CN109949816A - Robot voice processing method and processing device, cloud server - Google Patents
Robot voice processing method and processing device, cloud server Download PDFInfo
- Publication number
- CN109949816A CN109949816A CN201910114891.6A CN201910114891A CN109949816A CN 109949816 A CN109949816 A CN 109949816A CN 201910114891 A CN201910114891 A CN 201910114891A CN 109949816 A CN109949816 A CN 109949816A
- Authority
- CN
- China
- Prior art keywords
- recognition result
- semantics recognition
- priority
- service provider
- audio data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 22
- 230000003993 interaction Effects 0.000 claims description 27
- 238000012163 sequencing technique Methods 0.000 claims description 16
- 235000013399 edible fruits Nutrition 0.000 claims description 6
- 238000000034 method Methods 0.000 abstract description 33
- 230000008569 process Effects 0.000 abstract description 28
- 238000010586 diagram Methods 0.000 description 9
- 238000012544 monitoring process Methods 0.000 description 6
- 230000004044 response Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
Landscapes
- Telephonic Communication Services (AREA)
Abstract
This application discloses a kind of robot voice processing method and processing devices, cloud server.This method includes obtaining the audio data acquired by robot；The audio data to multiple voice service provider server data-interfaces are forwarded to be handled；And receive by multiple voice service provider server data-interfaces return select the semantics recognition result to the semantics recognition result of the audio data and according to pre-set priority after issue.Speed slower technical problem when present application addresses robots for speech processes.Combine the voice semantic processes of multiple voice service quotient platforms as a result, the therefrom more accurate result of selection semantic results and more timely as a result, to more meet feedback expected from user with reaching in accuracy in speed by the application.
Description
Technical field
This application involves robots, speech processes field, in particular to a kind of robot voice processing method and dress
It sets, cloud server.
Background technique
Robot voice is answered, and refers to the process of progress interactive voice between robot and user.
Inventors have found that due to network or hardware problem, robot is slower for semantics recognition speed.Further, machine
The accuracy rate for the semantic understanding content that device people returns is poor, user can not be accurately obtained expected answer.
Speed slower problem when for robot in the related technology for speech processes not yet proposes effective solution at present
Certainly scheme.
Summary of the invention
The main purpose of the application is to provide a kind of robot voice processing method and processing device, cloud server, with solution
Speed slower problem when certainly robot is for speech processes.Can significantly improve robot voice answer result reasonability with
And reduce the delay of semantic understanding.
To achieve the goals above, according to the one aspect of the application, a kind of robot voice processing method is provided, is used
In acquisition speech recognition result.
Robot voice processing method according to the application includes: the audio data for obtaining and being acquired by robot；Forwarding
The audio data to multiple voice service provider server data-interfaces are handled；And it receives by multiple voices
Service provider server data-interface return to the semantics recognition result of the audio data and selected according to pre-set priority
It is issued after selecting out the semantics recognition result.
Further, it receives by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and select according to pre-set priority to issue after the semantics recognition result include: to receive by multiple institutes
What predicate sound service provider server data-interface returned knows to the semantics recognition result of the audio data and according to semanteme
Other result accuracy rate priority processing grade selects the semantics recognition result；And/or it receives and is provided by multiple voice services
Quotient's server data interface return to the semantics recognition result of the audio data and according to semantics recognition result return speed
Priority processing grade selects the semantics recognition result.
Further, it receives by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and while selecting the semantics recognition result according to semantics recognition result return speed priority also wrap
It includes: the semantics recognition result to the audio data returned according to each voice service provider server data-interface
Corresponding high priority processing grade is configured when return speed is fast；According to the user's request received, preferentially it is forwarded to described high preferential
The corresponding voice service provider server data-interface of process level.
Further, the semantics recognition result to the audio data and institute's predicate is selected according to pre-set priority
Adopted recognition result includes: to set low excellent for the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy
First process level；By the specific returned data of semantics recognition result, it is set as high priority processing grade；It is asked according to the user received
It asks, returns to user according to high priority processing grade.
Further, it receives by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and select according to pre-set priority issue after the semantics recognition result after further include: record institute
Predicate justice recognition result, and mark problem and answer；If being unsatisfactory for can be used for interacting field with user to the answer of described problem
The condition of scape then reduces the processing priority to described problem or the answer；And if the answer to described problem is full
It can be used for the condition with user's interaction scenarios completely, then improve the processing priority of described problem or the answer.
To achieve the goals above, according to the another aspect of the application, a kind of robot voice processing unit is provided, is used
In acquisition speech recognition result.
Robot voice processing unit according to the application includes: acquisition module, for obtaining through robot acquisition
Audio data；Forward modulus block, for forward the audio data to multiple voice service provider server data-interfaces into
Row processing；And receive and issue module, it receives by multiple voice service provider server data-interfaces returns to institute
It states the semantics recognition result of audio data and is issued after selecting the semantics recognition result according to pre-set priority.
Further, it includes: the first reception processing unit that the reception, which issues module, for receiving by multiple voices
Service provider server data-interface return to the semantics recognition result of the audio data and according to semantics recognition result
Accuracy rate priority processing grade selects the semantics recognition result；And/or second receive processing unit, for receive by multiple
The voice service provider server data-interface return to the semantics recognition result of the audio data and according to semanteme
Recognition result return speed priority processing grade selects the semantics recognition result；Speed-priority configuration unit, for according to every
The semantics recognition result return speed to the audio data that a voice service provider server data-interface returns
Corresponding high priority processing grade is configured when fast；Speed-priority processing unit, for according to the user's request received, preferential forwarding
To the corresponding voice service provider server data-interface of the high priority processing grade.
