CN109637523A - A kind of voice-based door lock for vehicle control method and device - Google Patents
A kind of voice-based door lock for vehicle control method and device Download PDFInfo
- Publication number
- CN109637523A CN109637523A CN201811626195.5A CN201811626195A CN109637523A CN 109637523 A CN109637523 A CN 109637523A CN 201811626195 A CN201811626195 A CN 201811626195A CN 109637523 A CN109637523 A CN 109637523A
- Authority
- CN
- China
- Prior art keywords
- voice
- target
- instruction
- vocal print
- model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- 230000001755 vocal effect Effects 0.000 claims abstract description 118
- 239000000284 extract Substances 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims description 22
- 238000004891 communication Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 15
- 238000013135 deep learning Methods 0.000 claims description 13
- 238000003062 neural network model Methods 0.000 claims description 12
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 7
- 238000012795 verification Methods 0.000 description 7
- 238000012512 characterization method Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/28—Constructional details of speech recognition systems
- G10L15/30—Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/04—Training, enrolment or model building
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/22—Interactive procedures; Man-machine interfaces
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/223—Execution procedure of a spoken command
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Lock And Its Accessories (AREA)
Abstract
This application discloses a kind of voice-based door lock for vehicle control method and device, this method comprises: after getting the target voice that target user issues in preset range around target vehicle, extract target vocal print feature therein, then, the target vocal print feature is compared with pre-stored first vocal print feature, when the two is consistent, target voice is input to the instruction constructed in advance and generates model, generate target voice instruction, and instructed according to the target voice, the door lock of target vehicle is controlled.It can be seen that, the application is first compared the vocal print feature of target voice and the voice vocal print feature of target vehicle owning user, when the two is consistent, show that target user is user belonging to target vehicle, it recycles instruction to generate model and generates target voice instruction, to control target vehicle door lock, so that user sends the control that instructs and can be realized to door lock for vehicle without using mobile phone, user experience is improved.
Description
Technical field
This application involves field of artificial intelligence more particularly to a kind of voice-based door lock for vehicle control method and dresses
It sets.
Background technique
Along with the continuous high speed development of social economy, vehicles number is increasing year by year, brings for people's lives
Great convenience, the number for holding driving license is also more and more, currently, China holds driver's license but do not buy the number of vehicle is more than
200000000, annual increase newly holds driving license number also 32,000,000 or so.Due to restricting driving, limiting the policies such as purchase, vehicle resources are driven with holding
According to number can not exactly match, as inlet flow rate persistently increases sharply, existing public transport may be unable to fully meet big
Many trip requirements, therefore, in recent years, the industry of hiring a car such as various timesharing leases is rapidly developed in the market, is rented in timesharing
In the industry of hiring a car such as rent, user needs that car door is unlocked and is latched under the premise of no car key.
Currently, existing solution sends instruction to mobile unit generally by cell phone network to control vehicle doorn
Lock carries out the communication modes such as Bluetooth pairing by mobile phone and vehicle to send instruction, but this side for sending instruction using mobile phone
Formula is unlocked and latches to car door, can not get rid of the dependence to mobile phone, when the both hands of user are occupied or winter it is cold
When, it is not easy to that car door is unlocked and is latched in this way, at this point, the experience of user can be seriously reduced, therefore,
Lacking in the prior art enables to user when both hands are occupied or winter it is cold, refers to without using mobile phone to send
Enable the mode that the control (unlock or latch) to door lock for vehicle can be realized.
Summary of the invention
The main purpose of the embodiment of the present application is to provide a kind of voice-based door lock for vehicle control method and device, energy
It is enough so that user is occupied or when winter it is cold in both hands, can be realized to send instruction to vehicle without using mobile phone
The mode of the control of door lock.
In a first aspect, the embodiment of the present application provides a kind of voice-based door lock for vehicle control method, comprising:
Target voice is obtained, the target voice is the language that target user is issued in preset range around the target vehicle
Sound;
Extract the target vocal print feature of the target voice;
The target vocal print feature is compared with pre-stored first vocal print feature, first vocal print feature is
The vocal print feature of the voice of the target vehicle owning user;
It is when the target vocal print feature is consistent with pre-stored first vocal print feature, the target voice is defeated
Enter to the instruction constructed in advance and generate model, to generate the target voice instruction of target user;
It is instructed according to the target voice, the door lock of the target vehicle is controlled.
