JP7237356B2 - CAD control support system - Google Patents

Description

The present invention relates to a CAD control support system that controls a CAD device by voice.

In housing design, a CAD device is used to create a floor plan. In the operation of this device, in addition to drawing, input processing of various items such as the type of material and layout of the structure of the house is performed. A wide variety of command lists are prepared for this work in the CAD device. However, it is extremely troublesome to display a list of drawing commands and component selection commands and select the corresponding command. For example, the user selects major items to medium items, displays a parts list, and selects the corresponding part from the list. In many cases, complicated selection operations are performed by opening the menu many times until the parts list is displayed. Part of this problem is solved by directly selecting the corresponding command by voice (Patent Document 1) (Patent Document 2).

JP 2018-49350 A JP-A-4-211828

By adopting the above technique, it is possible to operate a CAD device by directly designating drawing commands and component selection commands. However, CAD equipment for building design uses a wide variety of terminology. Speech recognition engines available on the market have extremely high recognition accuracy and thus have high utility value. This can be used, but in order to recognize technical terms, a dictionary must be prepared and sufficiently learned. Otherwise, erroneous conversion will occur frequently. Technical term dictionaries are not easy to maintain. In addition, since a certain amount of erroneous conversion must be allowed, a confirmation screen must be displayed each time so that an erroneous command is not selected, hindering improvement in operability.

The following configurations are means for solving the above problems.

<Configuration 1>
a storage device storing a command list in which a kana character string composed only of kana characters is used as a key and a kana character string is associated with a CAD control command including an instruction for calling a part to be drawn by the CAD device ;
a speech input device for sending input speech signals to a speech recognition engine for controlling a CAD device;
Kana character conversion means for obtaining a recognition result from a speech recognition engine and converting the recognition result into a kana character string composed only of kana characters;
The kana character string output from the kana character conversion means is used as a search key to search the command list, select the command for CAD control stored in correspondence, and send it to the CAD device. a sending command selection device;
Converts the recognition result by the speech recognition engine obtained by the kana character conversion means to a kana character string,
The above command selection device retrieves the command list,
The command selection device judges whether the command can be selected with the character string,
A voice CAD control support system characterized by sending selected commands to a CAD machine if possible.

<Configuration 2>
Converts the recognition result by the speech recognition engine obtained by the kana character conversion means to a kana character string,
The above command selection device retrieves the command list,
The command selection device judges whether the command can be selected with the character string,
If not, the command selector parses the string and
Search the command list again with the character string parsed by the command selection device,
The command selection device determines whether the command can be selected from the parsed character string,
If possible, send the selected command to the CAD device,
A CAD control support system by voice according to configuration 1, characterized by executing a process of performing error processing if it is not possible.

<Configuration 3>
The above command list is a one-to-one or N-to-one correspondence between unique kana character strings with no duplication and commands for CAD control.
The above-mentioned command selection device automatically includes kana character strings whose meanings are equivalent to those registered in the database when syntactic analysis is performed, in the N-to-one N kana character strings, so that the command list by learning is obtained. 2. The CAD control support system according to configuration 2, characterized by executing the optimization process of .

<Configuration 4 >
A computer program that causes a computer to function as the means according to any one of configurations 1 to 3 .

<Configuration 5 >
A computer-readable recording medium recording the computer program according to configuration 4 .

The recognition result by the speech recognition engine is mixed with Chinese characters, and in the case of technical terms, there are also conversion errors. If the pronunciation is correctly recognized, the effect of the conversion error can be eliminated by restoring the kana character string. By creating a command list in which kana character strings and commands are associated with each other in advance, commands can be easily called directly by voice. The conventional work of sequentially displaying various command menus and selecting a desired command can be greatly simplified.

If kana character strings and commands are associated one-to-one, there is no need to hit multiple commands and select them. This improves workability. Part selection is the most frequent and cumbersome operation of a CAD machine. If the part name can be specified directly and the corresponding part can be selected immediately, workability will be improved. If there are multiple names for the same part, operability can be further improved by establishing an N-to-1 correspondence.

1 is a functional block diagram of a CAD control support system of the present invention; FIG. FIG. 10 is an explanatory diagram showing an example of a command list; FIG. 10 is an explanatory diagram of command list optimization processing; 4 is an operation flowchart of the arithmetic processing device;

Hereinafter, embodiments of the present invention will be described in detail for each example.