Further, to issue module include: the first sequencing unit for the reception, for by semantics recognition result it is uncertain or
The fuzzy returned data of person's semantics recognition result, is set as low priority processing grade；Second sequencing unit is used for semantics recognition knot
The specific returned data of fruit is set as high priority processing grade, the first return unit, for pressing according to the user's request received
User is returned to according to high priority processing grade.
Further, the reception issues module, comprising: marking unit, for recording the semantics recognition as a result, and marking
Note problem and answer；Third sequencing unit is unsatisfactory for can be used for the item with user's interaction scenarios for the answer to described problem
When part, the processing priority to described problem or the answer is reduced；And the 4th sequencing unit, for described problem
When answer satisfaction can be used for the condition with user's interaction scenarios, the processing priority of described problem or the answer is improved.
To achieve the goals above, according to the another aspect of the application, a kind of cloud server is provided, including described
Robot voice processing unit.
Robot voice processing method and processing device, cloud server in the embodiment of the present application pass through machine using acquisition
The mode of the audio data of people's acquisition, by forwarding the audio data to multiple voice service provider server data-interfaces
It is handled, has reached and received by multiple voice service provider server data-interfaces returns to the audio data
Semantics recognition result and the purpose issued after the semantics recognition result is selected according to pre-set priority, to realize whole
Slow technology is asked when closing, the technical effect of optimization speech processes result, and then solving robot for speech processes
Topic.
Detailed description of the invention
The attached drawing constituted part of this application is used to provide further understanding of the present application, so that the application's is other
Feature, objects and advantages become more apparent upon.The illustrative examples attached drawing and its explanation of the application is for explaining the application, not
Constitute the improper restriction to the application.In the accompanying drawings:
Fig. 1 is the robot voice processing method flow diagram according to the application first embodiment；
Fig. 2 is the robot voice processing method flow diagram according to the application second embodiment；
Fig. 3 is the robot voice processing method flow diagram according to the application 3rd embodiment；
Fig. 4 is the robot voice processing method flow diagram according to the application fourth embodiment；
Fig. 5 is the robot voice processing method flow diagram according to the 5th embodiment of the application；
Fig. 6 is the robot voice processing device structure diagram according to the application first embodiment；
Fig. 7 is the robot voice processing device structure diagram according to the application second embodiment；
Fig. 8 is the robot voice processing device structure diagram according to the application 3rd embodiment；
Fig. 9 is the robot voice processing device structure diagram according to the application fourth embodiment.
Specific embodiment
In order to make those skilled in the art more fully understand application scheme, below in conjunction in the embodiment of the present application
Attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only
The embodiment of the application a part, instead of all the embodiments.Based on the embodiment in the application, ordinary skill people
Member's every other embodiment obtained without making creative work, all should belong to the model of the application protection
It encloses.
It should be noted that the description and claims of this application and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to embodiments herein described herein.In addition, term " includes " and " tool
Have " and their any deformation, it is intended that cover it is non-exclusive include, for example, containing a series of steps or units
Process, method, system, product or equipment those of are not necessarily limited to be clearly listed step or unit, but may include without clear
Other step or units listing to Chu or intrinsic for these process, methods, product or equipment.
It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Robot voice processing method in the application, for obtaining speech recognition result, comprising: acquisition passes through robot
The audio data of acquisition；The audio data to multiple voice service provider server data-interfaces are forwarded to be handled；With
And receive the semantics recognition knot to the audio data returned by multiple voice service provider server data-interfaces
Fruit simultaneously issues after selecting the semantics recognition result according to pre-set priority.In this application by combining multiple voice services
The voice semantic processes of quotient's platform as a result, therefrom the more accurate result of selection semantic results and more timely as a result, thus
Reach in speed and in accuracy and more meets feedback expected from user.
As shown in Figure 1, this method includes the following steps, namely S102 to step S106:
Step S102 obtains the audio data acquired by robot；
The audio data collected by the robot is obtained in server beyond the clouds.The robot terminal not
It needs to handle audio data, be requested according to the user that the robot receives, it beyond the clouds can be real-time in server
Obtain the audio data acquired by the robot.
Specifically, synchronous with cloud server after the audio data being carried out acquisition in real time as voice data stream,
The audio data can be stored in the cloud server.
It should be noted that cloud server preferred deployment, in the server in cloud, which can be in original machine
Increase individual cloud server on the basis of personal data server.
Step S104 forwards the audio data to multiple voice service provider server data-interfaces to be handled；
The audio data that cloud server will acquire externally is forwarded.
Specifically, transit server sends out audio data described to the voice service provider server beyond the clouds.It is described
Voice service provider server includes at least 2, can be to be multiple, not for voice service in embodiments herein
The number of provider server is specifically limited, as long as requirement can be forwarded.
Cloud server forwards the audio data to multiple voice service provider server data-interfaces to carry out voice
Identification is converted into text, carries out semantics recognition processing etc. according to text.The processing of this part is taken by specific voice service provider
Business device executes, and is served only for obtaining processing result from the voice service provider server in embodiments herein.
Step S106 is received by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and issue after according to pre-set priority selecting the semantics recognition result.
The cloud server is received by the data-interface return of multiple voice service provider servers to institute
After the semantics recognition processing result for stating audio data, the semantic knowledge is selected according to pre-set priority in the cloud server
Not as a result, being then issued to the robotic end for receiving user's request according to the semantics recognition result selected.
The cloud server according to pre-set priority selects the semantics recognition result and can pay close attention to recognition result
The no accurate and whether enough quick responses of recognition result.And it can and knowledge whether accurate by the monitoring recognition result
The other whether sufficiently fast speed response of result, it is available in multiple voice service provider servers, which is respectively suitable for
A little scenes.
