possible to cope with the input.

The application destination of the speech recognition apparatus is not limited to the above-described navigation system. For example, when the voice recognition device is used for an air conditioning system, it is conceivable to adjust the set temperature, select an air conditioning mode (cooling / heating / dry), or select a wind direction mode by voice input.

For example, as for the set temperature, the phrase "set the set temperature to 25 degrees" or "set temperature 25 degrees" or "set the temperature to 25 degrees" or "set the temperature to 25 degrees" is used. It is preferable to allow the phrase "please", and even simply the phrase "25 degrees". The same applies to the air conditioning mode and the wind direction mode.

The above navigation system and air conditioning system are not limited to the case where they are used as on-vehicle equipment, but may be, for example, a portable navigation device or an indoor air conditioner. However, when it is used for an in-vehicle device, downsizing of the voice recognition device is more effective when mounted on a vehicle. Of course, from such a viewpoint, the present invention can be similarly applied to in-vehicle devices other than the navigation system and the air conditioning system. For example,
Car audio equipment is effective. Also, considering a configuration in which opening and closing of the power window and adjustment of the mirror angle are instructed by voice, it is effective even in such a situation.

Although it has been described that there is an advantage peculiar to the in-vehicle device, the voice recognition device of the present invention is applied to a device which executes a predetermined process in accordance with a voice input instruction from a user. If it does, the same can be considered. For example, the present invention can be similarly applied to a portable information terminal device or an information terminal device set in a street or a parking area.

[0031]

Embodiments of the present invention will be described below with reference to the drawings. It is needless to say that the embodiments of the present invention are not limited to the following examples, and can take various forms as long as they belong to the technical scope of the present invention.

FIG. 1 shows a speech recognition apparatus 3 according to an embodiment of the present invention.
FIG. 1 is a block diagram illustrating a schematic configuration of a car navigation system 2 to which 0 is applied. The car navigation system 2 includes a position detector 4, a map data input device 6, an operation switch group 8, a control circuit 10 connected thereto, an external memory 12 connected to the control circuit 10, a display device 14, and a remote control sensor 15. And a voice recognition device 30.
The control circuit 10 is configured as a normal computer, and includes a well-known CPU, ROM, RAM,
/ O and bus lines for connecting these components.

The position detector 4 includes the gyroscope 1
8. GPS (Global Positioning System) that detects the position of the vehicle based on radio waves from the distance sensor 20 and satellites
m) has a GPS receiver 22. Each of the sensors 18, 20, and 22 has an error having a different property, and is configured to be used while being interpolated by a plurality of sensors. It should be noted that depending on the accuracy, a part of the above-described components may be used, and a rotation sensor for the steering wheel, a wheel sensor for each rolling wheel, or the like may be used.

The map data input device 6 is a device for inputting various data including so-called map matching data, map data and landmark data for improving the accuracy of position detection. As a medium, CD-R
Although OM is generally used, another medium such as a digital video disk (DVD) or a memory card may be used.

The display device 14 is a color display device. On the screen of the display device 14, the vehicle current position mark input from the position detector 4, the map data input from the map data input device 6, and the map data It is possible to superimpose and display additional data such as a guidance route displayed above and a mark of a set point described later.

The present car navigation system 2
Is input from the remote control sensor 15 via a remote control terminal (hereinafter referred to as a remote control) 15a or the position of the destination by the operation switches 8;
It also has a so-called route guidance function that automatically selects an optimal route from the current position to the destination and forms and displays a guidance route. As a technique for automatically setting the optimum route, a technique such as the Dijkstra method is known.
The operation switch group 8 is, for example, a touch switch integrated with the display device 14 or a mechanical switch, and is used for various inputs.

The voice recognition device 30 is used when the operation switch group 8 or the remote controller 15a is used for manually instructing a destination or the like. This is a device that allows the user to specify a destination or the like.

The speech recognition device 30 includes a speech recognition unit 31 as “recognition means”, a dialog control unit 32 as “command output means”, a speech synthesis unit 33, and a speech input unit 3
4, a microphone 35 as "audio input means", and a PTT
A (Push-To-Talk) switch 36 and a speaker 37 are provided.

The voice recognition unit 31 performs a voice recognition process on the voice data input from the voice input unit 34 in accordance with an instruction from the dialog control unit 32, and returns the recognition result to the dialog control unit 32. That is, the voice data acquired from the voice input unit 34 is collated using the stored dictionary data, and a higher-level comparison target pattern having a higher degree of matching as compared with a plurality of comparison target pattern candidates is determined. Output to The recognition of the word sequence in the input speech is performed as follows. That is, the audio data input from the audio input unit 34 is sequentially subjected to acoustic analysis to extract acoustic features such as cepstrum coefficients calculated by LPC analysis, for example.
The acoustic feature time series data obtained by this acoustic analysis is obtained. And the well-known DP matching method, HMM
(Hidden Markov model) or a neural network, this time series data is divided into several sections,
It is determined which word each section corresponds to stored in the dictionary data.