In the CAD control support system 10 of the present invention shown in FIG. 1, a user 11 uses a CAD device 12 to create, for example, a design drawing of a house while using a voice input control function. Here, in order to utilize this function, the arithmetic processing unit 23 of the CAD device 12 includes a voice input device 28, a communication device 29, a kana character conversion means 32, and a command selection device 34, in addition to the original control function of CAD. and are provided.

The storage device 22 of the CAD device 12 also stores a command list 20 in which the kana character strings 16 and commands 18 for CAD control are associated with each other. The CAD device 12 is also connected to a server device 14 via a communication network (not shown). The server device 14 improves its speech recognition performance day by day by learning using a large amount of data, and uses a device that provides a speech conversion service on the market.

A voice input device 28 has a microphone or the like, and has a function of sending a voice signal 24 input by the user 11 to the voice recognition engine 26 in order to control the CAD device 12 . The kana character converting means 32 has a function of acquiring the recognition result 30 by the speech recognition engine 26 and converting the recognition result 30 into a kana character string 16 composed only of kana.

The communication device 29 has a function of controlling the transmission and reception of signals between the voice input device 28 and the text converting means 32 and the server device 14 . The command selection device 34 searches the command list 20 using the kana character string 16 output from the kana character converting means 32 as a search key, and selects the command stored in association with the kana character string 16. is selected and sent to the CAD device 12.

Next, description will be made with reference to FIG. The user 11 creates a floor plan including, for example, a living room, kitchen, hallway, entrance, toilet, etc., selects the types of doors and windows, and selects the material of the floorboards.

Conventionally, the CAD device was operated as follows without using voice operation. First, for example, the layout of the living room, the kitchen, etc. is determined, and the sample figure list is displayed. After arranging the figure of the main room, for example, select a part such as a kitchen unit. Select a type, select a manufacturer, and then select a product number to display the desired part. The same work is done when selecting parts such as windows and toilets.

On the other hand, according to the present invention, the following operations are performed by voice operation. For example, when there are several types in the living room, the target figure can be directly selected by pronouncing "ribin gutaipuei" (living type A). Even if technical terms such as ``Ikken majikiri fusuma'' (one-partition fusuma) are listed, the desired figure can be specified directly. Create a command list like that.

A recognition result 30 by the speech recognition engine 26 is a character string containing Chinese characters. Commands for controlling CAD equipment are highly technical. Therefore, many kanji conversion errors may occur. Looking at the result of voice recognition and correcting the conversion error each time would halve the advantage of the voice input operation.

Therefore, in the present invention, the recognition result 30 by the speech recognition engine 26 is converted into a kana character string 16 composed only of kana. If the pronunciation is correctly recognized, the conversion error can be eliminated by restoring the kana character string 16 . Furthermore, a command list 20 in which kana character strings 16 and commands are associated with each other is created in advance. You can just add phonetic kana to all commands and use them.

It is preferable to prepare this command list 20 as a database in which each unique kana character string 16 without duplication is associated with a command 18 for CAD control. For example, a command list may be created so that the same command is selected regardless of whether (type A) is "type" or "type". As for "kana", either hiragana or katakana may be used. It is sufficient if only the pronunciation is converted into a character string.

As described with reference to FIG. 2, kana character strings and commands may be associated one-to-one or N-to-one. As long as the target command can be called directly with a unique Kana character string without duplication, it does not matter if the Kana character string is long. A kana character string may be set by combining words that are easy for the user to remember. If multiple words come to mind to invoke the same command, an N-to-1 correspondence eliminates the need for the user to memorize all the kana character strings accurately.

For example, you can say "madorizu" (floor plan) to display a screen with various floor plans arranged side by side, or say "kitchin no heimenzu" (kitchen floor plan), A screen on which floor plans of various kitchens are arranged may be displayed. Then select one of the drawings with the mouse. In addition, it is also possible to call up a specific figure directly, such as "Toreyu Nitto Seihin Bangouei Biishi Ichi Ni San" (toilet unit product number ABC123). It can be widely used for commands for calling various parts to be drawn by the CAD device 12 .

The drawing commands (instructions for specifying data processing operations) are the same. can be specified. In either case, the desired command can be directly invoked easily and without error by voice. Therefore, the conventional work of sequentially displaying various command menus and selecting a desired command can be greatly simplified.

