JP7624571B2 - Voice recognition system, voice recognition device, voice recognition method, and program used for facility inspection, etc. - Google Patents

Description

The present invention relates to a voice recognition system, a voice recognition device, a voice recognition method, and a program used for equipment inspection, etc.

Technology is known that assists inspectors when inspecting equipment, etc.

For example, there is a known technology that uses a tablet terminal or handheld terminal to scan and manage barcodes attached to equipment. By using such barcodes, the status of the equipment can be quickly determined. In addition, there is a known technology that prevents human error such as forgetting to inspect or missing an inspection by digitizing, and supports the implementation of reliable inspections (for example, Non-Patent Document 1, etc.).

"Facility Patrol Inspection System", [online], [searched on March 11, 2020], Internet <URL: http://www.tm-es.co.jp/service-product/services/products/mimawari-kun-mit.html>

In conventional technology, a user may not be able to input the inspection results by voice until all audio explaining the inspection has been output. As a result, the user cannot input voice, start voice recognition processing, or output the results of voice recognition processing until the device has finished outputting the audio, which can result in long wait times for voice input.

The present invention was made in consideration of the above problems, and aims to reduce the time required to input voice.

The speech recognition system according to each embodiment of the present invention includes the following configuration:

A speech recognition system (for example, speech recognition system 1) includes:
A voice input unit (e.g., step S03) for inputting a first voice (e.g., the first voice SD1);
A voice recognition unit (e.g., step S05) that performs voice recognition based on the first voice;
A registration means (e.g., step S01) for registering a second voice (e.g., the second voice SD2);
An output means for outputting the second sound (e.g., step S02);
The voice input device further includes a limiting means (eg, step S04) for limiting the output of the second voice when the first voice is input while the second voice is being output.

This configuration can reduce the time it takes to input voice.

It is also desirable for the voice recognition system to be used for equipment inspection (see, for example, Figure 6).

This configuration allows inspections to be carried out efficiently.

In addition, the voice recognition system
an input data generating means for generating first input data indicating the content of the first voice based on a speech recognition result of the first voice;
a storage means for storing second input data used to check the first input data;
It is preferable that the configuration further includes a determination means for comparing the first input data with the second input data to determine whether or not the first input data is abnormal.

In addition, the voice recognition system
The second input data is
Inputted before the first input data (for example, the previous result V21),
The determination means is
It is desirable to determine that an abnormality exists if the first input data and the second input data are different, or if the first input data is outside an allowable range (for example, allowable range V22) with respect to the second input data.

In addition, the voice recognition system
The second input data is
A value indicating a normal value or a normal range (for example, normal value V23),
The determination means is
It is preferable that if the first input data and the second input data are different, or if the first input data is outside the normal range, it is determined that an abnormality has occurred.

In this way, if there is a difference when compared with the previous result V21, i.e., the most recent inspection result, it is determined that there is an abnormality in the equipment. When such a check is performed, the voice recognition system 1 can notify the user of the equipment in which the abnormality is occurring.

The tolerance range V22 may also be configured to provide a margin for the previous result V21. In particular, when a numerical value is subject to inspection, minute fluctuations in the numerical value are often not abnormal. Therefore, if an acceptable range can be set in this way, abnormalities can be determined with high accuracy.

In addition, the voice recognition system
A speech recognition system including a first information processing device and a second information processing device,
The first information processing device,
Performing speech recognition using a first dictionary (e.g., the first dictionary D211);
The second information processing device is
The speech recognition means recognizes the first speech by using a second dictionary (e.g., the second dictionary D212) (e.g., FIG. 17 ),
The second dictionary is preferably a dictionary for the field of inspection.

Since the second information processing device is a portable information processing device, the memory area MEM tends to be smaller than that of the first information processing device.

The second dictionary D212 is a dictionary for the inspection field. Therefore, the data capacity of the second dictionary D212 can be made smaller than that of the first dictionary D211. Therefore, it is preferable that the second dictionary D212 is a dictionary that can be stored even in an information processing device with a memory area MEM that has a smaller storage capacity than the first information processing device, such as the second information processing device.

The dictionary for the inspection field is a dictionary suitable for voice recognition of terms frequently used in inspection. For example, a dictionary for the inspection field is set with numerical values, terms used in inspection results, names of equipment, etc. In this way, when a dictionary limited to terms frequently used in inspection is used, voice recognition can be realized in a small memory area, and voice recognition can be performed with high accuracy to generate the first input data V10.

In addition, the voice recognition system
The apparatus further includes a noise canceling unit that cancels noises (e.g., a first noise NZ1, a second noise NZ2, and a third noise NZ3) included in the first sound,
The voice recognition means
performing voice recognition using the third voice in which the noise has been attenuated by the noise canceling means;
It is desirable that the noise canceling means sets frequency bands to be cancelled (for example, a first frequency band FR1 and a second frequency band FR2) for each site or position (for example, as shown in FIG. 18).

In this way, noise can be canceled in a way that is appropriate for each site or location. As a result, voice recognition can be performed using the third voice in which the noise contained in the voice has been attenuated, thereby improving the accuracy of voice recognition.

In addition, the voice recognition system
It is desirable for the limiting means to limit the output of the second voice by stopping the output of the second voice, reducing the volume of the second voice, gradually reducing the volume of the second voice, starting the output of a voice to be output after the second voice, or increasing the output speed of the second voice (e.g., step S04).

When this type of restriction process PR1 is performed, it is possible to reduce the output of unnecessary audio.

In addition, the voice recognition system
The registration means registers a plurality of inspection items in association with the second sound,
A group setting means for setting a group (for example, group GS) that includes the inspection items (for example, FIG. 13 );
an omission operation means for inputting an omission operation (e.g., an omission operation C3) for omitting the group;
It is preferable that when the omission operation is input, output of the second sound based on the inspection item belonging to the group corresponding to the omission operation is omitted.

For example, equipment may not need to be inspected if it is not in operation. If inspection of such equipment can be omitted, inspections can be carried out more efficiently.

In addition, the voice recognition device (for example, the mobile terminal 11)
A voice input means for inputting a first voice;
a voice recognition unit that performs voice recognition based on the first voice;
A registration means for registering a second voice;
an output means for outputting the second sound;
The device further includes a limiting means for limiting the output of the second sound when the first sound is input while the second sound is being output.

Also, a speech recognition method (for example, FIG. 3) performed by a speech recognition system,
a speech input step of inputting a first speech by the speech recognition system;
a speech recognition step in which a speech recognition system performs speech recognition based on the first speech;
a registration step in which the voice recognition system registers the second voice;
an output step of the speech recognition system outputting the second speech;
The voice recognition system includes a limiting step of limiting output of the second voice when the first voice is input while the second voice is being output.

It is also a program for causing a computer to execute the speech recognition method (for example, FIG. 3).

Each embodiment of the present invention can reduce the time required to input voice.