The dialogue control unit 32 sends, for example, a navigation instruction to the control circuit 10 which issues an instruction to generate a response voice to the voice synthesis unit 33 and executes processing of the system itself, based on the recognition result and the internal state managed by itself. When the map is displayed in the gate process, a process for instructing to enlarge or reduce the map is executed. As a result, if the voice recognition device 30 is used, it is possible to instruct a destination to the navigation device by voice input without manually operating the operation switch group 8 or the remote controller 15a.

The voice input unit 34 converts the surrounding voice captured by the microphone 35 into digital data and outputs the digital data to the voice recognition unit 31. In this embodiment,
This is a usage method in which a user inputs a voice through the microphone 35 while pressing the PTT switch 36. Specifically, the voice input unit 34 determines whether or not the PTT switch 36 has been pressed. When the PTT switch 36 has been pressed, the voice input unit 34 executes voice input processing via the microphone 35. If not, the voice input process is not executed. Therefore, only the voice data input via the microphone 35 while the PTT switch 36 is pressed is output to the voice recognition unit 31.

Here, the voice recognition unit 31 and the dialogue control unit 32
Will be described in more detail with reference to FIG.
As shown in FIG. 2, the voice recognition unit 31 includes a collating unit 31a as a “recognition unit” and a dictionary unit 31 as a “dictionary unit”.
b, and the dialog control unit 32 includes a post-processing unit 32
a, a communication control unit 32b and a dictionary control unit 32c.

In the voice recognition section 31, a collation section 31a
Performs collation on the voice data obtained from the voice input unit 34 using the word dictionary data stored in the dictionary unit 31b, and recognizes a word sequence in the input voice. Specifically, audio data input from the audio input unit 34 is sequentially subjected to acoustic analysis to extract an acoustic feature (for example, cepstrum), and to obtain acoustic feature time-series data obtained by the acoustic analysis. Then, the time series data is divided into several sections by a well-known DP matching method, for example, and a process as “recognition means” for determining which word stored in the dictionary unit 31b corresponds to each section is performed. Execute.

The recognition result of the collating unit 31a is transmitted to the dialogue control unit 3
2 is sent to the post-processing unit 32a. The post-processing unit 32a transmits the data to the control circuit 10 via the communication control unit 32b, for example, to execute a post-determination process instructing to perform a predetermined process, or transmits the voice data to the voice synthesis unit 33. A process for instructing to sound is executed.

The dictionary control unit 32c of the dialog control unit 32
Is provided with a dictionary description file 321 containing a dictionary description described in a dictionary definition format corresponding to each command used in the navigation system 2. Construct word dictionary data based on the description file. The constructed word dictionary data is stored in the dictionary unit 31b, and is used for the above-described speech recognition.

Here, a method in which the word dictionary construction section 322 constructs word dictionary data based on the dictionary description file in the dictionary description file 321 will be described in detail. In order to facilitate understanding, here, a description will be given taking “map enlargement” as an example of the command. (1) First, a phrase that is assumed that the user inputs the command by voice is set as a standard comparison target pattern. In detail, a word specifying a command target indicating a target object or a target function of a command, a particle of an object case “を”, a word specifying a command operation indicating the operation content of a command, a word specifying the command operation It is constructed by sequentially connecting verbs added later in language custom. As a specific example of the command “enlarge map”, “enlarge map” can be considered.

That is, the word specifying the command or the command target indicating the target function of the command “map enlargement” is “map”, and the word specifying the command operation indicating the operation content of the command is “magnify”. is there. As a verb added in the language custom after the word “enlargement” that specifies the command action, a compound verb “Please” is given. Therefore, "Enlarge the map" by adding the case particle "" is set as a standard comparison target pattern. (2) Then, a pattern that satisfies at least one or more of the following conditions with respect to the standard comparison pattern “enlarge the map” is used as an alternative comparison pattern having substantially the same contents. Set.

A synonym is used for a word that specifies a command target or a command operation. The inflected form is used for verbs added in a language custom after a word specifying a command action.

The object case particle "" is omitted. The verb added in the language custom after the word that can specify the command operation is omitted. When the verb added in the language custom after the word that can specify the command operation is “Please”, only “Please” is omitted.