Optionally, building described instruction generates model, comprising:
Obtain the sample voice of the target vehicle owning user;
Extract the phonetic feature of the sample voice;
According to the phonetic feature of the sample voice and the corresponding instruction label of the sample voice, train described in generating
Instruction generates model.
Optionally, the phonetic feature and the corresponding instruction label of the sample voice according to the sample voice,
Training generates described instruction and generates model, comprising:
Obtain deep learning neural network model;
Using the phonetic feature and the corresponding instruction label of the sample voice of the sample voice, to the depth
It practises neural network model to be trained, generates described instruction and generate model.
Optionally, the method also includes:
Obtain the verifying voice of the target vehicle owning user;
Extract the phonetic feature of the verifying voice;
The phonetic feature input described instruction of the verifying voice is generated into model, the instruction for obtaining the verifying voice is raw
At result;
When the instruction generation result cue mark result corresponding with the verifying voice of the verifying voice is inconsistent,
The verifying voice is re-used as the sample voice, model is generated to described instruction and is updated.
Optionally, the method also includes:
It establishes and communicates to connect with server-side in advance;
By the communication connection, first vocal print feature that the server-side is sent is received.
Optionally, after the target vehicle has replaced owning user, the method also includes:
Delete pre-stored first vocal print feature.
Second aspect, the application provide a kind of voice-based vehicle door lock controller, comprising:
Target voice acquiring unit, for obtaining target voice, the target voice is target user in target vehicle week
Enclose the voice issued in preset range;
Vocal print feature extraction unit, for extracting the target vocal print feature of the target voice;
Vocal print feature comparing unit, for comparing the target vocal print feature and pre-stored first vocal print feature
Right, first vocal print feature is the vocal print feature of the voice of the target vehicle owning user;
Phonetic order generation unit, for when the target vocal print feature and pre-stored first vocal print feature one
When cause, the target voice is input to the instruction constructed in advance and generates model, to generate the target voice instruction of target user;
Door lock for vehicle control unit controls the door lock of the target vehicle for being instructed according to the target voice
System.
Optionally, described device further include:
Sample voice acquiring unit, for obtaining the sample voice of the target vehicle owning user;
Fisrt feature extraction unit, for extracting the phonetic feature of the sample voice;
Model generation unit, for according to the sample voice phonetic feature and the corresponding instruction of the sample voice
Label, training generate described instruction and generate model.
Optionally, the model generation unit includes:
Model obtains subelement, for obtaining deep learning neural network model;
Model generates subelement, for the phonetic feature and the corresponding finger of the sample voice using the sample voice
Label is enabled, the deep learning neural network model is trained, described instruction is generated and generates model.
Optionally, described device further include:
Voice acquisition unit is verified, for obtaining the verifying voice of the target vehicle owning user;
Second feature extraction unit, for extracting the phonetic feature of the verifying voice;
Instruction results obtaining unit is obtained for the phonetic feature input described instruction of the verifying voice to be generated model
The instruction for obtaining the verifying voice generates result;
Model modification unit, for generating result instruction corresponding with the verifying voice when the instruction of the verifying voice
When marking result inconsistent, the verifying voice is re-used as the sample voice, model is generated to described instruction and is carried out more
Newly.
Optionally, described device further include:
Communication connection establishment unit is communicated to connect for establishing in advance with server-side;
Vocal print feature receiving unit, for receiving first sound that the server-side is sent by the communication connection
Line feature.
Optionally, after the target vehicle has replaced owning user, described device further include:
Vocal print feature deletes unit, for deleting pre-stored first vocal print feature.