FIG. 1 is a diagram illustrating an example of a system configuration of a voice recognition system 1. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing device. FIG. 4 is a diagram illustrating an example of an overall process in the first embodiment. FIG. 2 is a diagram showing an example of a registration database D1. FIG. 11 is a diagram illustrating an example of a processing result in the first embodiment. FIG. 4 is a diagram showing an example of a screen display of a mobile terminal in the first embodiment. FIG. 2 is a diagram illustrating an example of a functional configuration according to the first embodiment. FIG. 11 is a diagram illustrating an example of an overall process in the second embodiment. FIG. 11 is a diagram illustrating an example of a processing result in the second embodiment. FIG. 11 is a diagram showing an example of a screen display of a mobile terminal in the second embodiment. FIG. 11 is a diagram illustrating an example of a functional configuration according to a second embodiment. FIG. 13 is a diagram showing a modified example in which group setting and omission are performed. FIG. 13 is a diagram illustrating an example of group settings. FIG. 13 is a diagram showing an example in which second input data is used as a previous result. FIG. 13 is a diagram illustrating an example of setting a tolerance range. FIG. 13 is a diagram showing an example in which second input data is set as a normal value or a normal range. FIG. 23 is a diagram showing a modified example in which a first dictionary D211 and a second dictionary D212 are used. FIG. 13 is a diagram illustrating an example of setting a frequency band to be canceled for each site. FIG. 13 is a diagram showing a first modified example of interruption and release. FIG. 13 is a diagram showing a second modified example of interruption and release. FIG. 13 is a diagram showing a modified example of audio input and audio output.

The optimal and minimal form for implementing the invention will be described below with reference to the drawings. Note that when the same reference numerals are used in the drawings, they indicate similar configurations, and duplicated explanations will be omitted. Also, the specific examples shown in the drawings are merely examples, and the configuration may further include configurations other than those shown in the drawings.

First Embodiment
<Overall configuration example>
1 is a diagram showing an example of a system configuration of a voice recognition system 1. For example, the voice recognition system 1 includes a server 10, a mobile terminal 11 which is an example of a voice recognition device, and an earphone 12.

The server 10, mobile terminal 11, and earphones 12 are connected via a network NW.

As shown in the figure, the user 13 wears earphones 12 and inspects the equipment. The user 13 also inspects the equipment while holding a mobile terminal 11. Meanwhile, the server 10 is installed in a location different from the site where the equipment inspection is performed. Therefore, the setting values, data, etc. are input in advance to the server 10, and when the inspection is performed, the mobile terminal 11 acquires the setting values, data, etc. via the network NW.

The equipment to be inspected is, for example, air conditioning equipment. Therefore, the inspection site is a machine room in a company. During the inspection, the values indicated by measuring instruments such as a pressure gauge, a voltmeter, an ammeter, a chemical level gauge, a thermometer, and a hygrometer are checked by the user 13.

The server 10 and the mobile terminal 11 are information processing devices. For example, the information processing device is a device with the following hardware configuration:

<Hardware configuration example>
2 is a diagram showing an example of the hardware configuration of an information processing device. For example, an information processing device such as a server 10 has a hardware configuration including a CPU (Central Processing Unit, hereinafter referred to as "CPUHW1"), a storage device HW2, a network interface HW3, an input device HW4, an output device HW5, and an interface HW6.

CPUHW1 is an example of a calculation device and a control device.

The memory device HW2 is a main memory device, an auxiliary memory device, etc.

The network interface HW3 is a communication device that transmits and receives data to and from external devices via a network.

The input device HW4 is a device for inputting user operations. For example, the input device HW4 is a mouse and a keyboard.

The output device HW5 is a device for outputting the processing results to the user. For example, the output device HW5 is a display, etc.

Interface HW6 is a connector for connecting peripheral devices.

The earphones 12 are an input device for inputting voice. Voice uttered by the user 13 and input through the earphones 12 is then input to the mobile terminal 11, where processing such as voice recognition is performed. The earphones 12 are also an output device for outputting voice. Therefore, the earphones 12 output voice to the user 13 through processing by the mobile terminal 11.

Hereinafter, the audio input by the user 13 through the earphone 12 is referred to as the "first audio SD1." On the other hand, the audio output to the user 13 through the earphone 12 is referred to as the "second audio SD2."

<Overall processing example>
3 is a diagram showing an example of the overall processing in the first embodiment. Below, as shown in the figure, the overall processing will be explained by dividing it into "pre-processing" and "main processing". The "pre-processing" is a process that is performed before the "main processing" and is a preparation process for operating the speech recognition system 1. Therefore, the "pre-processing" and the "main processing" do not have to be performed consecutively, and it is sufficient that the "pre-processing" is completed before the "main processing" starts.

<Example of pre-processing>
In the "pre-processing", the following processing is carried out:

In step S01, the voice recognition system 1 registers, for example, inspection items and data for the second voice SD2 to construct a registration database D1. That is, the registration database D1 indicates the contents to be read out as the second voice SD2 during the inspection. In addition, the correspondence between the inspection items and which data of the input voice should be output as the second voice SD2 and at what timing is also registered in the registration database D1. For example, the registration procedure is performed in this manner. In this way, in the "pre-processing", the voice recognition system 1 inputs and sets data to enable the second voice SD2 to be output in the "main processing". Specifically, the registration database D1 is constructed as follows.

Figure 4 shows an example of registration database D1. As shown in the figure, registration database D1 is composed of "item numbers" and "inspection items."

For example, the content output by the second audio SD2 is the content entered in "inspection item." Specifically, the name of the equipment that identifies the equipment to be inspected (for example, in the example where the "item number" is "1," it is "fire pump") is entered in "inspection item." Therefore, when the name of the equipment is read out by the second audio, the user 13 can identify the equipment that is to be inspected by listening to the second audio SD2.

Furthermore, as shown in the figure, detailed inspection content is input in "Inspection item". Specifically, the detailed inspection content is "Pressure value" in the example where "Item number" is "1". When such detailed inspection content is read out by the second voice, the user 13 knows that by listening to the second voice SD2, he or she will inspect the pressure value and input the pressure value with the first voice SD1.

In addition, in step S01, the voice recognition system 1 inputs the dictionary D2. That is, in the "pre-processing", the voice recognition system 1 inputs and sets data to enable voice recognition, that is, to recognize the voice input in the first voice SD1 as words in the "main processing".

<Example of this process>
"This process" begins when inspection begins.

In step S02, the voice recognition system 1 performs an output procedure to output the registered second voice SD2.

Hereinafter, it is assumed that the first voice SD1 is input while step S02 is being executed, i.e., while the second voice SD2 is being output. In such a case, the voice recognition system 1 proceeds to step S03.

In step S03, the voice recognition system 1 performs a voice input procedure to input the first voice SD1. That is, in step S03, the voice recognition system 1 detects that the user 13 has spoken. In this way, when the first voice SD1 is input, the voice recognition system 1 proceeds to step S04.

In step S04, the voice recognition system 1 performs a restriction procedure to restrict the output of the second voice SD2.

In step S05, the voice recognition system 1 performs a voice recognition procedure to perform voice recognition based on the first voice SD1.

