JPH0331176A - Information input device for elevator - Google Patents

Abstract

PURPOSE:To operate an elevator with voice without reducing the service for the elevator operation by providing a means which reports the sound recognition result for a fixed period and preferentially complies with the voice command being reported. CONSTITUTION:A sound input microphone is installed in an elevator cage and at a hall, and the voice command issued to this microphone and the outside noise are identified by a filter section 7. The temporary display of the recognition result for this voice commend is reported for a fixed period by a reporting section 16. The voice command being reported has priority, if the sound recognition result is judged as abnormal, another command is given again to cancel the temporary display being displayed, and the recognition result of the newly inputted voice command is temporarily displayed again. The operation command is corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to an apparatus for operating an elevator by recognizing the voice of an elevator user.

(Prior art) Elevators are installed in various buildings, but the actions that elevator users directly operate are: 1. Press the call button in the elevator hall.

2. Press the destination button inside the elevator car.

3. Press the button to open and close the elevator door There are three types of items that can be considered.

These are the sequences in which the user presses the corresponding button and the elevator responds as a result.

However, the buttons installed in elevators are not always easy to use, and often leave users dissatisfied.

For example, if you are carrying luggage in both hands, you must unload your luggage before pressing the button. Also, when used by a blind person, the user must manually search for the small buttons. Additionally, if a person is standing in front of the button, it may be difficult to operate it.

As a method to overcome these drawbacks, an information input device that recognizes and processes a user's voice has been proposed.

This system places a microphone for voice input in the place where a conventional call button is installed, analyzes the voice, and performs the corresponding action if it is determined to be a command to operate the elevator. For example, if a user pronounces ``Gokai-J'', the call response lamp for the 5th floor will be turned on and the elevator will be moved to the 5th floor.

This is exactly the same as pressing the conventional call button for the 5th floor.

Speech recognition processing stores pre-registered words as a dictionary, and compares the results of frequency analysis of the input voice with the data stored in the dictionary, and finds the closest voice that is more similar than a certain percentage. The corresponding processing is performed when the

As a voice recognition method used in elevators, a method called speaker-independent word recognition is generally used, and it is possible to recognize the voice of any speaker, that is, anyone.

Also, the audio content is "Hirake", "Tojiro", and "Gokai".
etc. are recognized in word units. The number of registered words is approximately 16, and several types of words are stored in the dictionary for speech that is difficult to recognize and distinguish. For example, the first floor “Ikkai” “
The recognition rate is improved by having multiple dictionaries such as Ichikai-j.

As described above, the voice recognition method allows general elevator operations. However, the environment when registering speech into a dictionary is done in a quiet place, and if there is an elevator in a similar environment, the recognition rate will be over 90%, but the actual environment where the elevator is placed is dark. The recognition rate is decreasing due to the loud noise.

In order to prevent this, processing such as not performing a recognition operation unless the voice exceeds a certain level has been implemented, resulting in a practical information input device.

FIG. 5 shows an external view when voice recognition is used in an elevator hall. A hall indicator 2 and a hall call button 3 are generally installed near the elevator door 1, and in this case, a voice recognition microphone 4 is installed. If an elevator user speaks into this microphone, the elevator can be operated. For example, if the user utters ``sea urchin'', the upward direction hall call button 3 will light up, indicating that the call has been registered for the elevator.

Next, the configuration of the speech recognition processing section will be explained using FIG. 6. Since the audio input from the microphone 4 is an analog value, the A/D converter 6 converts it into a digital value corresponding to the amplitude of the audio for processing by the CPU 5, and a bandpass filter unit 7 divides the digital value into frequency domains. As a result of limiting the voice band with, for example, the sampling frequency is 12KHz, 1
An array of 2-bit digital values is obtained. However, since this value is a continuous value for the voice, the voice section detection unit 8 extracts only the voice features and converts it into a spectral sequence of 8 m5ec units to compress the amount of information. In order to perform extraction, silent parts and effective speech are separated and, as a result, the data of the speech to be actually recognized is compiled.

The sampling unit 9 normalizes the length of input speech, which differs for each user, according to the length of the speech section of the dictionary pattern used for recognition.

