JPH0367890A - Voice control device of crane - Google Patents

Abstract

PURPOSE:To improve a normal recognition factor by recognizing a command word being based on a degree of difference between a plurality of standard patterns, in which a recognition condition of a voice recognizing device is multi- registered, and an input characterized pattern. CONSTITUTION:In a voice recognizing device 6 of the ground set up device A, a plurality of standard patterns are registered in each command word, and at the time of operation, a characterized pattern of voice of the operator command word, voice-input from a portable moving station 2, is compared and collated with an amplitude pattern in each frequency with a degree of difference zero when perfect agreement is obtained. In the case of difference, the command word is recognized being based on a degree of difference, next deciding whether the command word, recognized by a control device 7, agrees or not with a predetermined command order based on a crane operation rule, when the decision is right, a predetermined control is performed by radio transmission to an on-crane set up device B. Thus, misoperation is prevented.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a crane voice control device for remotely controlling a crane using command words issued by an operator.

[Prior art]

In recent years, various types of crane voice control devices using voice recognition devices have been developed in order to rationalize and save labor in crane equipment operations.

For example, as disclosed in Japanese Unexamined Patent Application Publication No. 61-267689, an operator carries a mobile radio equipped with a microphone, and the command words uttered by the operator are transmitted to the voice via an audio radio device. transmitted to a recognition device. In the voice recognition device, the standard voice patterns of multiple types of command words are registered in advance by the operator, and these and the characteristic patterns of the voice of the command words uttered by the operator during driving are combined into 11 ((combined) The standard pattern is recognized. When the command word is recognized by the standard pattern, a signal corresponding to the command content is output to the crane control device, which controls the crane drive unit and responds to the command content. The vehicle is configured to perform a specific driving operation.

[Problem to be solved by the invention]

By the way, in the above-mentioned devices, the most important thing is the command word recognition ability of the voice recognition device. Since the environment in which the device is used is often high noise, it is necessary to reliably recognize the command word uttered by the commander without making a mistake. - The device mentioned above has a function of answering the command word once recognized to the operator, and after confirming this, the operator issues the command word to actually start the drive. If you set the recognition conditions of the speech recognition device very strictly to prevent misidentification due to noise, etc.
Due to subtle differences in the operator's utterances, if the corresponding command word is not registered, it is determined to be rejected, causing a problem that the operation cannot be started. On the other hand, if the recognition conditions are set relatively roughly, the voice recognition device is likely to misrecognize the command word, and the operator may have to repeat the command until a matching command word is answered. If you miss or mishear the answer bank and start the operation, there is also the problem that an unintended driving operation will be performed.

For this reason, the conventional device suffers from frequent rejections and malfunctions, has low safety and operability, and is hindered from being put to practical use.

Furthermore, as a measure to prevent malfunctions, it is considered effective to utilize the regularity of crane operation. Although there is some description regarding this in the above-mentioned Japanese Patent Application Laid-open No. 61-267689, it lacks specific description and is not clarified.

The present invention has been made in view of the above circumstances, and it sets a judgment criterion based on the degree of difference from the standard Bakun in the recognition conditions of the voice recognition device, and also prevents command words that are misrecognized due to the order of commands according to crane operation rules. The purpose of this invention is to provide a highly practical crane voice control device that reduces the rejection rate and misidentification rate by detecting the following.

rI! ! ! [Means for Solving the Problems] The crane voice control device according to the present invention compares the input characteristic pattern with a standard pattern of voice of a plurality of command words registered in advance by a specific speaker, and determines a standard pattern that matches. In a crane voice control device equipped with a voice recognition device that recognizes a pattern and outputs a signal corresponding to the standard pattern to the crane control section, the crane voice control device includes a registration means for multiple-registering standard patterns for each command word, and a registering means for multiple-registering standard patterns for each command word, and a registration means for multiple registering standard patterns for each command word. a dissimilarity calculation means for separately calculating the dissimilarity between the characteristic pattern and the standard pattern of each command word registered in the registration means; The degree of difference calculated by the degree of difference calculation means is 1.
A selection means for selecting the lowest first standard pattern and the second lowest standard pattern, and a ratio 1 for comparing the command words corresponding to the first and second standard patterns selected by the selection means. a biting means; a first recognition means that recognizes the first standard patterns as matching when the command words of the second standard patterns are equal and the degree of difference of the first standard patterns is smaller than the first standard value; 1. When the command words of the second standard patterns are different, the degree of difference of the first standard pattern is smaller than the second & quasi-value which is lower than the 1st jJiyp value, and the degree of difference of the first and second standard patterns is The first invention is characterized by comprising a second recognition means for recognizing the first standard pattern as a match when the difference is greater than a third reference value lower than the second ii4B pattern, and a storage means for storing a standard pattern that is recognized by the user and a corresponding signal is output; and a standard pattern to be recognized next from the standard pattern stored in the storage means is estimated based on a command order according to a predetermined driving rule. and means for refusing to output a signal to the corresponding crane control unit when a standard pattern other than the standard pattern estimated by the estimation means is next recognized. The characteristics of