Further, by passing through network interface after the completion of handling the semantic results beyond the clouds in server beyond the clouds again
Mode the result returned according to default priority processing grade returned to robot handled.To realize speech recognition result
The accuracy of semanteme and timely response speed.
It can be seen from the above description that the application realizes following technical effect:
Robot voice processing method and processing device, cloud server in the embodiment of the present application pass through machine using acquisition
The mode of the audio data of people's acquisition, by forwarding the audio data to multiple voice service provider server data-interfaces
It is handled, has reached and received by multiple voice service provider server data-interfaces returns to the audio data
Semantics recognition result and the purpose issued after the semantics recognition result is selected according to pre-set priority, to realize whole
Slow technology is asked when closing, the technical effect of optimization speech processes result, and then solving robot for speech processes
Topic.
According to the embodiment of the present application, as preferred in the present embodiment, taken as shown in Fig. 2, receiving by multiple voices
Business provider server data-interface return to the semantics recognition result of the audio data and selected according to pre-set priority
It is issued after the semantics recognition result out and includes:
Step S202 is received by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and select the semantics recognition result according to semantics recognition result accuracy rate priority processing grade；
It receives by multiple voice service provider server data-interfaces returns in the cloud server to institute
After the semantics recognition result for stating audio data, according to semantics recognition result accuracy rate priority processing grade, the semantic knowledge is selected
Other result.That is, being accurately preferably five-star processing priority with semanteme.
For example, if the cloud server receives the semantics recognition that the voice service provider server A is returned
As a result a is more accurate than the semantics recognition result a returned from the voice service provider server B, then the voice is selected to take
Be engaged in provider server A.
Specifically, after robot terminal is responsible for carrying out audio collection, audio stream is uploaded into cloud server in real time
In speech processing system, the audio stream that acquisition is received in the speech processing system of cloud server is forwarded to the language of more families in real time
Reason and good sense solution platform, while monitoring semantics recognition result and the semantic speed returned that more semantic platforms return.If semantic
It is higher to return to accuracy rate, then can pay the utmost attention to select.
And/or step S204, it receives by multiple voice service provider server data-interfaces returns to described
The semantics recognition result of audio data simultaneously selects the semantics recognition according to semantics recognition result return speed priority processing grade
As a result.
It receives by multiple voice service provider server data-interfaces returns in the cloud server to institute
After the semantics recognition result for stating audio data, according to semantics recognition result return speed priority processing grade, the semanteme is selected
Recognition result.That is, being preferably five-star processing priority with return speed.
For example, if the cloud server receives the semantics recognition that the voice service provider server A is returned
As a result a is faster than the semantics recognition result a returned from the voice service provider server B, then selects the voice service
Provider server A.
Specifically, after robot terminal is responsible for carrying out audio collection, audio stream is uploaded into cloud server in real time
In speech processing system, the audio stream that acquisition is received in the speech processing system of cloud server is forwarded to the language of more families in real time
Reason and good sense solution platform, while monitoring semantics recognition result and the semantic speed returned that more semantic platforms return.If semantic
Return speed is very fast, then can pay the utmost attention to select.
Preferably, it receives by multiple voice service provider server data-interfaces returns to the audio data
Semantics recognition result and select the same of the semantics recognition result according to semantics recognition result return speed priority processing grade
When can consider to receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
Recognition result simultaneously selects the semantics recognition result according to semantics recognition result accuracy rate priority processing grade.
According to the embodiment of the present application, as preferred in the present embodiment, taken as shown in figure 3, receiving by multiple voices
Business provider server data-interface return to the semantics recognition result of the audio data and returned according to semantics recognition result
Speed-priority grade is returned when selecting the semantics recognition result further include:
Step S302, according to each voice service provider server data-interface return to the audio data
Semantics recognition result return speed it is fast when configure corresponding high priority processing grade；
What the voice service provider server data-interface returned returns the semantics recognition result of the audio data
Return speed it is fast when be with return speed when configuring corresponding high priority processing grade be preferably treatment principle.
Step S304 requests according to the user received, is preferentially forwarded to the corresponding institute's predicate of the high priority processing grade
Sound service provider server data-interface.
According to the user's request received in the robotic end, server is preferably forwarded to the high preferential place beyond the clouds
It manages the corresponding voice service provider server data-interface of grade and carries out subsequent speech recognition processing.
Specifically, the voice when priority for returning the result speed conditions for semanteme is handled, in cloud server
Processing system can be counted according to the speed that each voice service provider returns the result, if corresponding voice service mentions
It is very fast for the semantic results return speed of quotient, just improve the priority that the voice service provider semanteme returns the result, behind language
The result of justice request preferentially uses the result of the voice service provider.
According to the embodiment of the present application, as preferred in the present embodiment, as shown in figure 4, described to the audio data
Semantics recognition result simultaneously selects the semantics recognition result according to pre-set priority and includes:
Step S402 sets low for the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy
Priority processing grade；
Low priority processing grade is set by the uncertain returned data of semantics recognition result, while the semanteme being known
The fuzzy returned data of other result is also configured as low priority processing grade.
It should be noted that or semantics recognition result uncertain for the semantics recognition result is obscured and can be built in advance
Vertical semantic base is to judge whether the semantics recognition result is determining or fuzzy.
The specific returned data of semantics recognition result is set high priority processing grade by step S404；
High priority processing grade is set by the specific returned data of semantics recognition result.
It should be noted that can explicitly pre-establish semantic base for the semantics recognition to judge the semanteme
Whether recognition result is specific.
Step S406 requests according to the user received, returns to user according to high priority processing grade.