As a result, the audio data is converted into data of 256 points, and the CPU 5 calculates the similarity by comparing it with the 256 points of dictionary patterns stored in the dictionary storage unit 10, and the dictionary pattern with the highest degree of similarity is recognized. The configuration is such that the result is output to the operation output section 17. This operation output section is often a general digital output circuit, and the program storage section 12 is a program in which the above procedure is programmed. The data from the operation output section 11 is sent to the elevator control section 13.
is input to perform various controls.

FIG. 7 illustrates the above-mentioned speech recognition processing process as a speech waveform.

Step 1 is the audio waveform input from microphone 4,
Since this waveform contains various noises in addition to the words to be recognized, as in step 2, the speech section detecting section 8 extracts the necessary word sections.

Since this process has to be performed by detecting that audio with the most characteristic amplitude has been input, there is a possibility of false detection if there is loud noise. Step 3 is the output result of the bandpass filter section 7, which is data in the frequency band of 16 channels.

(Problem to be Solved by the Invention) As explained above, if the voice commands of the elevator users are used to recognize the voice and operate the elevator, it is possible to perform contactless processing, which has an excellent effect, but on the other hand, it is contactless. Therefore, users were often left dissatisfied if their own voices were not processed accurately.

For example, if a user utters "Nikai" but recognizes that it is on the third floor, this not only causes dissatisfaction of the user who issued the command, but also causes the service to deteriorate as a result of unnecessary operation and dissatisfaction of other users.

In addition, if the voice recognition process is repeated many times to obtain the correct result, a large number of erroneous calls will occur, and there is a possibility that the voice recognition process itself will be judged to be malfunctioning and will no longer be used.

Another disadvantage is that in cases where the device cannot be operated, it is not possible to determine whether the user's voice is poor or the device is malfunctioning.

For example, even though the third floor is a non-stop floor, even if you command "Sankai", no processing is done, resulting in confusion for users.

In addition, as a solution to this problem, a method has been proposed in which the user's commands are responded to with voice, as in JP-A-57-571.73, but responding with voice causes noise to other users, so it is quieter. I was not welcomed in the elevator hall,
Furthermore, when multiple voice recognition devices are installed, there is a drawback that malfunctions may occur due to voice responses.

An object of the present invention is to notify the user of the voice recognition result uttered by the elevator user as a temporary display, and to show the user how to correct erroneous recognition or erroneous operation commands.

[Structure of the invention]

(Means for Solving the Problem) The present invention includes a voice input microphone installed in an elevator car and a hall, a voice recognition unit that processes the voice, an elevator control device that actually processes the recognition result, and an operation command that is displayed. In contrast to conventional information input devices consisting of conventional devices,
A device for temporarily displaying voice recognition results is added to a conventional device for displaying operation commands to provide a device for achieving the above purpose.

Here, the conventional devices for displaying operation commands are the hall call button 3 shown in FIG. The results were shown using lamps, etc. Therefore, even when an operation command is given by voice as in the present invention, the notification to the user can be configured in exactly the same manner as in the past.

(Function) By the way, since the temporary display of voice recognition is to confirm the operation commanded to the user, the notification method is set to a display format different from the final registered display format, and the display result for a certain period of time, Next, the results that would have been temporarily displayed if there was no voice command are actually registered. Additionally, if there is another voice command during the temporary display, the operation command can be corrected by canceling the previously displayed temporary display and temporarily displaying the recognition result of the voice command input this time. It has become.

(Example) An embodiment according to the present invention will be described using FIG. 4.

FIG. 4 shows a case where a voice recognition microphone 4 is installed in an elevator hall and the user's destination floor is registered by voice.

In this example, the destination floor is input before getting on the elevator, so there is no need to register a call in the elevator car after getting on the elevator, and the call is automatically created within the elevator control device.

Further, in this embodiment, an OK lamp 20 indicating that the voice recognition process was successfully performed and a reject lamp 21 indicating that the voice recognition could not be performed are added to the sides of the voice recognition microphone 4.

The operation of these lamps is shown below.

If the operation process is not performed using voice recognition, the elevator user selects the floor he or she wants to go to from the hall destination call button 3 and presses the button, as in the past, and the elevator control device 1
3 is input, and as a result, the lamp of the button of the pressed floor lights up. Once the elevator arrives, the destination call operated earlier is transferred to the call in the elevator car, allowing the user to be moved to the desired floor.