[Effect]

In the crane voice control device according to the first invention, standard patterns are multiple-registered for each command word by a specific speaker in advance. When the operator utters a command word, the degree of difference between this characteristic pattern and the multiple registered standard pattern is calculated separately, and the standard pattern with the lowest degree of difference calculated among them and the second standard pattern are calculated. The standard patterns for which the lowest degree of dissimilarity has been calculated, in other words, the standard patterns with the highest and second highest degrees of similarity to the feature pattern are selected, respectively, and their corresponding command words are compared. Here, if the command words are equal, and the degree of difference of the first standard pattern is smaller than the first reference value, it is assumed that the first standard pattern matches the voice pattern of the command word uttered by the operator. Recognized. In other words, even if the first and second standard patterns are equal, if the degree of difference of the second standard pattern is greater than the first reference value, it is rejected as there is no matching standard pattern. On the other hand, even if the command words corresponding to the first and second standard patterns are different, the difference degree of the first standard pattern is smaller than the second reference value which is lower than the first base IJ= value, and the first and second The difference in the degree of dissimilarity between the standard patterns is the second
When the first standard pattern is greater than the third standard value, which is lower than the standard value, the first standard pattern is recognized as matching the voice characteristic pattern of the command word uttered by the operator.

In the crane voice control device according to the second invention, a standard pattern that has been previously recognized and a corresponding signal has been output is stored, and the standard pattern that is next recognized from this standard pattern is in a command order according to a predetermined operation rule. Estimated based on If the next recognized standard pattern is other than the estimated standard pattern, a signal corresponding to the recognized display pattern is not output to the crane control unit.

〔Example〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be specifically described below based on drawings showing embodiments thereof. FIG. 1 is a block diagram showing the configuration of a crane voice control device according to the present invention.

This embodiment shows an example of application to an overhead traveling crane, which is roughly divided into a mobile phone 1 equipped with a portable mobile station 23 carried by the operator, and transmits and receives signals by radio to the portable mobile station 2. It is composed of a ground-mounted device A and a crane-mounted device B that performs wireless signal transmission and reception with the ground-mounted device A.

The handphone 1 includes a voice input microphone 1a and a voice response phone 1b, and is worn on the head of an operator. The operator's command input into the voice input microphone la is transmitted from the portable mobile station 2 as a radio wave in the form of a voice signal, and is received by the voice input system antenna 3 of the ground installation device A.

The audio input antenna 3 constitutes an audio wireless device together with the antenna coupler 4 and the audio input fixed station 5, and the audio signal received by the audio input antenna 3 is transmitted to the antenna coupler 4 and the audio input fixed station 5. are transmitted to the speech recognition device 6 via the respective channels.

The voice recognition device 6 is, for example, manufactured by NEC Corporation (DP-300 manufactured by 11).
0, etc. can be used, and multiple types (three or more types) of standard speech patterns for each command word can be multiple-registered, and the operator can multiple-register multiple types of each command word in advance. This is to deal with the fact that when operating a crane, the operator cannot always utter the same voice pattern as the standard pattern in which the command word has been registered. By registering standard patterns of different types, command words uttered by a specific operator can be easily recognized.

The voice recognition device 6 is connected to a control device 7, and the control device 7 stores a command order based on the command word recognition conditions and crane operation rules, which will be described later, by the voice recognition device 6.
It has the function of performing final command word recognition using these and determining whether or not to actually perform the action corresponding to the recognized command word.