A kind of voice-based door lock for vehicle control method and device provided by the embodiments of the present application are used getting target
After the target voice that family is issued in preset range around the target vehicle, the target vocal print feature of target voice can be extracted,
Then, which is compared with pre-stored first vocal print feature, wherein the first vocal print feature refers to
The vocal print feature of the voice of target vehicle owning user, then, when target vocal print feature and pre-stored first vocal print feature
When consistent, target voice can be input to the instruction constructed in advance and generate model, referred to generating the target voice of target user
It enables, in turn, can be instructed according to the target voice, the door lock of target vehicle is controlled.As it can be seen that the embodiment of the present application will first mention
The vocal print feature of the voice of the vocal print feature of the target voice of taking-up and pre-stored target vehicle owning user has carried out pair
Than showing that target user is user belonging to target vehicle, then target voice can be input to preparatory structure when the two being consistent
The instruction built generates model, the target voice instruction of target user is generated, so as to realize using the phonetic order to vehicle
The control of door lock, and then enable to user when both hands are occupied or winter it is cold, it is sent without using mobile phone
The control to door lock for vehicle can be realized in instruction, improves user experience.
Detailed description of the invention
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is the application
Some embodiments for those of ordinary skill in the art without creative efforts, can also basis
These attached drawings obtain other attached drawings.
Fig. 1 is a kind of flow diagram of voice-based door lock for vehicle control method provided by the embodiments of the present application;
Fig. 2 is the flow diagram that building instruction provided by the embodiments of the present application generates model;
Fig. 3 is the flow diagram that a kind of instruction provided by the embodiments of the present application generates model verification method;
Fig. 4 is a kind of composition schematic diagram of voice-based vehicle door lock controller provided by the embodiments of the present application.
Specific embodiment
Currently, hiring a car in industry in timesharing lease etc., user needs to solve car door under the premise of no car key
It locks and latches.Specifically, first it is mobile end by the way that application program of hiring a car (Application, abbreviation APP) will be equipped with
It holds (such as smart phone) and the vehicle rented to carry out Bluetooth pairing mode, establishes communication connection between the two, it is then possible to
Based on the communication connection, instruction is sent to control vehicle to the mobile unit for renting vehicle using the mobile terminal (such as smart phone)
Door lock (unlock or latch), but it is this sent by mobile terminal (such as smart phone) to vehicle instruction in the way of to car door into
Row unlock and latch, the dependence to mobile terminal (such as smart phone) can not be got rid of always, when the both hands of user hold thing or
When person's winter it is cold, it is not easy to that car door is unlocked and is latched in this way, at this point, user can seriously be reduced
Experience.
To solve drawbacks described above, the embodiment of the present application provides a kind of voice-based door lock for vehicle control method, is obtaining
After getting the target voice that target user is issued in preset range around the target vehicle, the target of target voice can be extracted
Then vocal print feature the target vocal print feature is compared with pre-stored first vocal print feature, wherein the first vocal print
Feature refers to the vocal print feature of the voice of target vehicle owning user, then, when target vocal print feature and pre-stored the
When one vocal print feature is consistent, target voice can be input to the instruction constructed in advance and generate model, to generate target user's
Target voice instruction, in turn, can instruct according to the target voice, control the door lock of target vehicle., it is seen then that the application
Embodiment be in such a way that vocal print feature compares, identification target user whether be target vehicle owning user, when identifying
Target user is that the owning user of target vehicle is, can generate model using the instruction constructed in advance, generate target user
Target voice instruction, to realize the control to door lock for vehicle, and then enable to user both hands are occupied or winter
When it is cold, the control that instructs and can be realized to door lock for vehicle is sent without using mobile phone, improves user experience.
To keep the purposes, technical schemes and advantages of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
First embodiment
It is a kind of flow diagram of voice-based door lock for vehicle control method provided in this embodiment referring to Fig. 1, it should
Method the following steps are included:
S101: target voice is obtained, wherein target voice is the voice that target user issues around target vehicle.
In the present embodiment, the vehicle for carrying out door lock control will be needed to be defined as target vehicle, and will be to target vehicle
The user that door lock carries out voice control is defined as target user, meanwhile, by target user around the target vehicle in preset range
The voice of sending is defined as target voice.It should be noted that the present embodiment does not limit the languages of target voice, for example, target
The voice etc. that voice can be the voice of Chinese composition or English is constituted.
Wherein, preset range refers to whether the voice to judge that target user issues is control around target vehicle
The critical range of the voice of target vehicle, for example, the critical range can be set to 3 meters, if target user and target vehicle
Distance in the critical range, then may indicate that the voice that target user issues is door lock in order to control target vehicle, need
The volume to be illustrated is that the specific value of the preset range can be set according to the actual situation, and the application is not limited this.