Next, when performing voice recognition processing, when the elevator user approaches the voice recognition microphone 4 and utters an operation, the voice recognition process starts and the user's voice is analyzed. At this time, since the voice cannot be recognized unless it is louder than the surrounding noise, it is necessary to speak within about 30 cm from the voice recognition microphone 4.

Here, elevator users say the floor they want to go to,
When the voice recognition process is performed and it is determined that the voice is normal, the OK clamp 0 lights up and the destination call button 3 blinks. Further, if it is determined that the sound is abnormal, the ejection lamp 21 is turned on.

However, in this case, OK clamp 0, reject 1 to lamp 21 are the voice input to the voice recognition device and the dictionary storage unit 1.
When the contents of 0 are compared and the degree of similarity exceeds a certain standard, the OK lamp 20 lights up. Therefore, even if the user utters "Nikai", if the voice recognition device determines that it is the third floor, the OK clamp 0 may light up and the third floor lamp of the destination call button 3 may blink. This is called misrecognition. In this case, the destination button 3 is only displayed while it is blinking, and is a temporary display without actually controlling the elevator, so the user can see the result and if it is normal, it will be automatically registered after a while. All you have to do is wait, and if it is determined that it is wrong, you can correct it by saying it again.

OK clamp 0, reject 1 to lamp 21 are the results of a series of voice recognition processing. If the similarity of the input voice exceeds a certain standard, OK clamp 0, and if it is below, reject lamp 21 will be notified by the CPU 5. It is turned on for 1 to 2 seconds via the digital output section 16 to notify elevator users of the result.

At this time, if the similarity exceeds a certain standard, CPU 5 is OK.
The result is flashed and displayed not only on the clamp 0 but also on the destination call 3 registration lamp, and an operation is performed in which the elevator user is asked to judge whether the recognition result is normal or abnormal.

This destination call 3 is conventionally controlled by the elevator control unit 3, but by ORing the processing of the registration lamp with the output result from the voice recognition device, the present invention can be achieved without changing the elevator control unit 13. It becomes possible to implement In addition, the flashing display of the registration lamp is the conventional operation output section 1.
The data format is the same as that which was output from 1 to the elevator control unit 13, and the CPU 5 turns it on and off at a constant cycle to make it blink, generally at a cycle of 0.5 seconds. Further, this output and the conventional call registration lamp are configured by simply performing a wired OR. Therefore, if you press that button while the light is blinking, priority will be given to the constantly lit state.

FIG. 1 is a block diagram showing the configuration of an embodiment according to the present invention.

Comparing with FIG. 6, OK clamp 0, reject lamp 21, and destination call 3 are added.

This OK clamp 0, reject lamp 211 destination call 3
and a voice recognition microphone 4 are placed nearby so that they can be seen by the user when they speak.

The voice recognition device constantly observes that the voice of the dictionary pattern stored in the dictionary storage unit 10 is input, but in order to distinguish between operation commands and other noises, it performs processing according to the following procedure. ing.

(2) First, the setting of the gain control section 14 before the A/D conversion section 6 is increased so that voice input is not possible unless the user is close to the microphone.

2. Next, the voice section detection section 8 is set so that it will not be misrecognized even if the user makes a loud voice regardless of the operation command.

Therefore, if the voice interval of an operation command voice word is 1 second, if it is uttered slowly and takes 2 seconds, it will be treated as noise.

In FIG. 1, in addition to the block diagram of the conventional device, a gain controller 1-roll section 14 and a recognition result notification section 16 are added, and a call output section 15 is also added to the elevator control section 13.

The gain control unit 14 is a programmable volume, and has a structure that allows the CPU 5 to adjust the volume setting, so that the user's voice can be detected as described above. The recognition result notification section 16 is a digital output circuit, and is provided with means for temporarily displaying the destination call 3 displayed from the call output section 13 of the elevator control section 13 according to the present invention.

In this embodiment, the connection is simply made for blinking display. Here, the output signal from the recognition result notification unit 16 to the destination call 3 is O
After the K-clamp 0 is turned off, a similar signal is output from the operation output section 11. Also OK lamp 20, reject lamp 2
1 are both output from the recognition result notification section 16.

A processing flowchart according to the present invention is shown in FIG.

In step 1, the voice detected in the voice information input from the voice recognition microphone 4 is not only the user's voice, but also sounds that are not recognized as voice are collected as background noise. This data is used as comparison data for the speech recognition process performed in step 3. In step 2, as described above, the user's voice level is detected. In this process, it is also determined by the voice section detecting section 8 that the proper voice is inputted, and if the voice is proper, the voice recognition process of step 3 is performed.