According to the user's request received in robot terminal, beyond the clouds according to pre-set high preferential place in server
Reason grade result returns to user.
Specifically, for the processing of semantic results, the speech processing system in cloud server can screen the semanteme of return
As a result, arrange uncertain or relatively fuzzyyer answer for lowest priority, by the more explicitly row of answer be it is higher preferentially
Grade, then return to user according to priority as a result, and recorded in server beyond the clouds, when the cloud server collect
It makes marks after certain data to different answers, priority is reduced to minimum if answering and having regular logical mistake, answers mark
Standard thinks that mark improves priority.Further, it is also possible to carry out matching for different priorities according to label result according to different scenes
It sets.
According to the embodiment of the present application, as preferred in the present embodiment, taken as shown in figure 5, receiving by multiple voices
Business provider server data-interface return to the semantics recognition result of the audio data and selected according to pre-set priority
After being issued after the semantics recognition result out further include:
Step S502 records the semantics recognition as a result, and marking problem and answer；
After each cloud server is issued to the robot terminal semantic processes result, needed in server beyond the clouds
The semantics recognition is recorded as a result, and going out the problem in the semantics recognition result according to record result queue and corresponding answering
Case.
Step S504, if being unsatisfactory for can be used for reducing with the condition of user's interaction scenarios to the answer of described problem
To the processing priority of described problem or the answer；And
If being unsatisfactory for can be used for the condition with user's interaction scenarios to the answer of described problem, reduction pair can be corresponded to
Described problem or the processing priority of the answer.
Step S506 improves institute if the answer to described problem meets the condition that can be used for user's interaction scenarios
State the processing priority of problem or the answer.
If can satisfy the condition that can be used for user's interaction scenarios to the answer of described problem, then asked described in raising
Topic or the processing priority of the answer.
It should be noted that the answer satisfaction, which can be used for can be with the condition of user's interaction scenarios, refers to semantic processes knot
Whether fruit clear, completes, and whether clearly etc. whether semantic processes result meet logic, semantic processes result.It also needs to infuse
Meaning, answer satisfaction can be used for the condition of user's interaction scenarios can with problem in the semantic processes result and
Corresponding to reply whether meet user's interaction scenarios, answer contains day in the semantic processes result occurred when for example, ordering air ticket scene
Gas, then it is assumed that do not meet current interaction scenarios.For another example, answer contains city in the semantic processes result occurred when weather scene
Place name, then it is assumed that meet current interaction scenarios.
Specifically, for the processing of semantic results, the speech processing system in cloud server can screen the language of return
Adopted result such as can be processing record log, and uncertain or relatively fuzzyyer answer is arranged as lowest priority, will be compared
The specific row of answer is higher priority, then returns to user according to priority as a result, and server is remembered beyond the clouds
Record, makes marks to different answers after collecting certain data, and priority is reduced to minimum if answering and having regular logical mistake,
Answer standard thinks that mark improves priority.
According to the embodiment of the present application, as preferred in the present embodiment, robot voice processing method, for obtaining voice
Recognition result, which comprises
Obtain the audio data acquired by robot；
The audio data to multiple voice service provider server data-interfaces are forwarded to be handled；And receive by
Multiple voice service provider server data-interfaces return to the semantics recognition result of the audio data and according to
Pre-set priority issues after selecting the semantics recognition result.
Receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
Recognition result is simultaneously selected to issue after the semantics recognition result and includes: according to pre-set priority
Receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
Recognition result simultaneously selects the semantics recognition result according to semantics recognition result accuracy rate priority processing grade；
And/or it receives by multiple voice service provider server data-interfaces returns to the audio data
Semantics recognition result and select the semantics recognition result according to semantics recognition result return speed priority processing grade.
Receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
Recognition result and when selecting the semantics recognition result according to semantics recognition result return speed priority further include:
The semanteme of the audio data is known according to what each voice service provider server data-interface returned
Other result return speed configures corresponding high priority processing grade when fast；
According to the user's request received, preferentially it is forwarded to the corresponding voice service of the high priority processing grade and provides
Quotient's server data interface.
The semantics recognition result to the audio data simultaneously selects the semantics recognition knot according to pre-set priority
Fruit includes:
By the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy, it is set as low priority processing
Grade；
By the specific returned data of semantics recognition result, it is set as high priority processing grade；
According to the user's request received, user is returned to according to high priority processing grade.
Receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
After recognition result and selecting according to pre-set priority issues after the semantics recognition result further include:
The semantics recognition is recorded as a result, and marking problem and answer；
If being unsatisfactory for can be used for the condition with user's interaction scenarios to the answer of described problem, reduce to described problem
Or the processing priority of the answer；And
If meeting the condition that can be used for user's interaction scenarios to the answer of described problem, improve described problem or
The processing priority of the answer.
It should be noted that step shown in the flowchart of the accompanying drawings can be in such as a group of computer-executable instructions
It is executed in computer system, although also, logical order is shown in flow charts, and it in some cases, can be with not
The sequence being same as herein executes shown or described step.
According to the embodiment of the present application, a kind of for implementing the device of above-mentioned robot voice processing method, use is additionally provided
In obtaining speech recognition result, as shown in fig. 6, the device includes: to obtain module 10, for obtaining the sound acquired by robot
Frequency evidence；Forward modulus block 20, for forward the audio data to multiple voice service provider server data-interfaces into
Row processing；And receive and issue module 30, receive pair returned by multiple voice service provider server data-interfaces
The semantics recognition result of the audio data simultaneously issues after selecting the semantics recognition result according to pre-set priority.
It obtains in server in the acquisition module 10 of the embodiment of the present application and to be collected by the robot beyond the clouds
Audio data.It does not need to handle audio data in the robot terminal, the user received according to the robot
Request, can obtain the audio data acquired by the robot in real time beyond the clouds in server.