S102: the target vocal print feature of target voice is extracted.
In the present embodiment, get what target user was issued in preset range around the target vehicle by step S101
After target voice, the vocal print feature extracting method that can use existing or future appearance handles it, for example, can use
Linear predictor coefficient (Linear Prediction Coefficient, the abbreviation LPC) extracting method or mel-frequency of voice are fallen
Spectral coefficient (Mel-Frequency Cepstral Coefficients, abbreviation MFCC) extracting method, is extracted from target voice
The feature that its voiceprint can be characterized out, is defined as target vocal print feature for this feature here, answers in the target vocal print feature
Carry whole voiceprints of corresponding target voice.
S103: target vocal print feature is compared with pre-stored first vocal print feature.
It in the present embodiment, can be by target sound after the target vocal print feature that target voice is extracted by step S102
Line feature is compared with pre-stored first vocal print feature, wherein the first vocal print feature refers to using belonging to target vehicle
The vocal print feature of the voice at family, that is, rent the voice vocal print feature of the user of the target vehicle.
Wherein, in a kind of implementation of the present embodiment, which determined and be sent to by server-side
Mobile unit, concrete implementation process may include following step A1-A2:
Step A1: it establishes communicate to connect with server-side in advance.
In this implementation, the mobile unit of target vehicle and server-side are established communicate to connect in advance, such as can be with
By the radio connections such as WLAN (Wireless Local Area Networks, abbreviation WLAN) establish both sides it
Between communications conduit, to can communicate the interaction of information therebetween.
Step A2: by the communication connection, first vocal print feature that server-side is sent is received.
In this implementation, user's (i.e. owning user of target vehicle) of target vehicle is rented by the APP that hires a car
When renting target vehicle, the APP that hires a car needs to acquire the voice messaging of the user, and the voice messaging of user is sent by network
The voice of the owning user of target vehicle is obtained so that server-side carries out vocal print feature extraction to the voice messaging to server-side
Vocal print feature, be defined as the first vocal print feature.When the mobile unit of target vehicle is built with server-side in advance by step A1
After having stood communication connection, the first vocal print feature can be sent to the vehicle-mounted of target vehicle and be set by server-side by the communication connection
It is standby, to judge target user whether be target vehicle owning user.
S104: when target vocal print feature is consistent with pre-stored first vocal print feature, target voice is input to pre-
The instruction first constructed generates model, to generate the target voice instruction of target user.
In the present embodiment, if being carried out target vocal print feature and pre-stored first vocal print feature by step S103
After comparison, both discoveries are consistent, then target voice can be input to the instruction constructed in advance and generated in model, to generate
The target voice of target user instructs.
It should be noted that needing to construct an instruction generation model in advance to realize this step S104, specifically constructing
Journey can be found in the related introduction of subsequent second embodiment.
S105: instructing according to target voice, controls the door lock of target vehicle.
It in the present embodiment, can be according to the target after the target voice instruction that target user is generated by step S104
Voice quality controls the lock of target vehicle.Such as, it is assumed that pass through the target language of the step S104 target user generated
Sound instruction is " unlock ", that is, target user has issued " unlock " in the preset range around target vehicle, " unlocking ", " opens
Door " etc. the meanings one or more voices, then mobile unit can be instructed according to this " unlock " target voice, to target vehicle into
Row unlock;Or, it is assumed that being instructed by the target voice of the step S104 target user generated is " latching ", that is, target user
One or more voices of the meanings such as " shutdown ", " shutting ", " latching " are had issued in the preset range around target vehicle, then
Mobile unit " can be latched " according to this target voice instruction, latch to target vehicle.
It should be noted that in a kind of implementation of the present embodiment, when target vehicle has replaced owning user, i.e. mesh
Mark vehicle has stopped being rented by the target user, and is rented by other users, at this time, it may be necessary to delete target vehicle on-board equipment
First vocal print feature of upper storage, to store the vocal print feature of the subsequent user for renting target vehicle, then through the above steps
S101-S105 is realized based on the voice of the subsequent user for renting target vehicle, is controlled target vehicle door lock.