In step S06, the voice recognition system 1 outputs the voice recognition result of the first voice SD1 as the second voice SD2.

For example, when the overall processing described above is performed, the processing result will be as follows.

Figure 5 shows an example of a processing result in the first embodiment. Below, we will explain by comparing a "first item execution example" that does not impose restrictions and a "second item execution example" that imposes restrictions.

In the "first item execution example," the voice recognition system 1 first outputs the eleventh output EX11 by the second voice SD2. In this example, the eleventh output EX11 is a voice that successively indicates the name of the equipment to be inspected and the details of the inspection, such as "fire pump: pressure."

In response to the 11th output EX11, the user 13 performs an operation to input the 12th output EX12, which indicates the results of inspecting the equipment indicated by the 11th output EX11, to the speech recognition system 1 using the first voice SD1. Specifically, in this example, the user 13 reads out "12.3", which is the "pressure" value indicated by the pressure gauge of the "fire pump" equipment. The voice emitted by this reading becomes the 12th output EX12.

Next, in this example, the voice recognition system 1 performs voice recognition on the twelfth output EX12. Then, in the thirteenth output EX13, the details of the completed inspection and the voice recognition result are output by the second voice SD2. As shown in the figure, the first half of the contents output in the thirteenth output EX13 is a repetition of the eleventh output EX11. Next, in the second half of the thirteenth output EX13, the voice recognition result, i.e., the same contents as the twelfth output EX12, is output. In this way, if no restrictions are imposed and the voice recognition is performed correctly, the voice recognition system 1 operates in the following order: output of inspection items, etc., input of the inspection results, voice recognition, and output of the voice recognition results.

For a voice recognition system 1 capable of such operations, in the "second item execution example", while the second voice SD2 is being output, the user 13 inputs the first voice SD1 to limit the output of the second voice SD2.

In the illustrated example, the second audio SD2 to be restricted is set as the 21st output EX21 (step S02). The 21st output EX21 is the inspection item that follows the inspection item indicated by the 11th output EX11. If the 21st output EX21 is not restricted, the content is "Fire pump: exterior."

As shown in the figure, in the "second item execution example", while the 21st output EX21 is being output (step S02), the user 13 performs an operation to input the 22nd output EX22, which indicates the results of inspecting the equipment indicated by the 21st output EX21, to the speech recognition system 1 using the first voice SD1 (step S03).

In this way, when the 22nd output EX22 is input while the 21st output EX21 is being output, the speech recognition system 1 performs a restriction process PR1 that restricts the output of the 21st output EX21 (step S04). For example, the restriction process PR1 is a process that stops the output of the 21st output EX21.

The user 13 also reads out the result of checking the appearance of the "fire pump" equipment: "X - Liquid leaking" (step S03). The voice emitted by this reading becomes the 22nd output EX22.

Next, in this example, the voice recognition system 1 performs voice recognition on the 22nd output EX22 (step S05). Then, in the 23rd output EX23, the voice recognition system 1 outputs the content that the inspection has been completed and the voice recognition result by the second voice SD2, similar to the 13th output EX13 (step S06). As shown in the figure, the first half of the content output in the 23rd output EX23 is a repetition of the 21st output EX21. Next, in the second half of the 23rd output EX23, the voice recognition result, i.e., the same content as the 22nd output EX22, is continuously output.

<Screen display example>
6 is a diagram showing an example of a screen display of a mobile terminal in the first embodiment. For example, inspection is performed in the following order. However, the screen display, order, and input items shown in the figure are not essential.

Figure 6 (A) is a screen showing the "Main Menu." Hereinafter, the screen shown in Figure 6 (A) will be referred to as the eleventh screen PN11. When "Inspection Flow" is pressed on the eleventh screen PN11, the screen transitions to Figure 6 (B), which is a screen for preparing to perform an inspection.

Figure 6 (B) is a screen for selecting "Inspection flow list." Hereinafter, the screen shown in Figure 6 (B) will be referred to as the 12th screen PN12. The 12th screen PN12 is a screen for selecting the equipment to be inspected. Next, when equipment is selected on the 12th screen PN12, the screen transitions to Figure 6 (C), which is a screen for specifying the user 13 who will perform the inspection.

Figure 6 (C) is a screen for "selecting a worker." Hereinafter, the screen shown in Figure 6 (C) will be referred to as the thirteenth screen PN13. The thirteenth screen PN13 is a screen for selecting a user 13 who will be the worker who will perform the inspection. Next, when a user 13 is selected on the thirteenth screen PN13, the inspection begins and the screen transitions to Figure 6 (D).

Figure 6 (D) is a screen for the first inspection item. Hereinafter, the screen shown in Figure 6 (D) is referred to as the 14th screen PN14. The 14th screen PN14 shows the name of the equipment to be inspected and the progress of the inspection in the "Inspection Item List". The 14th screen PN14 also shows the "boiler pressure", which is the inspection content, in the "Inspection Item". For example, in response to the display of the 14th screen PN14, a second voice SD2 indicating the same content is output (step S02). Next, when the first voice SD1 indicating the inspection result is input (step S03), voice recognition is performed (step S05) and the screen transitions to Figure 6 (E).

Figure 6 (E) is a screen showing the first inspection result. Hereinafter, the screen shown in Figure 6 (E) will be referred to as the fifteenth screen PN15. The fifteenth screen PN15 differs from the fourteenth screen PN14 in that the numerical value "0.66" is displayed as the recognition result. This "0.66" is the result of voice recognition of the first voice SD1. When the voice recognition result is displayed on the screen in this way, it is easy for the user 13 to check the voice recognition result.

For example, the first inspection item is inspected as described above. The second and subsequent inspection items are processed in the same manner.

Figure 6 (F) is the screen for the second inspection item. Hereinafter, the screen shown in Figure 6 (F) is referred to as the 16th screen PN16. Whereas the first inspection item was "boiler internal pressure", the inspection item in the 16th screen PN16 is "forced fan current value". Then, in the 16th screen PN16, as in the case of the first inspection item, a second voice SD2 indicating the same content as in the screen is output (step S02). Next, when the first voice SD1 indicating the inspection result is input (step S03), voice recognition is performed (step S05) and the screen transitions to Figure 6 (G).

Figure 6 (G) is a screen showing the second inspection result. Hereinafter, the screen shown in Figure 6 (G) will be referred to as the 17th screen PN17. The 17th screen PN17 differs from the 16th screen PN16 in that the number "43" is displayed as the recognition result. This "43" is the result of voice recognition of the first voice SD1.

For example, when the first voice SD1 is input while the second voice SD2 indicating the same content as the screen is being output on the 14th screen PN14 or the 16th screen PN16, the voice recognition system 1 restricts the output of the second voice SD2 (step S04), starts voice recognition, and transitions to the next screen.

<Functional configuration example>
7 is a diagram showing an example of a functional configuration in the first embodiment. For example, the voice recognition system 1 has a functional configuration including a voice input unit 1F11, a voice recognition unit 1F12, a registration unit 1F13, an output unit 1F14, a restriction unit 1F15, an input data generation unit 1F16, a storage unit 1F17, a determination unit 1F18, a noise cancellation unit 1F19, a group setting unit 1F20, and an omission operation unit 1F21.