In step 3, correction by re-inputting the voice is possible according to the flowchart shown in FIG. 3, compared to the conventional voice recognition apparatus.

In step 3A of FIG. 3, a voice input after the recognition process is completed is detected, and if there is an input, the previous recognition result is invalidated and the recognition result of the current voice is prioritized. (Steps 3B and 3C) In step 4, data is compared between the input speech as a result of the speech recognition process and the dictionary storage unit 10, and if a degree of similarity exceeding a certain standard is found, the OK lamp 20 is turned on in step 5. Elevator users are notified that the voice is valid.

Furthermore, if the voice is below the standard, the reject lamp 21 is turned on in step 6 to indicate to the user that the voice is ambiguous, and a re-input is requested.

In step 7, it is determined that a valid voice command can be operated.

For example, when a non-stop floor is commanded, blinking cycle 2 is selected in step 9, and the user is notified that this is different from step 8 of the normal process.

15- =16- In step 10, the recognition result is blinked at the above-mentioned blinking cycle. Step 8 flashing period is 0.5 seconds, step 9
The blinking period is about 0.3 seconds.

In step 11, it is determined whether it is after the end of the voice command or before the operation, and if it is the voice command value, the above-mentioned ramp process is ended in step 12.

In other words, the OK lamp is turned off approximately one second after it is turned on, and the recognition result that has been blinking is sent to the elevator control device, which processes it as if a call had actually been made.

In this example, the elevator user stands in front of the voice recognition microphone, speaks the floor he or she wants to go to, checks the result by checking the lit state of the destination call button, and if the result is correct, it takes 1 second. Wait for a while and it will be registered.

Even if the result is incorrect, if you speak again while the destination call button is blinking, the previously inputted content, that is, the currently blinking data, will be erased and the currently input result will be displayed blinking.

This allows corrections to be made until the correct result is obtained, and since the elevator user can operate the elevator by voice and vision, it becomes possible to carry out the original voice recognition process, which is aimed at eliminating the need to operate buttons or the like. Furthermore, since the user can understand the difference between a normal input and an inoperable input due to the difference in blinking cycle, the user can correct errors in the commands he or she has uttered.

Although the above description has been made regarding the destination call button in the hall, the same expression can be expressed with call buttons in the ascending direction and descending direction, such as in a normal elevator. The same applies to the call button, door open/close button, monthly extension button, etc. in the elevator car.

Next, although the notification of the voice recognition result was explained by flashing the registration button, it can also be expressed by changing the color of the registration button light.

It can also be represented by a variable character display device such as liquid crystal or plasma.

〔Effect of the invention〕

In the present invention, a provisional display of the recognition result is notified in response to a voice command, and if the result is determined to be normal, the operation is performed, and if it is abnormal, it can be easily corrected by speaking again, thereby reducing service degradation in elevator operation. It has the advantage of being able to be operated by voice without having to ask.

Furthermore, even if the voice command is incorrect, it is always notified through the display, so the user can operate the system without hesitation.

[Brief explanation of drawings]

FIG. 1 is a block diagram according to the present invention, FIG. 2 is a flowchart according to the present invention, FIG. 3 is a flowchart of the speech recognition section according to the present invention, FIG. 4 is an external view of the embodiment, and FIG. 5 is an external view of the conventional device. Figure 6 is a block diagram of the conventional device, and Figure 7 is a block diagram of the conventional device.
The figure is a speech waveform diagram of the speech recognition processing process. 3... Destination call, 4... Microphone, 5...
・・CPU, 6- A/D conversion section,
7... Band pass filter section, 8... Voice section detection section, 9 - Sampling section,
10...Dictionary storage unit, 11...Operation output unit, 1
2...Program storage section, 13.Elevator control section, 14.15.021.Gain controller 1 to roller section, output section to be called, 16.OK lamp, reject lamp. recognition result notification unit,

Claims (1)

[Claims] A microphone for recognizing voices in an elevator that recognizes and operates the voice of a passenger in an elevator car or hall, a means for distinguishing between external noise and voice commands, a means for notifying voice recognition results for a certain period of time, and a means for notifying An information input device for an elevator, comprising means for prioritizing voice commands, and allowing passengers to judge and correct voice recognition results.