To sum up, a kind of voice-based door lock for vehicle control method provided in this embodiment, exists getting target user
After the target voice issued in preset range around target vehicle, the target vocal print feature of target voice can be extracted, then,
The target vocal print feature is compared with pre-stored first vocal print feature, wherein the first vocal print feature refers to target
The vocal print feature of the voice of vehicle owning user, then, when target vocal print feature is consistent with pre-stored first vocal print feature
When, target voice can be input to the instruction constructed in advance and generate model, to generate the target voice instruction of target user, into
And can be instructed according to the target voice, the door lock of target vehicle is controlled.As it can be seen that the embodiment of the present application will first extract
The vocal print feature of target voice and the vocal print feature of the voice of pre-stored target vehicle owning user compared, when
When the two is consistent, shows that target user is user belonging to target vehicle, then target voice can be input to and to be constructed in advance
Instruction generates model, the target voice instruction of target user is generated, so as to realize using the phonetic order to door lock for vehicle
Control, and then enable to user when both hands are occupied or winter it is cold, instruction sent without using mobile phone
The control to door lock for vehicle can be realized, improve user experience.
Second embodiment
The specific building process for generating model to the instruction referred in first embodiment is introduced the present embodiment.It utilizes
The instruction constructed in advance generates model, the target voice instruction of target user can be generated, to the door lock to target vehicle
It is controlled.
Referring to fig. 2, it illustrates the flow diagram that building instruction provided in this embodiment generates model, which includes
Following steps:
S201: the sample voice of target vehicle owning user is obtained.
In the present embodiment, model is generated in order to construct instruction, needs to carry out a large amount of preparation in advance, firstly, needing
The sample voice of target vehicle owning user is collected, for example, the voice of user described in 10 sections of target vehicles can be collected in advance
Data, and each section of voice data of the target vehicle owning user being collected into is first passed through as sample voice, and in advance
The label for manually marking out the instruction type of the target vehicle owning user of these sample voices characterization, generates to training instruction
Model.
S202: the phonetic feature of sample voice is extracted.
It in the present embodiment, can not be straight after the sample voice that target vehicle owning user is got by step S201
It connects and generates instruction generation model for training, but need to extract the phonetic feature of these sample voices, wherein sample voice
The extraction of phonetic feature refers to for sample voice being converted into one group of feature vector with obvious physics, to characterize sample language
The semantic information of sound can use LPC feature extracting method or MFCC extracting method extract in characteristic extraction procedure,
And then can use the phonetic feature of the sample voice extracted, training obtains instruction and generates model.
S203: according to the phonetic feature of sample voice and the corresponding instruction label of sample voice, training generates instruction life
At model.
It in the present embodiment, further, can should after the phonetic feature that sample voice is extracted by step S202
The label result of the corresponding instruction type of phonetic feature and sample voice of sample voice is as model training data, training life
Model is generated at instruction.
Specifically, a kind of to be optionally achieved in that, this step S203 is " according to the phonetic feature and sample of sample voice
The specific implementation process of the corresponding instruction label of this voice, training generation instruction generation model " may include following step B1-
B2:
Step B1: deep learning neural network model is obtained.
In this implementation, available to one deep learning neural network model generates mould as initial instruction
Type, such as convolutional neural networks model, to by subsequent step B2, training generates instruction and generates model.
Step B2: using the corresponding instruction label of phonetic feature and sample voice of sample voice, to deep learning mind
It is trained through network model, generates instruction and generate model.
It, can be successively after getting deep learning neural network model as initial instruction generation model by step B1
One group of sample voice and the corresponding instruction label manually marked of the sample voice are extracted from model training data, are carried out more
Model training is taken turns, until meeting training termination condition, at this point, producing instruction generates model.
Specifically, when carrying out epicycle training, the target voice in first embodiment can be replaced with to epicycle extraction
Sample voice, this can be generated according to the implementation procedure in first embodiment by generating model by current initial order
The instruction type of sample voice characterization.It is then possible to the instruction type is compared with corresponding artificial annotation results, and root
Model parameter is updated according to the difference of the two, until meeting preset condition, then stops the update of model parameter, completes to refer to
The training for generating model is enabled, a trained instruction is generated and generates model.