The voice input means 1F11 and the output means 1F14 are realized, for example, by earphones 12. The voice recognition means 1F12, the registration means 1F13, the restriction means 1F15, the input data generation means 1F16, the storage means 1F17, the determination means 1F18, the noise cancellation means 1F19, the group setting means 1F20, and the omission operation means 1F21 are realized by the cooperation of the arithmetic units, storage units, input devices, and output devices, such as the CPU HW1, the storage unit HW2, the input unit HW4, the output unit HW5, and the interface HW6, of the mobile terminal 11 or the server 10.

In particular, if the user 13 performing the inspection is accustomed to inspections, the user 13 may be familiar with the details of the inspection, such as the contents and sequence of the inspection, even if they are not guided by the second voice SD2. In such a case, the user 13 may want to finish the inspection quickly by inputting the inspection results in the first voice SD1 before all of the second voice SD2 has been output. Therefore, as in the above example, if the first voice SD1 is input while the second voice SD2 is being output, it is desirable to limit the output of the second voice SD2. In this way, if the output of the second voice SD2 is limited and the first voice SD1 can be input, the user 13 does not have to wait for the second voice SD2 to finish being output, and the time required to input voice can be shortened.

Second Embodiment
The second embodiment is realized by a voice recognition system 1 having the same system configuration as the first embodiment. The hardware configurations of the mobile terminal 11 and the server 10 in the second embodiment are also the same as those in the first embodiment. The second embodiment differs from the first embodiment in the overall processing. The following description will focus on the differences and omit redundant description.

<Overall processing example>
8 is a diagram showing an example of the overall processing in the second embodiment. Compared with the first embodiment, the second embodiment is different in that steps S21 to S23 are performed. Also, the "pre-processing" in the second embodiment is the same as that in the first embodiment.

In step S02, the voice recognition system 1 outputs the registered second voice SD2.

Below, an example will be described in which the interrupt operation C1 is input while step S02 is being executed, i.e., while the second voice SD2 is being output. In such a case, the voice recognition system 1 proceeds to step S21. Note that the type of operation that the interrupt operation C1 is is set in advance.

In step S21, the voice recognition system 1 inputs an interruption operation C1. In this manner, when the interruption operation C1 is input, the voice recognition system 1 interrupts the output of the second voice SD2.

When a cancel operation C2 is input to cancel the interruption while the output of the second voice SD2 is interrupted, the voice recognition system 1 proceeds to step S22. Note that the type of operation that the cancel operation C2 is is set in advance.

In step S22, the voice recognition system 1 releases the interruption based on the release operation C2.

In step S23, the voice recognition system 1 resumes output of the second voice SD2.

For example, when the overall processing described above is performed, the processing result will be as follows.

Figure 9 is a diagram showing an example of a processing result in the second embodiment. Below, we will explain an example in which the 21st output EX21 of the "Second item execution example" in the second embodiment is the target of interruption.

For example, suppose that an interruption operation C1 is input while an output voice indicating the 21st output EX21 is being output (step S02). When the interruption operation C1 is input, the voice recognition system 1 performs an interruption process PR2 (step S21). Hereinafter, the point in time when the interruption operation C1 is input and interrupted is referred to as the "first point in time."

As shown in the figure, the interruption process PR2 is a process that stops the output midway through the 21st output EX21. Furthermore, the interruption process PR2 is a process that prevents output from being performed after the interruption until the cancel operation C2 is input.

Next, when the cancel operation C2 is input, the voice recognition system 1 cancels the interruption (step S22). Specifically, when the cancel operation C2 is input, the voice recognition system 1 outputs an output voice indicating the 200th output EX200, and resumes output of an output voice indicating the 21st output EX21.

The 200th output EX200 is an output sound that indicates the contents of the inspection that was completed at the first point in time. This output sound reminds the user 13 of how much of the inspection has been completed.

Then, after the 200th output EX200 is output, the output of the 21st output EX21 is resumed (step S23).

<Screen display example>
10 is a diagram showing an example of a screen display of a mobile terminal in the second embodiment. For example, inspection is performed in the following order. However, the screen display, order, and input items shown in the figure are not essential.

Figure 10 (A) is a screen showing the first inspection result. Hereinafter, the screen shown in Figure 10 (A) will be referred to as the 21st screen PN21. The 21st screen PN21 is an example of a screen that is displayed when the first inspection is "steam temperature" and "120" is entered as the inspection result. As shown in the figure, the voice recognition result is output by text output B2. Then, if the voice recognition result is incorrect when looking at the text output B2, it is desirable to be able to press the re-input button B1 and re-input the inspection result.

Similarly, the second, third, and fourth inspections are performed in that order. In this example, the 22nd screen PN22 shown in FIG. 10(B) is an example of the screen that is displayed when the second inspection result is entered.

The 23rd screen PN23 shown in FIG. 10(C) is an example of a screen that is displayed when the third inspection result is input. And the 24th screen PN24 shown in FIG. 10(D) is an example of a screen that is displayed when the fourth inspection result is input.

<Functional configuration example>
11 is a diagram showing an example of a functional configuration in the second embodiment. For example, the voice recognition system 1 has a functional configuration including a voice input unit 1F11, a voice recognition unit 1F12, an output unit 1F14, a release unit 1F101, and an interruption unit 1F100.

The voice input means 1F11 and the output means 1F14 are realized, for example, by earphones 12. The voice recognition means 1F12, the release means 1F101, and the interruption means 1F100 are realized by the cooperation of the arithmetic units, memory units, input devices, and output devices, such as the CPU HW1, the memory unit HW2, the input unit HW4, the output unit HW5, and the interface HW6, of the mobile terminal 11 or the server 10.

For example, as shown in FIG. 10, an inspection is performed based on multiple inspection items that are set in advance. During the flow of such an inspection work, an interrupting task may occur, for example, the telephone rings or someone speaks to the user. When such an interruption occurs, it is desirable for the user 13 to be able to interrupt the inspection with an interrupt operation C1. In this way, if the user 13 can interrupt the inspection with the interrupt operation C1, the user 13 can interrupt the inspection and then perform the interrupting task. Then, after completing the interrupting task, the user 13 can resume the inspection with a cancel operation C2.

<Modification>
The inspection items may be either "standardized" items that are input and checked in the form of numerical values or "o" or "x" (or "YES" or "NO", etc.), or "non-standardized" items that the user 13 inputs as comments such as "missing". Also, images taken by the user 13 may be attached to the inspection results.

The voice recognition system 1 may be used for purposes other than equipment inspection. In other words, the object to be inspected may be a device other than equipment. The voice recognition system 1 may be used for purposes other than inspection.

When the voice recognition system 1 is used for equipment inspection, the time required for equipment inspection can be shortened, making the work more labor-saving.