The command target is omitted. The above standard comparison pattern "Enlarge the map"
Consider each condition for. The condition may be, for example, "large" or "zoom" as a synonym for the word "enlarge" that specifies the command operation.
Also, as a condition, please do after "enlarge"
Is added, the verb "Suru" is used in conjunction with "Shi".
And an imperative form of "shiro". Further, since the condition "o" is omitted, "enlarge the map" corresponds to "enlarge the map". And
As for the condition, "Please" after "Enlargement" is omitted. Therefore, for example, “enlarge the map” corresponds to “enlarge the map”. The condition is
Only "Please" of "Please" is omitted. Therefore, for example, “enlarge the map” corresponds to “enlarge the map”. Further, since the condition is omitted from the command target, "Please enlarge" is applicable.
It should be noted that the case where two or more of the conditions (1) to (4) are satisfied naturally applies. Therefore, for example, when both of the conditions for specifying the omission are satisfied, only “map enlargement” is performed for “enlarge map”.

As described above, when the above-mentioned standard comparison target pattern “enlarge the map” is specifically exemplified as an alternative comparison target pattern satisfying at least one or more of the following conditions, the following (a) ) To (t). (A) Enlarge map (b) Enlarge map (c) Enlarge map (d) Enlarge map (e) Enlarge map (f) Enlarge map (g) Map (H) Zoom map (i) Zoom map (j) Zoom map (k) Map zoom (l) Zoom (m) Zoom (n) Zoom (o ) Increase (p) Increase (q) Increase (r) Zoom (s) Zoom (t) Zoom In this way, the standard comparison pattern "Enlarge the map" On the other hand, 20 alternative comparison target patterns (a) to (t) are set, and these 21 word dictionary data correspond to one command “map enlargement”. The dictionary description file to be described collectively is [Map [ (Enlargement | large | zoom) [to [please] becomes].

In this specific example, () in (enlargement | large | zoom) indicates a priority order, and |
Indicates an option. In other words, (Enlarge | Large | Zoom) means selecting one of "Enlarge", "Large" or "Zoom", "Large" has priority over "Zoom", and "Large" has priority over "Large". "Expansion" has priority. [] Indicates omission possible, and [Map []]
In such a case, it is possible to omit only "O" to obtain a "map" or to omit "map" entirely. (3) As described above, in this specific example, the standard comparison target pattern “Please enlarge the map” is the same as (a) to (t).
21 comparison target patterns including zero alternative comparison target patterns are set. Instead of setting these as 21 word dictionary data as they are, word dictionary data is constructed as follows. . In other words, when having the same syllable data, the tree structure is allocated so as to have the same parent, but the number of paths reaching the vertices is not necessarily one. The syllable data constituting the target pattern and the identification data indicating the end of the word are assigned to each vertex according to the preceding forward scanning.

Here, "directed graph format similar to tree structure"
This basically conforms to the tree structure, but the number of paths reaching the vertices is not necessarily one, but may be plural. Will be called as follows. By providing syllable data assigned to each vertex and identification data indicating the end of a word, if the syllable data is connected in the order of parent and child in a tree structure, it becomes word data.

For example, assuming that the standard pattern to be compared is “enlarge the map” described above, word dictionary data having all the alternative patterns to be compared is constructed as shown in FIG. . By constructing the word dictionary data in this way, a speech recognition device using the word dictionary data can, for example, "enlarge the map", "enlarge the map", or "zoom the map". ,
Words such as "enlarge the map", "enlarge the map", and "zoom" have different words, but correspond to the same command of "enlarge the map" for speech input with words that have substantially the same content. It can be specified as a user, and the usability is improved for the user. Nevertheless, there is no need to store these multiple phrases as individually corresponding word dictionary data, such as "map", "enlarge", "large", "zoom" or "please". Since words can be constructed as shared word dictionary data, a memory for word dictionary data can be realized with a smaller capacity.

By the way, under the condition (2) that is permitted to be set as an alternative comparison pattern having substantially the same content as the standard comparison pattern, the command may be, for example, “map enlargement”. In the case of
As for the command "map" and the command operation "magnify", the use of synonyms was only permitted, and the omission was not a condition. This is because if the object and operation of the command are not known, the command is meaningless.

However, the command object may be omitted if it can be specified by any method. That is, in a situation where a map is specified as a target of a command, if only "enlargement" is input, it can be understood that it is map enlargement. Therefore, at least one of the six conditions obtained by adding the condition in addition to the above conditions, that is, the condition that “the command object is omitted” is added.
A pattern that satisfies one or more conditions may be set as an alternative pattern to be compared having substantially the same content as the standard pattern to be compared.