Through the foregoing embodiment, the sample voice training that can use target vehicle owning user generates instruction and generates mould
Type, then further, the verifying voice that can use target vehicle owning user generate model to the instruction of generation and verify.
Model verification method is generated to instruction provided by the embodiments of the present application with reference to the accompanying drawing to be introduced.
Referring to Fig. 3, it illustrates a kind of flow charts of instruction generation model verification method provided by the embodiments of the present application, such as
Shown in Fig. 3, this method comprises:
S301: the verifying voice of target vehicle owning user is obtained.
In practical applications, instruction generation model is verified in order to realize, it is necessary first to obtain belonging to target vehicle
The verifying voice of user, wherein the verifying voice of target vehicle owning user refers to be used to carry out instruction to generate model
The verifying voice of verifying can continue to execute step 302 after getting the verifying voice of target vehicle owning user.
S302: the phonetic feature of verifying voice is extracted.
It in practical applications,, can not be straight after getting the verifying voice of target vehicle owning user by step S301
It connects and generates model for verifying instruction, but need to extract the phonetic feature of the verifying voice of target vehicle owning user, wherein
The feature extraction of the verifying voice of target vehicle owning user refers to verifying voice and is converted into one group with obvious physics
Feature vector, the semantic information to characterize verifying voice can use LPC feature extracting method in characteristic extraction procedure
Or MFCC extracting method extracts, and then can use the phonetic feature of the verifying voice extracted, verifying obtains instruction life
At model.
S303: the phonetic feature input instruction for verifying voice is generated into model, the instruction for obtaining verifying voice generates result.
It,, further, can after the phonetic feature for extracting verifying voice by step S302 during specific implementation
Model is generated so that the phonetic feature input instruction of voice will be verified, obtains the command verification of verifying voice as a result, can continue in turn
Execute step S304.
S304:, will when the instruction generation result cue mark result corresponding with verifying voice for verifying voice is inconsistent
Verifying voice is re-used as sample voice, generates model to instruction and is updated.
In practical applications, by step S303, after the command verification result for obtaining verifying voice, when the verifying voice
When command verification result artificial annotation results corresponding with verifying voice are inconsistent, which can be re-used as sample
Voice generates model to instruction and is updated.
Through the foregoing embodiment, the verifying voice that can use target vehicle owning user, which generates model to instruction, to be had
Effect card can adjust more new command in time and give birth to when verification result artificial annotation results corresponding with verifying voice are inconsistent
At the parameter of model, and then help to improve instruction generation precision and accuracy that instruction generates model.
It should be noted that the process that above-mentioned building and verifying instruction generate model can be executed in server-side, it can also be with
Target vehicle mobile unit end execute, when server-side executes, after the completion of model training, server-side can by with
The model is sent to the mobile unit of target vehicle, to realize base by the communication connection between the mobile unit of target vehicle
The door lock control of target vehicle is carried out in the voice of target user, the specific position that constructs can be set according to the actual situation,
The embodiment of the present application is not limited this.
To sum up, model is generated using instruction made of the present embodiment training, can use the vocal print letter of characterization target voice
The vocal print feature of breath generates corresponding target voice instruction, to realize the control to door lock for vehicle, and then target is enabled to use
When both hands are occupied or winter it is cold, instruction is sent without using mobile phone can be realized control to door lock for vehicle at family
System, improves user experience.
3rd embodiment
A kind of voice-based vehicle door lock controller will be introduced in the present embodiment, and related content refers to above-mentioned
Embodiment of the method.
It referring to fig. 4, is a kind of composition schematic diagram of voice-based vehicle door lock controller provided in this embodiment, it should
Device includes:
Target voice acquiring unit 401, for obtaining target voice, the target voice is target user in target vehicle
The voice issued in surrounding preset range;
Vocal print feature extraction unit 402, for extracting the target vocal print feature of the target voice;
Vocal print feature comparing unit 403, for by the target vocal print feature and pre-stored first vocal print feature into
Row compares, and first vocal print feature is the vocal print feature of the voice of the target vehicle owning user;
Phonetic order generation unit 404, for when the target vocal print feature and the pre-stored first vocal print spy
When levying consistent, the target voice is input to the instruction constructed in advance and generates model, to generate the target voice of target user
Instruction;
Door lock for vehicle control unit 405 carries out the door lock of the target vehicle for being instructed according to the target voice
Control.