Specifically, synchronous with cloud server after the audio data being carried out acquisition in real time as voice data stream,
The audio data can be stored in the cloud server.
It should be noted that cloud server preferred deployment, in the server in cloud, which can be in original machine
Increase individual cloud server on the basis of personal data server.
The audio data that cloud server will acquire in the forwarding modulus block 20 of the embodiment of the present application carries out external
Forwarding.
Specifically, transit server sends out audio data described to the voice service provider server beyond the clouds.It is described
Voice service provider server includes at least 2, can be to be multiple, not for voice service in embodiments herein
The number of provider server is specifically limited, as long as requirement can be forwarded.
Cloud server forwards the audio data to multiple voice service provider server data-interfaces to carry out voice
Identification is converted into text, carries out semantics recognition processing etc. according to text.The processing of this part is taken by specific voice service provider
Business device executes, and is served only for obtaining processing result from the voice service provider server in embodiments herein.
The reception of the embodiment of the present application issues the reception of cloud server described in module 30 and is mentioned by multiple voice services
After the semantics recognition processing result to the audio data returned for the data-interface of quotient's server, in the cloud server
The semantics recognition, which is selected, according to pre-set priority connects as a result, being then issued to according to the semantics recognition result selected
Receive the robotic end of user's request.
The cloud server according to pre-set priority selects the semantics recognition result and can pay close attention to recognition result
The no accurate and whether enough quick responses of recognition result.And it can and knowledge whether accurate by the monitoring recognition result
The other whether sufficiently fast speed response of result, it is available in multiple voice service provider servers, which is respectively suitable for
A little scenes.
Further, by passing through network interface after the completion of handling the semantic results beyond the clouds in server beyond the clouds again
Mode the result returned according to default priority processing grade returned to robot handled.To realize speech recognition result
The accuracy of semanteme and timely response speed.
According to the embodiment of the present application, as preferred in the present embodiment, as shown in fig. 7, the reception issues module includes:
First receives processing unit 301, for receiving by multiple voice service provider server data-interfaces returns to institute
It states the semantics recognition result of audio data and selects the semantics recognition according to semantics recognition result accuracy rate priority processing grade
As a result；And/or second receive processing unit 302, for receive by multiple voice service provider server data-interfaces
Return to the semantics recognition result of the audio data and selected according to semantics recognition result return speed priority processing grade
The semantics recognition result；Speed-priority configuration unit 303, for according to each voice service provider server data
Interface return it is fast to the semantics recognition result return speed of the audio data when configure corresponding high priority processing grade；Speed
Priority processing unit 304, for it is corresponding described to be preferentially forwarded to the high priority processing grade according to the user's request received
Voice service provider server data-interface.
The first of the embodiment of the present application receives in processing unit 301 and receives in the cloud server by multiple voices
After the semantics recognition result to the audio data that service provider server data-interface returns, according to semantics recognition result
Accuracy rate priority processing grade selects the semantics recognition result.That is, being accurately preferably five-star processing with semanteme
Priority.
For example, if the cloud server receives the semantics recognition that the voice service provider server A is returned
As a result a is more accurate than the semantics recognition result a returned from the voice service provider server B, then the voice is selected to take
Be engaged in provider server A.
Specifically, after robot terminal is responsible for carrying out audio collection, audio stream is uploaded into cloud server in real time
In speech processing system, the audio stream that acquisition is received in the speech processing system of cloud server is forwarded to the language of more families in real time
Reason and good sense solution platform, while monitoring semantics recognition result and the semantic speed returned that more semantic platforms return.If semantic
It is higher to return to accuracy rate, then can pay the utmost attention to select.
The second of the embodiment of the present application receives in processing unit 302 and receives in the cloud server by multiple voices
After the semantics recognition result to the audio data that service provider server data-interface returns, according to semantics recognition result
Return speed priority processing grade selects the semantics recognition result.That is, being preferably five-star place with return speed
Manage priority.
For example, if the cloud server receives the semantics recognition that the voice service provider server A is returned
As a result a is faster than the semantics recognition result a returned from the voice service provider server B, then selects the voice service
Provider server A.
Specifically, after robot terminal is responsible for carrying out audio collection, audio stream is uploaded into cloud server in real time
In speech processing system, the audio stream that acquisition is received in the speech processing system of cloud server is forwarded to the language of more families in real time
Reason and good sense solution platform, while monitoring semantics recognition result and the semantic speed returned that more semantic platforms return.If semantic
Return speed is very fast, then can pay the utmost attention to select.
Preferably, it receives by multiple voice service provider server data-interfaces returns to the audio data
Semantics recognition result and select the same of the semantics recognition result according to semantics recognition result return speed priority processing grade
When can consider to receive the semanteme to the audio data returned by multiple voice service provider server data-interfaces
Recognition result simultaneously selects the semantics recognition result according to semantics recognition result accuracy rate priority processing grade.
Voice service provider server data-interface described in the speed-priority configuration unit 303 of the embodiment of the present application returns
Return it is fast to the semantics recognition result return speed of the audio data when be to return when configuring corresponding high priority processing grade
Speed-priority is treatment principle.
Basis is asked in the user that the robotic end receives in the speed-priority processing unit 304 of the embodiment of the present application
It asks, server is preferably forwarded to the corresponding voice service provider server data of the high priority processing grade and connects beyond the clouds
Mouth carries out subsequent speech recognition processing.
Specifically, the voice when priority for returning the result speed conditions for semanteme is handled, in cloud server
Processing system can be counted according to the speed that each voice service provider returns the result, if corresponding voice service mentions
It is very fast for the semantic results return speed of quotient, just improve the priority that the voice service provider semanteme returns the result, behind language
The result of justice request preferentially uses the result of the voice service provider.