FIG. 3 is an explanatory diagram of command list optimization processing, and FIG. 4 is an operation flowchart of the arithmetic processing unit including this processing.
For example, instead of pronouncing "toreyu ni t sei hin bang uei ibi ishii ichi ni san", the user pronounces "tore yu ni t de seihin bangou" as shown in FIG. 3(a). gaeibiishiiichinisan", or as shown in (d), "seihinbangougaeibiishiiichinisannotoreyunitto" I want to select commands without hesitation.

For this reason, it is good to associate as many conceivable equivalent kana strings as possible with the desired command. Developing this further, using the syntactic analysis function by AI, automatically include kana strings with equivalent meanings in N-to-one N kana strings when syntactically parsed. , it can also optimize the command list by learning.

A simple processing example in it will be described using a flowchart. First, at step S11 in FIG. 4, voice input by the voice input device 28 is accepted. In step S12, the communication device 29 transmits the speech signal to the speech recognition engine 26. FIG. Then, the communication device 29 receives the recognition result in step S13.

In step S14, the kana character converting means 31 converts the received recognition result into a kana character string. In step S15, the command selection device 34 refers to the command list 20. FIG. At step S16, it is determined whether the command can be selected.

If the result of this determination is YES, the process shifts to the process of executing the command. If NO, the process proceeds to step S17. Here, in step S17, the command selection device 34 parses the syntax. For example, as shown in step S18, particles are detected and deleted from the kana character strings shown in (b) and (d) of FIG.

After that, in step S19, the command list 20 is referred again to select the appropriate command. If it can be selected, it is determined in step S20 that the kana character strings shown in (b) and (d) of FIG. 3 are equivalent to the kana character strings in the command list. If it is determined to be equivalent, the process proceeds to step S21, and if not, error processing is performed.

In error processing, it is preferable to display a comment to the effect that recognition has failed on the CAD device 12 . In step S21, the kana character string containing the particles determined to be equivalent is added to the command list as shown in FIG. The CAD device 12 then executes the appropriate command. The processing in step S21 may be performed in a batch.

In the case of the kana character string in FIG. 3(d), it is only necessary to determine whether the command list 20 includes a kana character string that includes all three words obtained by deleting the particles. . If more than one match is found, the user should be asked to select a command, or error processing should be performed.

For example, if the relationship between the recognized kana character string and the selected command is always displayed on the CAD device 12, the user can pay attention to the pronunciation and expressions, and the CAD device 12 can be operated more conveniently by voice. become able to.

10 CAD control support system 11 user 12 CAD device 14 server device 16 kana character string 18 command for CAD control 20 command list 22 storage device 23 arithmetic processing device 24 voice signal 26 voice recognition engine 28 voice input device 29 communication device 30 recognition Result 32 Kana character conversion means 34 Command selection device

Claims (5)

a storage device storing a command list in which a kana character string composed only of kana characters is used as a key and a kana character string is associated with a CAD control command including an instruction for calling a part to be drawn by the CAD device ;
a speech input device for sending input speech signals to a speech recognition engine for controlling a CAD device;
Kana character conversion means for obtaining a recognition result by a speech recognition engine and converting the recognition result into a kana character string composed only of kana characters;
The kana character string output from the kana character conversion means is used as a search key to search the command list, select the command for CAD control that has been associated and stored, and send it to the CAD device. a sending command selection device;
Converts the recognition result by the speech recognition engine obtained by the kana character conversion means to a kana character string,
The above command selection device retrieves the command list,
The command selection device judges whether the command can be selected with the character string,
A voice CAD control support system characterized by sending selected commands to a CAD machine if possible. Converts the recognition result by the speech recognition engine obtained by the kana character conversion means to a kana character string,
The above command selection device retrieves the command list,
The command selection device judges whether the command can be selected with the character string,
If not, the command selector parses the string and
Search the command list again with the character string parsed by the command selection device,
The command selection device determines whether the command can be selected from the parsed character string,
If possible, send the selected command to the CAD device,
2. The CAD control support system by speech according to claim 1, wherein processing for performing error processing is executed if it is not possible. The above command list is a one-to-one or N-to-one correspondence between unique kana character strings with no duplication and commands for CAD control.
The above-mentioned command selection device automatically includes kana character strings whose meanings are equivalent to those registered in the database when syntactic analysis is performed, in the N-to-one N kana character strings, so that the command list by learning is obtained. 3. The CAD control support system according to claim 2, wherein the optimization processing of is executed. A computer program that causes a computer to function as the means according to any one of claims 1 to 3. 5. A computer-readable recording medium recording the computer program according to claim 4.

Images