The voice recognition system 1 is preferably used for inspections in the machine room of air conditioning equipment. When inspecting the machine room of air conditioning equipment, the inspection points are often scattered throughout the machine room. Furthermore, when a large number of large pieces of equipment are installed, the meters to be inspected for each piece of large equipment are often also scattered throughout the machine room. In such machine rooms, the inspection items are often diverse. Furthermore, when performing such a large number of inspections, the user moves around a lot to carry out the inspections. Furthermore, in some cases, the user may have to move around by climbing a ladder, etc.

Furthermore, when inspecting, the user also performs incidental tasks such as turning on the lights (and turning them off when finished) or opening the door to the power panel. For this reason, there is a desire to eliminate manual input of inspection results as much as possible. Therefore, if the voice recognition system 1 can be used to input inspection results by voice, it will be easier for the user to inspect.

In addition, when there is a lot of movement, such as in the case of inspections in machine rooms as described above, there is a high possibility that other workers and managers will talk to the inspector while he or she is performing the inspection. Therefore, if the voice recognition system 1 can pause and cancel the inspection, the inspection can be stopped and resumed, making the inspection more efficient.

Depending on the layout of the equipment and meters, multiple instruments may be installed in one location. In such installation situations, many inspection results are often input in one location. Therefore, if the user waits until the device finishes outputting audio before inputting audio, the waiting time for inputting audio increases, and the inspection takes time. On the other hand, if the device is configured to limit audio output while allowing audio input, the user does not have to wait for audio output to finish, and the time required to input audio can be reduced. This allows the inspection to be performed efficiently.

The earphones 12 used to output sound are preferably of the cochlear type. Cochlear type earphones 12 are shaped to block external sounds when worn. For example, at work sites where equipment is in operation, the environment is often noisy due to the sounds of the equipment operating. Therefore, with cochlear type earphones 12, the output sound is easy to hear even in noisy environments. Also, compared to bone conduction type earphones, cochlear type earphones can often be made smaller.

The order and the necessity of inspection for each inspection item may be set. For example, the second audio SD2 is output in the order of the "item numbers" in the registration database D1. Therefore, before the inspection is performed, the order in which the inspections are performed may be changed in "pre-processing" by changing the setting of the "item numbers" or the correspondence between the "inspection items" and the "item numbers."

It may also be possible to set "inspection items" that are not to be inspected to "OFF."

The items to be inspected may vary depending on the operating status of the equipment. Therefore, if the order and necessity of inspections can be set, the contents of the inspection can be flexibly changed to match the operating status on the day of inspection.

The information processing device is not limited to the hardware configuration shown above. For example, the information processing device may have a hardware configuration that further includes a calculation device, a control device, a storage device, an input device, an output device, and peripheral devices other than those shown above. Furthermore, the input device and the output device may be integrated into a touch panel or the like. Note that the mobile terminal 11 and the server 10 may be devices with different hardware configurations. Furthermore, the connection with the peripheral devices and other devices may be wired or wireless.

The mobile terminal 11 may also be an information processing device such as a tablet, smartphone, or mobile computer.

<Modification of Restriction Processing>
Note that the restriction process PR1 is not limited to a process of stopping the output of the second sound SD2. For example, the restriction process PR1 may be a process of limiting the second sound SD2 by lowering the volume of the second sound SD2. The restriction process PR1 may also be a process of limiting the second sound SD2 by gradually lowering the volume of the second sound SD2. In this way, the restriction process PR1 may limit the output of the second sound SD2 by lowering the volume more than usual.

Alternatively, the restriction process PR1 may be a process that restricts the second audio SD2 that starts to be output as the audio to be output next to the second audio SD2. In this way, the restriction process PR1 may restrict the output of the second audio SD2 by moving from the inspection item currently indicated by the second audio SD2 to the next inspection item.

Alternatively, the restriction process PR1 may be a process that restricts the second audio SD2 by increasing the output speed of the second audio SD2. In other words, the restriction process PR1 may restrict the output of the second audio SD2 by playing the second audio SD2 in fast forward.

When this type of restriction process PR1 is performed, it is possible to reduce the output of unnecessary audio.

<Modification for Group Setting and Omission>
In addition, some of the inspections may be omitted. For example, the following process may be omitted.

Figure 12 shows a modified example of group setting and omission. In the following, an example will be described in which the equipment to be inspected is "first equipment," "second equipment," "third equipment," and "fourth equipment," and inspections are performed in this order.

In this example, similar to the first embodiment, the inspection items for the first equipment are output in the 31st output EX31. Also, in this example, the inspection of the first equipment including the 31st output EX31 is subject to omission.

When omission is to be performed, the user 13 performs the omission operation C3. For example, the omission operation C3 is an operation such as pressing a specific button or inputting a pre-set word such as "omission" in the first voice SD1. Note that the type of operation that the omission operation C3 is is set in advance.

As shown in the figure, when an omission operation C3 is performed while the 31st output EX31 is being output, the voice recognition system 1 performs an omission process PR3 that omits the inspection of the first equipment, including the 31st output EX31. For example, the omission process PR3 is a process that stops the output of the 31st output EX31 and proceeds to the inspection of the second equipment. Therefore, when the omission operation C3 is performed, the voice recognition system 1 starts outputting the 32nd output EX32, which indicates the inspection items for the second equipment.

Which of the pre-registered inspection items corresponds to the inspection of the first equipment is determined based on, for example, a group setting such as the following:

Figure 13 is a diagram showing an example of group settings. For example, a case will be described where inspection items such as those shown in the figure have been registered in advance. In this example, the "inspection items" with "item numbers" of "1" and "2" are inspections of the first equipment. Similarly, the "inspection items" with "item numbers" of "3" through "5" are inspections of the second equipment. Furthermore, the "inspection item" with "item number" of "6" is an inspection of the third equipment. Furthermore, the "inspection item" with "item number" of "7" is an inspection of the fourth equipment.

A group GS is set for these multiple inspection items. For example, as shown in the figure, the voice recognition system 1 performs a group setting procedure in which a numerical value is entered into "group" to set the group GS. In this example, "inspection items" with "item numbers" of "1" and "2" are grouped together and set as a group called "G1". Similarly, in this example, "inspection items" with "item numbers" of "3" through "5" are grouped together and set as a group called "G2". Note that groups GS are not limited to per equipment, and may be set as one group, such as "G3," which groups together the inspection items for the third and fourth equipment.

As shown in FIG. 12, for example, when the omission operation C3 is performed on the 31st output EX31 corresponding to the inspection of the first equipment, the voice recognition system 1 determines that the omission operation C3 is an operation corresponding to the group "G1."

As shown in FIG. 13, the group GS of "G1" includes the "inspection items" with the "item numbers" "1" and "2." Therefore, the voice recognition system 1 performs an omission procedure to omit the second voice SD2 indicating the "inspection items" with the "item numbers" "1" and "2."

Note that the omission operation C3 may be an operation of inputting the name of the group or facility to be omitted.

For example, equipment may not need to be inspected if it is not in operation. If inspection of such equipment can be omitted, inspections can be carried out more efficiently.