According to the embodiment of the present application, as preferred in the present embodiment, as shown in figure 8, the reception issues module includes:
First sequencing unit 305, for being set as the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy
Low priority processing grade；Second sequencing unit 306, for being set as high preferential place for the specific returned data of semantics recognition result
Manage grade, the first return unit 307, for returning to user according to high priority processing grade according to the user's request received.
The uncertain returned data of the semantics recognition result is arranged in first sequencing unit 305 of the embodiment of the present application
For low priority processing grade, while the returned data that the semantics recognition result obscures is also configured as low priority processing grade.
It should be noted that or semantics recognition result uncertain for the semantics recognition result is obscured and can be built in advance
Vertical semantic base is to judge whether the semantics recognition result is determining or fuzzy.
The specific returned data of semantics recognition result is set in second sequencing unit 306 of the embodiment of the present application
High priority processing grade.
It should be noted that can explicitly pre-establish semantic base for the semantics recognition to judge the semanteme
Whether recognition result is specific.
According to the user's request received in robot terminal in first return unit 307 of the embodiment of the present application, in cloud
User is returned to according to pre-set high priority processing grade result in the server of end.
Specifically, for the processing of semantic results, the speech processing system in cloud server can screen the semanteme of return
As a result, arrange uncertain or relatively fuzzyyer answer for lowest priority, by the more explicitly row of answer be it is higher preferentially
Grade, then return to user according to priority as a result, and recorded in server beyond the clouds, when the cloud server collect
It makes marks after certain data to different answers, priority is reduced to minimum if answering and having regular logical mistake, answers mark
Standard thinks that mark improves priority.Further, it is also possible to carry out matching for different priorities according to label result according to different scenes
It sets.
According to the embodiment of the present application, as preferred in the present embodiment, as shown in figure 9, the reception issues module, wrap
It includes: marking unit 308, for recording the semantics recognition as a result, and marking problem and answer；Third sequencing unit 309, is used for
When being unsatisfactory for can be used for the condition with user's interaction scenarios to the answer of described problem, reduce to described problem or the answer
Processing priority；And the 4th sequencing unit 310, it can be used for and user's interaction scenarios for the answer satisfaction to described problem
Condition when, improve the processing priority of described problem or the answer.
It is semantic to be issued to the robot terminal for each cloud server in the marking unit 308 of the embodiment of the present application
After processing result, need to record the semantics recognition beyond the clouds in server as a result, and being gone out according to record result queue described
Problem and corresponding answer in semantics recognition result.
If being unsatisfactory for can be used for the answer of described problem and user in the third sequencing unit 309 of the embodiment of the present application
The condition of interaction scenarios can then correspond to the processing priority reduced to described problem or the answer.
If can satisfy to the answer of described problem can be used for and use in 4th sequencing unit 310 of the embodiment of the present application
The condition of family interaction scenarios then improves the processing priority of described problem or the answer.
It should be noted that the answer satisfaction, which can be used for can be with the condition of user's interaction scenarios, refers to semantic processes knot
Whether fruit clear, completes, and whether clearly etc. whether semantic processes result meet logic, semantic processes result.It also needs to infuse
Meaning, answer satisfaction can be used for the condition of user's interaction scenarios can with problem in the semantic processes result and
Corresponding to reply whether meet user's interaction scenarios, answer contains day in the semantic processes result occurred when for example, ordering air ticket scene
Gas, then it is assumed that do not meet current interaction scenarios.For another example, answer contains city in the semantic processes result occurred when weather scene
Place name, then it is assumed that meet current interaction scenarios.
Specifically, for the processing of semantic results, the speech processing system in cloud server can screen the language of return
Adopted result such as can be processing record log, and uncertain or relatively fuzzyyer answer is arranged as lowest priority, will be compared
The specific row of answer is higher priority, then returns to user according to priority as a result, and server is remembered beyond the clouds
Record, makes marks to different answers after collecting certain data, and priority is reduced to minimum if answering and having regular logical mistake,
Answer standard thinks that mark improves priority.
In another embodiment of the application, cloud server, including robot voice as mentioned processing are additionally provided
Device, wherein the realization principle and beneficial effect of the robot voice processing unit as described above, no longer go to live in the household of one's in-laws on getting married herein
It states.
Obviously, those skilled in the art should be understood that each module of above-mentioned the application or each step can be with general
Computing device realize that they can be concentrated on a single computing device, or be distributed in multiple computing devices and formed
Network on, optionally, they can be realized with the program code that computing device can perform, it is thus possible to which they are stored
Be performed by computing device in the storage device, perhaps they are fabricated to each integrated circuit modules or by they
In multiple modules or step be fabricated to single integrated circuit module to realize.In this way, the application be not limited to it is any specific
Hardware and software combines.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (10)
1. a kind of robot voice processing method, which is characterized in that for obtaining speech recognition result, which comprises
Obtain the audio data acquired by robot；
The audio data to multiple voice service provider server data-interfaces are forwarded to be handled；And
Receive the semantics recognition to the audio data returned by multiple voice service provider server data-interfaces
As a result it and after selecting the semantics recognition result according to pre-set priority issues.
2. robot voice processing method according to claim 1, which is characterized in that receive by multiple voice services
Provider server data-interface return to the semantics recognition result of the audio data and selected according to pre-set priority
It is issued after the semantics recognition result and includes:
Receive the semantics recognition to the audio data returned by multiple voice service provider server data-interfaces
As a result and according to semantics recognition result accuracy rate priority processing grade the semantics recognition result is selected；
And/or receive the language to the audio data returned by multiple voice service provider server data-interfaces
Adopted recognition result simultaneously selects the semantics recognition result according to semantics recognition result return speed priority processing grade.