This is established in a situation where, for example, in a mode in which some operation is performed on the map at the output destination of the command, the command target itself can be specified and can be omitted. . Dictionary part 3
The explanation of 1b ends here.

In the voice recognition section 31, a collation section 31a
Is a post-processing unit of the dialogue control unit 32, which compares the voice data acquired from the voice input unit 34 with a higher comparison pattern having a higher degree of coincidence with a plurality of comparison target pattern candidates stored in the dictionary unit 31b. 32a. Then, the post-processing unit 32a stores the upper comparison target pattern and, for example, sends data to the control circuit 10 and instructs to perform a predetermined process when the above-mentioned predetermined determination instruction is issued. Or a process of instructing the voice synthesizing unit 33 to transmit voice data to generate sound. In this case, the data to be sent to the control circuit 10 may be all of the upper comparison target patterns as the final recognition result, or only the uppermost one of them. As described above, according to the car navigation system 2 of the present embodiment, when the command to enlarge the map is instructed, for example, in addition to the standard comparison target pattern of “Please enlarge the map”, “ , "Zoom the map", "enlarge the map", "enlarge the map", "zoom", etc. The above 20 alternative comparison target patterns (a) to (t), a total of 21
, The same command can be specified. In other words, a voice input using a word having a different wording but having substantially the same content can be specified as a command corresponding to the same command of “enlarge map”, and the user can use the device more conveniently. Nevertheless, there is no need to memorize these multiple phrases as individually corresponding word dictionary data. For example, "map", "enlarge", "large", "zoom" or "please"
Such common words can be constructed as shared word dictionary data, so that a memory for word dictionary data can be realized with a smaller capacity. [Alternative Embodiment] Although the above-described word dictionary data is for the command "map enlargement", a command "current location display" will be described as another specific example.

In this case, for example, the following 15 patterns (A) to (Y) can be considered as the alternative comparison patterns that are the standard comparison patterns “Please display the current location”. (B) Display the current position (b) Display the current position (c) Display the current position (d) Display the current position (e) Display the current position (f) Display the current position (g) Display the current position Display the position (H) Display the current position (R) Display the current position (N) Display the current position (L) Display the current position (ヲ) Where is the current position (W) Where is the present location (f) Where is the present (yo) Where are you now? As the word dictionary data including all of these standard comparison target patterns and alternative comparison target patterns, a structure as shown in FIG. 4 is constructed.

In the case of this alternative embodiment, the word "current position" can be substituted for the word "current location", and the words "please" and "please" for "please display". In addition to setting the alternative comparison target pattern in consideration of the fact that a part is omitted, the command “display current location” is used from the viewpoint that the command is used when the user wants to know the current location. The questionable usages used in ordinary conversations such as "where are you now?" And "now where" are also set as alternative comparison patterns. In addition, "where is your current location", which combines them, is also set.
Of course, other alternative comparison target patterns may be set from the same viewpoint. [Others] In the above-described embodiment, the example in which the voice recognition device 30 is applied to the car navigation system 2 has been described. However, the application destination is not limited to the car navigation system 2 described above. For example, when the voice recognition device is used for an air conditioning system, it is conceivable to adjust the set temperature, select an air conditioning mode (cooling / heating / dry), or select a wind direction mode by voice input.

For example, regarding the set temperature, the phrase “set the set temperature to 25 degrees” or “set temperature 25 degrees” or “set the temperature to 25 degrees” or “set the temperature to 25 degrees” is used. It is preferable to allow the phrase "please", and even simply the phrase "25 degrees". The same applies to the air conditioning mode and the wind direction mode.

The above-described navigation system and air conditioning system are not limited to the case where they are used as on-vehicle equipment, but may be, for example, a portable navigation device or an indoor air conditioner. However, when it is used for an in-vehicle device, downsizing of the voice recognition device is more effective when mounted on a vehicle. Of course, from this perspective,
The present invention can be similarly used for in-vehicle devices other than the navigation system and the air conditioning system. For example, a car audio device is effective. Also, considering a configuration in which opening and closing of the power window and adjustment of the mirror angle are instructed by voice, it is effective even in such a situation.

Although it has been described that there is an advantage unique to an in-vehicle device, the present invention is applied to a speech recognition apparatus which executes a predetermined process according to a speech input instruction from a user. If it does, the same can be considered. For example, the present invention can be similarly applied to a portable information terminal device or an information terminal device set in a street or a parking area.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a car navigation system as one embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration of a speech recognition unit and a dialogue control unit in the speech recognition device.