<Modification for Checking>
It is desirable that the voice recognition system 1 is configured to check the inspection results and determine whether or not there is an abnormality in the equipment. Hereinafter, the value to be checked, i.e., the data input by the first voice SD1 and showing the voice recognition result, will be referred to as "first input data". On the other hand, the data used to check the first data will be referred to as "second input data".

When the first input data is generated through the first data generation procedure, i.e., through voice recognition, the voice recognition system 1 performs a judgment procedure in which the first input data is compared with the second input data to judge whether or not there is an abnormality.

The second input data only needs to be stored by the storage procedure before the first input data, i.e., before the first input data is checked. For example, the second input data is stored as follows:

Figure 14 is a diagram showing an example in which the second input data is the previous result. Below, an example will be described in which the inspection items are the same as those in the first embodiment. For example, the second input data is the previous inspection result (hereinafter referred to as "previous result V21") that is stored corresponding to each "inspection item" as shown in the figure as "second input data."

Note that the previous result V21 is based on the assumption that no abnormalities were found in the previous inspection. Therefore, if an abnormality was found in the previous inspection, the previous inspection result or a normal value may be entered in the previous result V21.

In other words, the voice recognition system 1 compares the first input data, i.e., the current inspection result, with the previous result V21 to determine whether there is a difference. Then, if the first input data and the second input data differ, the voice recognition system 1 determines that there is an abnormality.

The range that is determined to be normal may have a certain degree of tolerance. In other words, the voice recognition system 1 determines that the first input data is abnormal if it is outside the tolerance range for the second input data. For example, the tolerance range is set as follows:

Figure 15 is a diagram showing an example of setting an acceptable range. For example, the acceptable range V22 indicates the range within which the result is judged to be "normal" even if it deviates from the previous result V21, as in the "acceptable range" shown in the figure. Note that this example is an example in which the acceptable range V22 is set to "none" so that all cases where there is no match with the previous result V21, such as when the "item number" is "2", are judged to be "abnormal".

In this way, the tolerance range V22 may be configured to provide a margin for the previous result V21. In particular, when a numerical value is subject to inspection, minute fluctuations in the numerical value are often not abnormal. Therefore, if an acceptable range can be set in this way, abnormalities can be determined with high accuracy.

In this way, if there is a difference when compared with the previous result V21, i.e., the most recent inspection result, it is determined that there is an abnormality in the equipment. When such a check is performed, the voice recognition system 1 can notify the user of any equipment in which an abnormality has occurred.

The check may also be performed by setting a value indicating a normal value or normal range in the second input data as follows:

Figure 16 is a diagram showing an example in which the second input data is set to a normal value or normal range. For example, a value such as the illustrated "second input data normal value or normal range" (hereinafter referred to as "normal value V23") may be set in advance.

When the normal value V23 is set, the voice recognition system 1 can grasp the value or range that is considered normal. Therefore, the voice recognition system 1 compares the first input data with the normal value V23, and if the first input data and the second input data are different, or if the first input data is outside the normal range, it determines that there is an abnormality. When such a check is performed, the voice recognition system 1 can notify equipment in which an abnormality is occurring.

<Modification using first and second dictionaries>
In addition, data, programs, etc. may be downloaded to the mobile terminal 11, and the entire processing may be performed by the mobile terminal 11. For example, the communication environment at the site where the inspection is performed may be poor. That is, the communication environment at the site may be such that it is difficult to communicate between the mobile terminal 11 and the server 10.

Therefore, a configuration may be adopted in which setting values and the like input in advance are downloaded from the server 10 before the start of the inspection, and the entire process can be completed without communication between the mobile terminal 11 and the server 10 while the inspection is being performed. In this case, when the inspection is completed and an environment is created in which communication can be performed between the mobile terminal 11 and the server 10, the inspection results and the like may be uploaded from the mobile terminal 11 to the server 10.

In such cases, it is desirable that the dictionary used for voice recognition is a dictionary for the inspection field, as follows:

Figure 17 is a diagram showing a modified example using a first dictionary D211 and a second dictionary D212. Hereinafter, an information processing device that can be used via a network NW, such as the server 10, is referred to as a "first information processing device." In contrast to the first information processing device, an information processing device such as a mobile terminal 11 that can be used by a user 13 on-site is referred to as a "second information processing device."

As shown in the figure, the dictionary used for voice recognition in the first information processing device is called the "first dictionary D211." On the other hand, the dictionary used for voice recognition in the second information processing device is called the "second dictionary D212."

Since the second information processing device is a portable information processing device, the memory area MEM tends to be smaller than that of the first information processing device.

The second dictionary D212 is a dictionary for the inspection field. Therefore, the second dictionary D212 can have a smaller data capacity than the first dictionary D211. For this reason, it is preferable that the second dictionary D212 is a dictionary that can be stored even in an information processing device with a memory area MEM that has a smaller storage capacity than the first information processing device, such as the second information processing device.

The dictionary for the inspection field is a dictionary suitable for voice recognition of terms frequently used in inspection. For example, a dictionary for the inspection field is set with numerical values, terms used in inspection results, names of equipment, etc. In this way, when a dictionary limited to terms frequently used in inspection is used, voice recognition can be realized in a small memory area, and voice recognition can be performed with high accuracy to generate the first input data V10.

<Modification of noise cancellation>
The voice recognition system 1 is preferably configured to cancel noise contained in the first voice SD1, and preferably performs voice recognition based on a voice (hereinafter referred to as a "third voice") obtained by canceling the noise of the first voice SD1.

Noise cancellation is performed by attenuating noise in the target frequency band using, for example, a low-pass filter, a band-pass filter, or a combination of multiple filters. For example, the filter is realized by a digital filter or a filtering circuit.

Furthermore, it is desirable to set the frequency bands to be canceled individually for each site or location. Below, we explain an example of setting the frequency bands to be canceled for each site.

Figure 18 shows an example of setting frequency bands to be cancelled for each site. Hereinafter, Figure 18(A) is for the "first site" and Figure 18(B) is for the "second site." Note that the frequency bands to be cancelled may be set based on location information using GPS or the like.

For example, at the first site, if a certain period of voice (assuming that no voice is being produced by the user 13) is input and frequency analysis (e.g., FFT (Fast Fourier Transform)) is performed on the voice input data, the frequency analysis results shown in Figures 18(A) and 18(B) are obtained.

Figure 18 (A) shows an example in which noise (hereinafter, the noise generated at the first site is referred to as "first noise NZ1") occurs in the frequency band shown in Figure 18 (A) (hereinafter, referred to as "first frequency band FR1") at the first site.

Figure 18 (B) is an example where noise (hereinafter, the two noises occurring at the second site are referred to as "second noise NZ2" and "third noise NZ3") occurs at the second site in the frequency band shown in Figure 18 (B) (hereinafter, referred to as "second frequency band FR2").

Noise may vary from site to site or location. For example, if different types of equipment are in operation at each site, the equipment may emit different sounds, and the frequency band in which noise occurs may also be different. In this example, the first site generates a first noise NZ1, while the second site generates a second noise NZ2 and a third noise NZ3. In this example, the first site generates noise in a lower frequency band than the second site.