The control device 7 is connected to a radio ground station 8, and the radio ground station 8 is connected to an antenna coupler 9 equipped with a control output system antenna 10. The operation output signal corresponding to the command word that is finally recognized and permitted to operate by the control device 7 is transmitted from the operation output system antenna 10, and is transmitted from the antenna connected to the radio station 12 of the crane on-board device B. It is received at 11. The radio station 12 is connected to a sequencer 13,
The sequencer 13 inputs the operation output signal as a digital signal, performs logic processing on operating conditions and operating time,
Crane control the final operation output! Output to 14.

The crane control panel 14 inputs the final operation output and outputs a drive signal to a corresponding drive section (not shown).

In addition, while the control device 7 outputs an operation output signal, the control device 7 outputs the result of acceptance or rejection of the command word, that is, if the driving operation is permitted, the command word is pronounced, and if the driving action is rejected, a term such as an error is voiced by a speech synthesizer (not shown). This is output as a signal, transmitted from the antenna coupler 4 and the audio input system antenna 3, received by the portable mobile station 2, and sent to the operator from the audio response phone 1b of the headset 1. Now, in the apparatus of the present invention configured as described above, the recognition conditions by the speech recognition device 6 and the control device 7 will be explained first. As mentioned above, the voice recognition device 6 has a plurality of standard patterns registered for each command word, and these standard patterns are compared with the voice characteristic pattern of the command word input during driving by using the amplitude pattern for each frequency. This is done by comparing. Here, when the standard pattern and the characteristic pattern completely match, the degree of difference (hereinafter referred to as a score value) is zero, and the greater the degree of difference, the higher the score value. Note that this comparison can of course be performed based on similarity. Calculate the score values of all standard patterns for the feature patterns input in this way. Then, the standard pattern with the 2 lowest score value and the standard pattern with the second lowest score value are selected. Now, normally, if the command words corresponding to these standard patterns (hereinafter referred to as the first candidate term and second candidate term) are the same, the standard pattern of the first candidate term is recognized, and if they are different, it is said that they are unrecognizable. This may lead to rejection.

However, with this, the false recognition rate and rejection rate are extremely high, and it is difficult to maintain a normal recognition rate (95%).
It becomes impossible to apply voice control devices.

Therefore, in the apparatus of the present invention, the recognition conditions are set as shown below.

(1) In the case of the first candidate term - the second candidate term, the score value of the first candidate term <3. If so, the first candidate term is recognized.

If the score value of the first candidate term is ≧S, then it is a jet.

(2) If the first candidate term ≠ the second candidate term, the score value of the first candidate term <S2 and the score value of the first candidate term - the score value of the second candidate term> S3, the first candidate term Recognize.

If the score value of the first candidate term≧32 or the score value of the first candidate term−the score value of the second candidate term≦33, then select.

However, the score value is S,>S2>S3, and in this example, S, =100. S2=80. S3 =
Good judgment results were obtained when the value was set to 5.

That is, even if the first candidate term and the second candidate term are equal, if the score value of the edict for the first candidate is relatively high, such as S or more, there is a high possibility of misidentification, so a rejection determination is made. Furthermore, even if the first candidate term and the second candidate term are different, if the score value of the first candidate term is lower than 32 and the difference between the score value of the first and second candidate terms is higher than 33, then is likely to be an edict for the first candidate, so the first candidate term is recognized without being rejected.

Now, when the command word is recognized as described above, the control device 7 next determines whether the recognized command word conforms to a predetermined order of commands based on crane operation rules.

The control device 7 is configured to store the standard pattern of the command word that has already been recognized and generated the corresponding operation output signal, and when the next standard pattern is recognized, the command word of the stored standard pattern is stored. The next group of command words is estimated from the group of command words, and it is determined whether the command word of the next recognized standard pattern is included in this group of command words, that is, the rationality of the driving rule is checked. If the command word of the next recognized standard pattern is included, that command word is used as the final recognized command word and a corresponding operation output signal is generated, and if the command word is not included, the standard pattern recognized By determining that the command word was misrecognized and prohibiting the generation of the operation output signal, misrecognition of the command word is suppressed and abnormal operation is prevented.

Hereinafter, a detailed explanation will be given using timing charts of operation operations shown in FIGS. 2 to 4.

FIG. 2 shows an example of a single operation. First, "Higashi" is uttered from the direction instruction word group of the command word group shown in FIG. 6, and when this is recognized, the crane control panel
4 is set to the command holding state of execution operation. Then, the word "start" that commands the start of execution is uttered, and when this is recognized, the execution operation begins.After this, if the recognized command word is in accordance with the command order, the operation will stop. However, if a command word that does not conform to the command order is erroneously recognized, or if an unrecognized word is input and rejected, the execution operation is stopped.