FIG. 3 is an explanatory diagram showing dictionary data for a command of map enlargement stored in a dictionary unit in a voice recognition unit.

FIG. 4 is an explanatory diagram showing dictionary data corresponding to a command of present location display as another embodiment.

[Explanation of symbols]

2 ... Car navigation system 4 ... Position detector 6 ... Map data input device 8 ... Operation switch group 10 ... Control circuit 12 ... External memory 14 ... Display device 15 ... Remote control sensor 15a ... Remote control 16 ... Geomagnetic sensor 18 ... Gyroscope 20 ... Distance sensor 22 GPS receiver 30 Voice recognition device 31 Voice recognition unit 31a Matching unit 31b Dictionary unit 32 Dialogue control unit 32a Post-processing unit 32b Communication control unit 32c Dictionary control unit 33 Voice synthesis unit 34 voice input unit 35 microphone 36 PTT switch 37 speaker 321 dictionary description file 322 word dictionary construction unit

Claims (6)

[Claims] 1. A comparison target pattern corresponding to each word is stored as word dictionary data, and an input voice is compared with the comparison target pattern, a speech having a high degree of matching is regarded as a recognition result, and substantially the same. A plurality of recognition results indicating the contents are specified as corresponding to the same command, and the following procedure (in order to construct the word dictionary data used in the voice recognition device that outputs the specified command to an external device) ( A method for constructing word dictionary data for a speech recognition device, wherein the word dictionary data is constructed according to 1) to (3). (1) First, for each of the commands, a word that specifies a command target indicating a target object or a target function of the command, a target particle “wo”, a word specifying a command operation indicating the operation content of the command, A standard comparison target pattern configured by sequentially connecting verbs added linguistically after the word specifying the command operation is set. (2) Next, a pattern satisfying at least one or more of the following conditions with respect to the standard comparison target pattern:
It is set as an alternative comparison target pattern having substantially the same contents. Synonyms are used for words that identify the command target or command action. The inflected form is used for verbs added in a language custom after a word specifying a command action. The object particle "wo" is omitted. Verbs added linguistically after the word that can specify the command action are omitted. If the verb added in the language custom after the word that can identify the command action is "Please", "Please"
Only has been omitted. (3) When the standard comparison target pattern and the alternative comparison target pattern have the same syllable data, they are based on a tree structure allocated so as to have the same parent, but the number of paths reaching the vertex. Assigns syllable data constituting the pattern to be compared and identification data indicating the end of a word to each vertex in accordance with a preceding scan so as to correspond to a directed graph format similar to a tree structure which is not necessarily 1. 2. The word dictionary data construction method for a speech recognition device according to claim 1, wherein the alternative pattern to be compared in step (2) has substantially the same contents as a standard pattern to be compared. The speech recognition device according to claim 1, wherein the one to be set as satisfies at least one or more of the six conditions obtained by adding "the command object is omitted" to the above. How to build word dictionary data. 3. A voice input means for inputting voice, a dictionary means for storing a comparison target pattern corresponding to each word as word dictionary data, and a voice input via the voice input means, A recognition unit that recognizes a pattern having a high degree of coincidence with the comparison target pattern stored in the unit as a recognition result; and, based on the recognition result by the recognition unit, recognizes a plurality of recognition results having substantially the same contents in the same manner. Claims 1. A speech recognition device comprising: a specifying unit that specifies a command corresponding to a command; and a command output unit that outputs the command specified by the specifying unit. The word dictionary data stored in the dictionary unit is: 1
A speech recognition device characterized by being constructed by the dictionary construction method according to (1). 4. The speech recognition apparatus according to claim 3, further comprising command target input means for inputting information on a command target that can be omitted from an output destination of the command, wherein word dictionary data stored in the dictionary means is provided. Is claim 2
A speech recognition device characterized by being constructed by the word dictionary construction method according to (1). 5. A voice recognition device according to claim 3, further comprising a navigation device, wherein said voice input means of said voice recognition device needs to be specified at least when said navigation device performs a navigation process. The command is used by a user to input an instruction of a predetermined navigation processing related command by voice, and the command output unit is configured to output the specified command to the navigation device. A navigation system characterized by the following. 6. A voice recognition device according to claim 4, further comprising a navigation device, wherein the navigation device obtains information about an omissible command target based on a currently executed navigation process. Command output means for outputting to the voice recognition device, the voice input means of the voice recognition device uses an instruction of a predetermined navigation process related command which needs to be specified at least when the navigation device performs the navigation process. A navigation system, wherein the command output unit is configured to output the specified command to the navigation device.