Therefore, at the first site, it is desirable to target the first frequency band FR1 for cancellation in order to attenuate the first noise NZ1. On the other hand, at the second site, it is desirable to target the second frequency band FR2 for cancellation in order to attenuate the second noise NZ2 and the third noise NZ3.

The first frequency band FR1 and the second frequency band FR2 are set, for example, by analyzing the sound for each site or location before an inspection is performed. The frequency bands may be set by inputting numerical values.

In this way, noise can be canceled in a way that is appropriate for each site or location. As a result, voice recognition can be performed using the third voice in which the noise contained in the voice has been attenuated, thereby improving the accuracy of voice recognition.

<Variations of Suspension and Cancellation>
The interruption process PR2 may be performed when the voice recognition system 1 recognizes a predetermined word by voice recognition, as follows.

Figure 19 shows a first modified example of interruption and cancellation. Below, an example will be described in which the second audio SD2 is output in the same way as in the second embodiment. Compared to the second embodiment, this modified example differs in that specific words are used for interruption and cancellation.

In this example, when the word "pause" (hereinafter referred to as "interrupted speech C11") is input in the first speech SD1, the speech recognition system 1 interrupts the output of the second speech SD2.

Then, after the interruption, the voice recognition system 1 cancels the interruption when the word "cancel" (hereinafter referred to as "cancel voice C21") is input as the first voice SD1.

The words used for the interruption sound C11 and the release sound C21 are set in advance.

In this way, being able to pause and cancel using specific words allows the user 13 to perform the operation even when, for example, holding a tool in their hand, improving operability.

Figure 20 shows a second modified example of interruption and cancellation. In this example, as shown in the figure, the operation of pressing a button on the earphone 12 is the interruption operation.

For example, the first button press (hereinafter referred to as the "first operation C12") causes the voice recognition system 1 to interrupt the output of the second voice SD2.

Next, by pressing the button a second time (hereinafter referred to as the "second operation C22"), the voice recognition system 1 releases the interruption.

Being able to operate using such earphones 12 eliminates the need to take out the mobile terminal 11.

As shown in the figure, the 200th output EX200 may be omitted. In this example, the speech recognition system 1 outputs the output speech from the point in time when the 21st output EX21 is started (hereinafter referred to as the "second point in time"), which is a point in time before the first point in time when the interruption occurred. The second point in time is set in advance, or is the initial point in time of the output speech that was output at the first point in time, etc. In this way, the output speech may be restated.

Also, the interruption and release may be realized by a button, for example, by interrupting while a specific button is pressed and releasing when the button is no longer pressed.

Additionally, the interruption and release may be linked to the start and end of a specified operation by the mobile terminal 11 or the like. For example, when the voice recognition system 1 detects an incoming call to the mobile terminal 11, it interrupts the output of the second voice SD2. The specified operation is set in advance. The specified operation may also be an operation by an external device.

<Modifications of Audio Input and Audio Output>
21 is a diagram showing a modified example of voice input and voice output. Hereinafter, an example will be described in which a numerical value of "12.3" (hereinafter referred to as "target numerical value 20") is handled by voice as shown in the figure.

In the first voice SD1, i.e., input, it is preferable that the target number 20 is input one digit at a time. Specifically, as shown in the figure, the pronunciation 21 that becomes the first voice SD1 is preferable to break down the target number 20 and read out one digit at a time, such as "ichi," "ni," "ten," and "san."

In the second audio SD2, i.e., in the output, it is desirable that the target numerical value 20 is output so as to express the entire numerical value. Specifically, as shown in the figure, it is desirable that the data audio 22 which becomes the second audio SD2 expresses the entire target numerical value 20 and is output so as to express the entire numerical value, such as "ten thousand and ten."

When the voice reads out one digit at a time, as in pronunciation 21, voice recognition can be performed with high accuracy. On the other hand, when data voice 22 expresses the entire number, it is easier for user 13 to understand the number by voice rather than when it is output one digit at a time.

<Modification of Offline Speech Recognition>
It is desirable to have a configuration that can recognize input speech even in an offline environment.

In construction and maintenance sites, the communication environment often makes it difficult to connect to the Internet. As a result, it is often difficult to use the cloud all the time. Therefore, if a voice recognition engine in the cloud is always used, voice recognition and the like may not be able to be performed at the work site. To avoid such situations, a configuration that uses a voice recognition engine that operates within the mobile terminal 11 is desirable. In particular, a configuration in which standard, short words are voice-recognized by the voice recognition engine that operates within the mobile terminal 11 is desirable.

With this configuration, the voice recognition system 1 can be used offline. Therefore, even in an environment where it is difficult to connect to the Internet, voice can be input using the voice recognition system 1.

Furthermore, it is preferable that the mobile terminal 11 is configured to record non-standard voice input such as comments by the user 13. In an environment connected to the Internet, such as an office, the mobile terminal 11 is preferably configured to transmit the recorded voice to a voice recognition engine on the cloud and convert it into text.

In this way, it is desirable to have a configuration that allows separate use of a voice recognition engine that can be used offline and a voice recognition engine that can be used online.

The voice recognition engine may be switched depending on the communication environment. Specifically, the mobile terminal 11 measures radio wave strength to determine whether the communication environment is good or bad. The quality of the communication environment may be registered in advance for each site, or previous judgment results may be stored. In this way, a configuration that takes the communication environment into consideration can prevent problems such as the voice recognition engine on the cloud being unable to be used.

In addition, in order to record offline as described above, it is preferable that the mobile terminal 11 is configured to have a memory unit and a recording data generation unit that can record input voice data, generate recording data, and store the recording data.

The storage unit and the recording data generation unit are not limited to being used offline. In other words, the storage unit and the recording data generation unit may generate and store recording data even in an environment where a cloud-based voice recognition engine can be used.

Since communication may be suddenly cut off, it is desirable to have a configuration that allows recording even if a cloud-based voice recognition engine is available. This configuration allows for backup.

<Other embodiments>
The embodiment is not limited to the above example. For example, the number of devices is not limited to the number shown in the above example. Therefore, each device in the above example may be a system of two or more devices. On the other hand, the device may be a single device. In addition, the type and combination of information processing devices may not be the devices shown above.

The embodiment is not limited to the above process. For example, the speech recognition method according to the present invention may be performed in an order other than that described above. Furthermore, the speech recognition method may be executed on multiple information processing devices. In other words, each step in the speech recognition method may be executed in a redundant, distributed, parallel, virtualized, or combination of these.

The embodiment may be realized by a program. That is, a computer such as an information processing device may control an arithmetic unit and a storage device based on the program to execute the above-mentioned method. The program may be recorded on a computer-readable recording medium and distributed. The recording medium may be a medium such as a magnetic tape, a flash memory, an optical disk, a magneto-optical disk, or a magnetic disk. Furthermore, the program may be distributed through a telecommunications line.