3. robot voice processing method according to claim 2, which is characterized in that receive by multiple voice services
What provider server data-interface returned returns to the semantics recognition result of the audio data and according to semantics recognition result
When speed-priority grade selects the semantics recognition result further include:
The semantics recognition knot to the audio data returned according to each voice service provider server data-interface
Fruit return speed configures corresponding high priority processing grade when fast；
According to the user's request received, it is preferentially forwarded to the corresponding voice service provider clothes of the high priority processing grade
Business device data-interface.
4. robot voice processing method according to claim 1, which is characterized in that the language to the audio data
Adopted recognition result simultaneously selects the semantics recognition result according to pre-set priority and includes:
By the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy, it is set as low priority processing grade；
By the specific returned data of semantics recognition result, it is set as high priority processing grade；
According to the user's request received, user is returned to according to high priority processing grade.
5. robot voice processing method according to claim 1, which is characterized in that receive by multiple voice services
Provider server data-interface return to the semantics recognition result of the audio data and selected according to pre-set priority
After being issued after the semantics recognition result further include:
The semantics recognition is recorded as a result, and marking problem and answer；
If being unsatisfactory for can be used for the condition with user's interaction scenarios to the answer of described problem, reduce to described problem or
The processing priority of the answer；And
If the answer to described problem meets the condition that can be used for user's interaction scenarios, described problem or described is improved
The processing priority of answer.
6. a kind of robot voice processing unit, which is characterized in that for obtaining speech recognition result, described device includes:
Module is obtained, for obtaining the audio data acquired by robot；
Modulus block is forwarded, for forwarding at the audio data to multiple voice service provider server data-interfaces
Reason；And
Reception issues module, receives by multiple voice service provider server data-interfaces returns to the audio number
According to semantics recognition result and issue after according to pre-set priority selecting the semantics recognition result.
7. robot voice processing unit according to claim 6, which is characterized in that the reception issues module and includes:
First receives processing unit, for receiving by multiple voice service provider server data-interfaces returns to institute
It states the semantics recognition result of audio data and selects the semantics recognition according to semantics recognition result accuracy rate priority processing grade
As a result；
And/or second receive processing unit, is returned for receiving by multiple voice service provider server data-interfaces
Selected to the semantics recognition result of the audio data and according to semantics recognition result return speed priority processing grade it is described
Semantics recognition result；
Speed-priority configuration unit, for according to each voice service provider server data-interface return to described
The semantics recognition result return speed of audio data configures corresponding high priority processing grade when fast；
Speed-priority processing unit, for it is corresponding to be preferentially forwarded to the high priority processing grade according to the user's request received
The voice service provider server data-interface.
8. robot voice processing unit according to claim 6, which is characterized in that the reception issues module and includes:
First sequencing unit, for being arranged the returned data that semantics recognition result is uncertain or semantics recognition result is fuzzy
For low priority processing grade；
Second sequencing unit, for being set as high priority processing grade for the specific returned data of semantics recognition result,
First return unit, for returning to user according to high priority processing grade according to the user's request received.
9. robot voice processing unit according to claim 6, which is characterized in that the reception issues module, comprising:
Marking unit, for recording the semantics recognition as a result, and marking problem and answer；
Third sequencing unit when being unsatisfactory for can be used for the condition with user's interaction scenarios for the answer to described problem, reduces
To the processing priority of described problem or the answer；And
4th sequencing unit improves institute when can be used for the condition with user's interaction scenarios for the answer satisfaction to described problem
State the processing priority of problem or the answer.
10. a kind of cloud server, which is characterized in that including the described in any item robot voice processing of such as claim 6 to 9
Device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114891.6A CN109949816A (en) | 2019-02-14 | 2019-02-14 | Robot voice processing method and processing device, cloud server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114891.6A CN109949816A (en) | 2019-02-14 | 2019-02-14 | Robot voice processing method and processing device, cloud server |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109949816A true CN109949816A (en) | 2019-06-28 |
Family
ID=67007609
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910114891.6A Pending CN109949816A (en) | 2019-02-14 | 2019-02-14 | Robot voice processing method and processing device, cloud server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949816A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110675871A (en) * | 2019-09-25 | 2020-01-10 | 北京蓦然认知科技有限公司 | A kind of speech recognition method and device |
| CN111190715A (en) * | 2019-12-31 | 2020-05-22 | 杭州涂鸦信息技术有限公司 | Distribution scheduling method and system of product service, readable storage medium and computer |
| CN112820295A (en) * | 2020-12-29 | 2021-05-18 | 华人运通(上海)云计算科技有限公司 | Voice processing device and system, cloud server and vehicle |