Here, the rationality of the driving rules is checked for ``Pahigashi'' and ``Start'', and the next command word that violates the order of commands according to the driving rules is the contradictory direction instruction term, ie, ``Nijini'', which commands westbound driving. ”, and if this is recognized, it is considered to be a misidentification, so it will be rejected.On the other hand, if a direction indicating word other than “nijiniji” or a stop word such as “tomato” is recognized, it will be permitted.In other words, Direction-indicating words other than ``Higashi'' and ``Nijiniji'' are permitted because they may be used in combination, and ``Higashi'', which is in the same direction as the direction in which you are driving, is permitted as a driving stop term in this case, just like ``Tomato''. be done.

Figure 3.4 shows an example of a compound operation, and in Figure 3G, after the execution operation is started with "Higashi" and "Start", the next command word that is recognized is "Makioroshi". Execution operation has been stopped due to misidentification or rejection. The command words judged as misidentified here are “start” 2 “tomato” “choi” “choi choi”
All direction indicating terms other than ``furetome'' and ``gaksun'' are recognized.For example, if ``gonamigawa'', which instructs southbound driving, is recognized, the direction indicating terms are consecutive along with ``makioroshi'', so it is determined to be a misrecognition and permission is granted. Not done.

In Fig. 4, after the execution operation has started, the lowering operation is also started by "Makioroshi" and "Start", and the execution and lowering operation are stopped due to a misrecognition or rejection at the next command word. There is. The command words that are judged to be misrecognized here are words other than ``Higashi,''``Makioroshi,'' 1 and ``Tomato.'' For example, even if ``Kitakita'', which instructs northbound driving, is recognized, it has already been determined that execution and lowering have been performed. Since the combined operation of
Since the inching commands such as the above are not executed until after one of the operations under execution has been stopped by "1. Male", etc., it is similarly determined to be a misrecognition.

FIG. 5 is an example of a system diagram of the order of commands summarizing the contents of the above-mentioned combined operation. First, each drive part of the crane is set to a driveable state by the "power-on state." Next, one of the direction instruction terms is uttered and recognized by "start," and operation in the recognized direction is started by "start." What is recognized and permitted in the next ■ is a direction indication term other than the same direction indication term as ■, and a direction indication term other than the conflicting direction indication term.When this is recognized and "start" is recognized next, the compound operation is started. is started. In ■, only the direction indicating words uttered with ■ or ■ and "tomato" are recognized, and if "tomato" is recognized, both the combined operations of ■ and ■ are stopped.■
If either one of the direction indicating words in ■ is recognized and permitted and only one of the driving is stopped, the recognition and permission in the next ■ will be the other direction indicating term that is not recognized in ■ and,
``1 Male''``Brake'',``Choichoi'',``Fletome'', and ``Gaxun''.

If "tomato" is recognized here, the other operation is also stopped. In addition, when "brake" is recognized, braking is applied to the driving operation that was stopped due to the recognition of ■, and "choice",
When "choi choi", "furetome", and "gaxun" are recognized, the corresponding inching motion is performed within 10 seconds.

■When the direction indication term of the other person currently driving is recognized,
What is recognized and permitted in the next ■ is the above-mentioned inching operation term and "brake", and when "brake" is recognized, similarly ■
Braking is applied to the stopped driving operation based on the recognition of
When the time is longer than 0 seconds, the initial power-on state is restored.

Table 1 below shows the test results of a voice recognition test based on the above-mentioned operating rule rationality check conducted at a small diameter pipe factory and a cold pipe factory.The noise level was higher at the cold pipe factory. big.

(Leaving space below) 9 Table 1 0 Table 2 As can be seen, the number of misidentifications detected by the rationality check in each factory was 0.3%, and it can be said that it is effective in reducing misperceptions.

Table 2 shows the comparison results of the speech recognition ability of the conventional device and the device of the present invention. The test results are shown respectively.

(Left below) As is clear from the table, in the device of the present invention, the rejection rate and false recognition rate are significantly reduced, so the normal recognition rate is 98%, which exceeds 95%, which is considered to have no practical problems.
It has been raised to .