The present invention is not limited to the above-described configuration of the embodiment shown above, and may be combined with other elements. These aspects may be changed without departing from the spirit of the present invention, and may be appropriately determined according to the application form.

1 Speech recognition system 1F100 Interruption means 1F101 Cancellation means 1F11 Speech input means 1F12 Speech recognition means 1F13 Registration means 1F14 Output means 1F15 Restriction means 1F16 Input data generation means 1F17 Storage means 1F18 Judgment means 1F19 Noise cancellation means 1F20 Group setting means 1F21 Omission operation means 10 Server 11 Mobile terminal 12 Earphone 13 User 20 Target numerical value 21 Pronunciation 22 Data voice C1 Interruption operation C2 Cancellation operation C3 Omission operation C11 Interruption voice C12 First operation C21 Cancellation voice C22 Second operation D211 First dictionary D212 Second dictionary EX11 Eleventh output EX12 Twelfth output EX13 Thirteenth output EX21 Twenty-first output EX22 Twenty-second output EX23 Twenty-third output EX31 Thirty-first output EX32 32nd output EX200 200th output FR1 First frequency band FR2 Second frequency band GS Group MEM Memory area NW Network NZ1 First noise NZ2 Second noise NZ3 Third noise PR1 Restriction process PR2 Interruption process PR3 Omission process SD1 First voice SD2 Second voice V10 First input data V21 Previous result V22 Allowable range V23 Normal value

Claims (14)

A voice recognition system for use in equipment inspection, comprising:
A voice input means for inputting a first voice;
a voice recognition unit that performs voice recognition based on the first voice;
a registration means for registering a plurality of inspection items, an inspection sequence, and a second sound in association with each other;
an output means for outputting the second sound;
a limiting means for limiting an output of the second sound when the first sound is input while the second sound is being output,
the output means outputs the second voice indicating the inspection items in accordance with the inspection order, and when the first voice indicating the inspection result of the inspection item is input, outputs the second voice indicating the inspection item and a voice recognition result of the first voice, and outputs the second voice indicating the next inspection item;
the limiting means stops output of the second voice and starts output of a voice to be output next to the second voice .
Voice recognition system. displaying a speech recognition result of the first speech indicating the inspection result on a screen of a mobile terminal;
When a re-input operation is performed on the mobile terminal, re-input of the inspection result displayed on the screen is accepted.
2. The speech recognition system of claim 1. A voice input means for inputting a first voice;
a voice recognition unit that performs voice recognition based on the first voice;
A registration means for registering a second voice;
an output means for outputting the second sound;
a limiting means for limiting an output of the second sound when the first sound is input while the second sound is being output,
The registration means registers a plurality of inspection items in association with the second sound,
A group setting means for setting a group of the inspection items;
and an omission operation means for inputting an omission operation for omitting the group,
When the omission operation is input, output of the second sound based on the inspection item belonging to the group corresponding to the omission operation is omitted.
Voice recognition system. 4. The voice recognition system according to claim 3, which is used for inspecting facilities. an input data generating means for generating first input data indicating the content of the first voice based on a speech recognition result of the first voice;
a storage means for storing second input data used to check the first input data;
5. The speech recognition system according to claim 1, further comprising a determination unit for comparing the first input data with the second input data to determine whether or not the first input data is abnormal. The second input data is
inputted before the first input data,
The determination means is
6. The speech recognition system according to claim 5 , wherein if the first input data and the second input data are different, or if the first input data is outside an allowable range for the second input data, it is determined that an abnormality has occurred. The second input data is
A value indicating a normal value or normal range,
The determination means is
6. The speech recognition system according to claim 5 , wherein if the first input data and the second input data are different, or if the first input data is outside the normal range, it is determined that there is an abnormality. A speech recognition system including a first information processing device and a second information processing device,
The first information processing device,
Performing speech recognition using the first dictionary;
The second information processing device is
the speech recognition means for recognizing the first speech using a second dictionary;
8. A speech recognition system according to claim 1 , wherein the second dictionary is a dictionary for the field of inspection. a noise canceling unit that cancels noise included in the first sound to generate a third sound,
The voice recognition means
performing speech recognition using the third speech;
9. The speech recognition system according to claim 1, wherein the noise canceling means sets a frequency band to be cancelled for each site or position. A voice recognition device used for inspecting equipment, comprising:
A voice input means for inputting a first voice;
a voice recognition unit that performs voice recognition based on the first voice;
a registration means for registering a plurality of inspection items, an inspection sequence, and a second sound in association with each other;
an output means for outputting the second sound;
a limiting means for limiting an output of the second sound when the first sound is input while the second sound is being output,
the output means outputs the second voice indicating the inspection items in accordance with the inspection order, and when the first voice indicating the inspection result of the inspection item is input, outputs the second voice indicating the inspection item and a voice recognition result of the first voice, and outputs the second voice indicating the next inspection item;
the limiting means stops output of the second voice and starts output of a voice to be output next to the second voice .
Speech recognition device. A voice input means for inputting a first voice;
a voice recognition unit that performs voice recognition based on the first voice;
A registration means for registering a second voice;
an output means for outputting the second sound;
a limiting means for limiting an output of the second sound when the first sound is input while the second sound is being output,
The registration means registers a plurality of inspection items in association with the second sound,
A group setting means for setting a group of the inspection items;
and an omission operation means for inputting an omission operation for omitting the group,
When the omission operation is input, output of the second sound based on the inspection item belonging to the group corresponding to the omission operation is omitted.
Speech recognition device. A voice recognition method performed by a voice recognition system used for inspecting equipment , comprising:
a speech input step of inputting a first speech by the speech recognition system;
a speech recognition step in which a speech recognition system performs speech recognition based on the first speech;
a registration step in which the voice recognition system registers a plurality of inspection items, an inspection sequence, and a second voice in association with each other;
an output step of the speech recognition system outputting the second speech;
a limiting step of limiting output of the second voice when the first voice is input while the second voice is being output,
The output step includes outputting the second voice indicating the inspection items according to the inspection order, and when the first voice indicating an inspection result of the inspection item is input, outputting the second voice indicating the inspection item and a voice recognition result of the first voice, and outputting the second voice indicating the next inspection item;
the restriction step includes stopping output of the second voice and starting output of a voice to be output next to the second voice .
Speech recognition methods. A speech recognition method performed by a speech recognition system, comprising:
a speech input step of inputting a first speech by the speech recognition system;
a speech recognition step in which a speech recognition system performs speech recognition based on the first speech;
a registration step in which the voice recognition system registers the second voice;
an output step of the speech recognition system outputting the second speech;
a limiting step of limiting output of the second voice when the first voice is input while the second voice is being output,
The registration step includes registering a plurality of inspection items in association with the second sound,
A group setting procedure for setting a group that includes the inspection items;
and an omission operation procedure for inputting an omission operation for omitting the group,
When the omission operation is input, output of the second sound based on the inspection item belonging to the group corresponding to the omission operation is omitted.
Speech recognition methods. A program for causing a computer to execute the speech recognition method according to claim 12 or 13 .

Images