| CN113053369A (en) * | 2019-12-26 | 2021-06-29 | 青岛海尔空调器有限总公司 | Voice control method and device of intelligent household appliance and intelligent household appliance |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2004055778A2 (en) * | 2002-12-13 | 2004-07-01 | Motorola Inc. | Method and apparatus for selective speech recognition |
| US20100004930A1 (en) * | 2008-07-02 | 2010-01-07 | Brian Strope | Speech Recognition with Parallel Recognition Tasks |
| CN103077718A (en) * | 2013-01-09 | 2013-05-01 | 华为终端有限公司 | Speech processing method, system and terminal |
| CN103117058A (en) * | 2012-12-20 | 2013-05-22 | 四川长虹电器股份有限公司 | Multi-voice engine switch system and method based on intelligent television platform |
| CN103413549A (en) * | 2013-07-31 | 2013-11-27 | 深圳创维－Rgb电子有限公司 | Voice interaction method and system and interaction terminal |
| CN103488529A (en) * | 2013-09-06 | 2014-01-01 | 上海骋娱传媒技术有限公司 | Method and equipment for controlling access of video resources |
| CN104795069A (en) * | 2014-01-21 | 2015-07-22 | 腾讯科技（深圳）有限公司 | Speech recognition method and server |
| CN108197191A (en) * | 2017-12-27 | 2018-06-22 | 神思电子技术股份有限公司 | A kind of scene of more wheel dialogues is intended to interrupt method |
-
2019
- 2019-02-14 CN CN201910114891.6A patent/CN109949816A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2004055778A2 (en) * | 2002-12-13 | 2004-07-01 | Motorola Inc. | Method and apparatus for selective speech recognition |
| US20100004930A1 (en) * | 2008-07-02 | 2010-01-07 | Brian Strope | Speech Recognition with Parallel Recognition Tasks |
| CN103117058A (en) * | 2012-12-20 | 2013-05-22 | 四川长虹电器股份有限公司 | Multi-voice engine switch system and method based on intelligent television platform |
| CN103077718A (en) * | 2013-01-09 | 2013-05-01 | 华为终端有限公司 | Speech processing method, system and terminal |
| CN103413549A (en) * | 2013-07-31 | 2013-11-27 | 深圳创维－Rgb电子有限公司 | Voice interaction method and system and interaction terminal |
| CN103488529A (en) * | 2013-09-06 | 2014-01-01 | 上海骋娱传媒技术有限公司 | Method and equipment for controlling access of video resources |
| CN104795069A (en) * | 2014-01-21 | 2015-07-22 | 腾讯科技（深圳）有限公司 | Speech recognition method and server |
| CN108197191A (en) * | 2017-12-27 | 2018-06-22 | 神思电子技术股份有限公司 | A kind of scene of more wheel dialogues is intended to interrupt method |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110675871A (en) * | 2019-09-25 | 2020-01-10 | 北京蓦然认知科技有限公司 | A kind of speech recognition method and device |
| CN110675871B (en) * | 2019-09-25 | 2021-03-05 | 北京蓦然认知科技有限公司 | Voice recognition method and device |
| CN113053369A (en) * | 2019-12-26 | 2021-06-29 | 青岛海尔空调器有限总公司 | Voice control method and device of intelligent household appliance and intelligent household appliance |
| CN111190715A (en) * | 2019-12-31 | 2020-05-22 | 杭州涂鸦信息技术有限公司 | Distribution scheduling method and system of product service, readable storage medium and computer |
| CN111190715B (en) * | 2019-12-31 | 2023-05-12 | 杭州涂鸦信息技术有限公司 | Distribution scheduling method and system for product service, readable storage medium and computer |
| CN112820295A (en) * | 2020-12-29 | 2021-05-18 | 华人运通(上海)云计算科技有限公司 | Voice processing device and system, cloud server and vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10162685B2 (en) | System and method for intelligent task management and routing | |
| CN109949816A (en) | Robot voice processing method and processing device, cloud server | |
| US7027586B2 (en) | Intelligently routing customer communications | |
| US9674361B2 (en) | System and method for intelligent task management in a workbin | |
| US11750743B1 (en) | Database allocation and analytics for service call centers | |
| AU2013305922B2 (en) | Hierarchical based sequencing machine learning model | |
| US10607167B2 (en) | System and method for intelligent task management and routing based on physiological sensor input data | |
| CN102819607B (en) | A kind of method and apparatus recommending user | |
| US20170344754A1 (en) | System and Method for Data Management and Task Routing Based on Data Tagging | |
| US20170104872A1 (en) | System and Method for Intelligent Task Management and Routing Based on Location Sensor Input Data | |
| CN108899082A (en) | Maintenance service management method, system, terminal and computer readable storage medium | |
| CN108200298A (en) | Method of calling and system based on call center system | |
| CN106203989A (en) | A kind of information processing method and device | |
| CN107508877A (en) | For the information recommendation method and device of vehicle device, storage medium, terminal | |
| CN110389697A (en) | Data interactive method and device, storage medium and electronic device | |
| CN107248002A (en) | A kind of data processing method, system and device | |
| CN104463722A (en) | Method and system for achieving intelligent voice dish ordering | |
| CN108733698A (en) | A kind of processing method and background service system of log information | |
| CN113472958A (en) | Method, system, electronic device and storage medium for receiving branch telephone in centralized mode | |
| CN109951578A (en) | A kind of method, apparatus and storage medium of inquiring contactor | |
| CN109412851A (en) | Link layer path detection method, apparatus and system | |
| CN103888423B (en) | Information processing method and information processing equipment | |
| CN106356056A (en) | Speech recognition method and device | |
| CN107832342A (en) | Robot chat method and system | |
| CN105205905B (en) | The method, apparatus and system that a kind of business hall is called out the numbers |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210705 Address after: Room 702, 7 / F, 67 North Fourth Ring Road West, Haidian District, Beijing Applicant after: BEIJING YUNJI TECHNOLOGY Co.,Ltd. Address before: Room 319, Gaoxin group building, 626 Huangshan Road, Gaoxin District, Hefei City, Anhui Province 230000 Applicant before: ANHUI YUNZHIYI INFORMATION TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room 702, 7th floor, NO.67, Beisihuan West Road, Haidian District, Beijing 100089 Applicant after: Beijing Yunji Technology Co.,Ltd. Address before: Room 702, 7 / F, 67 North Fourth Ring Road West, Haidian District, Beijing Applicant before: BEIJING YUNJI TECHNOLOGY Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190628 |