The reason why the normal recognition rate is approximately 1% lower in the cold pipe factory than in the small diameter pipe factory is mainly due to the difference in the frequency of occurrence of impulsive sounds during utterance. Therefore, in order to prevent the occurrence of rejection due to noise input, it is necessary to set the voice input operation parameters for the voice input bag W6 according to the noise conditions of the place where it is applied. For example, in order to avoid falsely detecting a '<5'+ impact sound as the start of audio input, take advantage of the fact that the input time is relatively short for the impact sound, which has a large peak, and eliminate it by setting a high integral value. can do.

In addition, the command words in this example are ``Nijiniji'',
Because we use easy-to-identify terms such as "゛mina ξ gawa",
It is effective in improving the normal recognition rate, and it is also difficult to input the end of beats for words that start with a voiceless fricative, such as "Higashi" start, and is likely to be rejected, but this can be countered by changing "Higashi" to "East", for example. It is also possible.

In this embodiment, the main devices such as the voice recognition device 6 and the control device 7 are installed on the ground, but they may be installed on the crane machine instead. However, the crane machine 2 is in a bad environment due to vibrations and temperature rises during crane operation, so it is desirable to install it on the ground side, where failures and malfunctions can be avoided and maintenance management is easy. In addition, when the main equipment is installed on the ground side, the crane operation output signal is transmitted to the top of the crane machine by wireless, but if an abnormality in the transmitted data or an abnormality in the wireless device occurs, it will be automatically transmitted. It is a good idea to take safety measures to make an emergency stop.

〔effect〕

As described above, in the crane voice control device according to the present invention, the recognition conditions of the voice recognition device are first determined based on the degree of difference between a plurality of multiple registered standard patterns and an input characteristic pattern. Therefore, command words that were conventionally recognized incorrectly can be rejected, and command words that were rejected can now be recognized, preventing malfunctions and improving the normal recognition rate.

In addition, the corresponding operation is ultimately permitted or rejected depending on whether the recognized command word conforms to the order of commands based on the crane operation rules, so malfunctions due to misrecognition can be reliably prevented. .

In this way, the device of the present invention can accurately reject unnecessary sounds and also detect false recognition, making it possible to significantly reduce the rejection rate and false recognition rate3, greatly improving safety and operability. improved,
The present invention has excellent effects such as being fully practical.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of the crane voice control device according to the present invention, FIGS. 2 to 4 are timing charts of operation operations, and FIG. 5 is a command sequence diagram for combined operation. FIG. 6 is an explanatory diagram of driving command terms.

Claims (1)

[Claims] 1. A standard pattern of speech of a plurality of command words registered in advance by a specific speaker is compared with the input characteristic pattern, a matching standard pattern is recognized, and this standard pattern is used. A crane voice control device equipped with a voice recognition device that outputs a corresponding signal to a crane control unit, a registration means for multiple-registering a standard pattern for each command word, and a characteristic pattern of an input signal and registered by the registration means. a dissimilarity calculation means that calculates the dissimilarity of the standard patterns of each command word separately; and a dissimilarity calculation means that calculates the dissimilarity of the standard patterns of each command word separately; A selection means for selecting the lowest first standard pattern and the second lowest standard pattern, and a comparison for comparing command words corresponding to the first and second standard patterns selected by the selection means. a means for recognizing the first standard pattern as matching when the command words of the first and second standard patterns are equal and the degree of difference of the first standard pattern is smaller than a first reference value; the recognition means, and when the command words of the first and second standard patterns are different, the degree of difference of the first standard pattern is smaller than a second reference value which is lower than the first reference value, and the first and second standard patterns are different from each other; and second recognition means that recognizes the first standard pattern as matching when the difference in the degree of difference between the standard patterns is greater than a third reference value that is lower than the second reference value. Crane voice control device. 2. Compare the input characteristic pattern with the standard voice patterns of multiple command words registered in advance by a specific speaker, recognize the matching standard pattern, and send the signal corresponding to this standard pattern to the crane control. The crane voice control device includes a voice recognition device that outputs a voice signal to a storage device, and includes a storage device for storing a standard pattern previously recognized and a corresponding signal output, and a storage device for storing a standard pattern that is recognized next from the standard pattern stored in the storage device. an estimating means for estimating a standard pattern based on a command sequence according to a predetermined operation rule; A crane voice control device characterized by comprising means for rejecting the output of a signal.