In a kind of implementation of the present embodiment, described device further include:
Sample voice acquiring unit, for obtaining the sample voice of the target vehicle owning user;
Fisrt feature extraction unit, for extracting the phonetic feature of the sample voice;
Model generation unit, for according to the sample voice phonetic feature and the corresponding instruction of the sample voice
Label, training generate described instruction and generate model.
In a kind of implementation of the present embodiment, the model generation unit includes:
Model obtains subelement, for obtaining deep learning neural network model;
Model generates subelement, for the phonetic feature and the corresponding finger of the sample voice using the sample voice
Label is enabled, the deep learning neural network model is trained, described instruction is generated and generates model.
In a kind of implementation of the present embodiment, described device further include:
Voice acquisition unit is verified, for obtaining the verifying voice of the target vehicle owning user;
Second feature extraction unit, for extracting the phonetic feature of the verifying voice;
Instruction results obtaining unit is obtained for the phonetic feature input described instruction of the verifying voice to be generated model
The instruction for obtaining the verifying voice generates result;
Model modification unit, for generating result instruction corresponding with the verifying voice when the instruction of the verifying voice
When marking result inconsistent, the verifying voice is re-used as the sample voice, model is generated to described instruction and is carried out more
Newly.
In a kind of implementation of the present embodiment, described device further include:
Communication connection establishment unit is communicated to connect for establishing in advance with server-side;
Vocal print feature receiving unit, for receiving first sound that the server-side is sent by the communication connection
Line feature.
In a kind of implementation of the present embodiment, after the target vehicle has replaced owning user, described device is also
Include:
Vocal print feature deletes unit, for deleting pre-stored first vocal print feature.
To sum up, a kind of voice-based vehicle door lock controller provided in this embodiment, exists getting target user
After the target voice issued in preset range around target vehicle, the target vocal print feature of target voice can be extracted, then,
The target vocal print feature is compared with pre-stored first vocal print feature, wherein the first vocal print feature refers to target
The vocal print feature of the voice of vehicle owning user, then, when target vocal print feature is consistent with pre-stored first vocal print feature
When, target voice can be input to the instruction constructed in advance and generate model, to generate the target voice instruction of target user, into
And can be instructed according to the target voice, the door lock of target vehicle is controlled.As it can be seen that the embodiment of the present application will first extract
The vocal print feature of target voice and the vocal print feature of the voice of pre-stored target vehicle owning user compared, when
When the two is consistent, shows that target user is user belonging to target vehicle, then target voice can be input to and to be constructed in advance
Instruction generates model, the target voice instruction of target user is generated, so as to realize using the phonetic order to door lock for vehicle
Control, and then enable to user when both hands are occupied or winter it is cold, instruction sent without using mobile phone
The control to door lock for vehicle can be realized, improve user experience.
As seen through the above description of the embodiments, those skilled in the art can be understood that above-mentioned implementation
All or part of the steps in example method can be realized by means of software and necessary general hardware platform.Based on such
Understand, substantially the part that contributes to existing technology can be in the form of software products in other words for the technical solution of the application
It embodies, which can store in storage medium, such as ROM/RAM, magnetic disk, CD, including several
Instruction is used so that a computer equipment (can be the network communications such as personal computer, server, or Media Gateway
Equipment, etc.) execute method described in certain parts of each embodiment of the application or embodiment.
It should be noted that each embodiment in this specification is described in a progressive manner, each embodiment emphasis is said
Bright is the difference from other embodiments, and the same or similar parts in each embodiment may refer to each other.For reality
For applying device disclosed in example, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place
Referring to method part illustration.
It should also be noted that, herein, relational terms such as first and second and the like are used merely to one
Entity or operation are distinguished with another entity or operation, without necessarily requiring or implying between these entities or operation
There are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant are intended to contain
Lid non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
The foregoing description of the disclosed embodiments makes professional and technical personnel in the field can be realized or use the application.
Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the application.Therefore, the application
It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one
The widest scope of cause.
Claims (12)
1. a kind of voice-based door lock for vehicle control method characterized by comprising
Target voice is obtained, the target voice is the voice that target user is issued in preset range around the target vehicle;
Extract the target vocal print feature of the target voice;
The target vocal print feature is compared with pre-stored first vocal print feature, first vocal print feature is described
The vocal print feature of the voice of target vehicle owning user;
When the target vocal print feature is consistent with pre-stored first vocal print feature, the target voice is input to
The instruction constructed in advance generates model, to generate the target voice instruction of target user;
It is instructed according to the target voice, the door lock of the target vehicle is controlled.
2. the method according to claim 1, wherein building described instruction generates model, comprising:
Obtain the sample voice of the target vehicle owning user;
Extract the phonetic feature of the sample voice;
According to the phonetic feature of the sample voice and the corresponding instruction label of the sample voice, training generates described instruction
Generate model.
3. according to the method described in claim 2, it is characterized in that, the phonetic feature and institute according to the sample voice
The corresponding instruction label of sample voice is stated, training generates described instruction and generates model, comprising:
Obtain deep learning neural network model;
Using the phonetic feature and the corresponding instruction label of the sample voice of the sample voice, to the deep learning mind
It is trained through network model, generates described instruction and generate model.
4. method according to any one of claims 2 to 3, which is characterized in that the method also includes:
Obtain the verifying voice of the target vehicle owning user;
Extract the phonetic feature of the verifying voice;
The phonetic feature input described instruction of the verifying voice is generated into model, the instruction for obtaining the verifying voice generates knot
Fruit;
When the instruction generation result cue mark result corresponding with the verifying voice of the verifying voice is inconsistent, by institute
It states verifying voice and is re-used as the sample voice, model is generated to described instruction and is updated.
5. method according to any one of claims 1 to 3, which is characterized in that the method also includes:
It establishes and communicates to connect with server-side in advance;
By the communication connection, first vocal print feature that the server-side is sent is received.
6. described the method according to claim 1, wherein after the target vehicle has replaced owning user
Method further include:
Delete pre-stored first vocal print feature.
7. a kind of voice-based vehicle door lock controller characterized by comprising
Target voice acquiring unit, for obtaining target voice, the target voice is that target user is pre- around target vehicle
If the voice issued in range;
Vocal print feature extraction unit, for extracting the target vocal print feature of the target voice;
Vocal print feature comparing unit, for the target vocal print feature to be compared with pre-stored first vocal print feature,
First vocal print feature is the vocal print feature of the voice of the target vehicle owning user;
Phonetic order generation unit, it is consistent with pre-stored first vocal print feature for working as the target vocal print feature
When, the target voice is input to the instruction constructed in advance and generates model, to generate the target voice instruction of target user;
Door lock for vehicle control unit controls the door lock of the target vehicle for being instructed according to the target voice.
8. device according to claim 7, which is characterized in that described device further include:
Sample voice acquiring unit, for obtaining the sample voice of the target vehicle owning user;
Fisrt feature extraction unit, for extracting the phonetic feature of the sample voice;
Model generation unit, for being marked according to the phonetic feature of the sample voice and the corresponding instruction of the sample voice
Label, training generate described instruction and generate model.
9. device according to claim 8, which is characterized in that the model generation unit includes:
Model obtains subelement, for obtaining deep learning neural network model;
Model generates subelement, for the phonetic feature and the corresponding instruction mark of the sample voice using the sample voice
Label, are trained the deep learning neural network model, generate described instruction and generate model.
10. according to the described in any item devices of claim 7 to 8, which is characterized in that described device further include:
Voice acquisition unit is verified, for obtaining the verifying voice of the target vehicle owning user;
Second feature extraction unit, for extracting the phonetic feature of the verifying voice;
Instruction results obtaining unit obtains institute for the phonetic feature input described instruction of the verifying voice to be generated model
The instruction for stating verifying voice generates result;
Model modification unit, for generating result cue mark corresponding with the verifying voice when the instruction of the verifying voice
As a result when inconsistent, the verifying voice is re-used as the sample voice, model is generated to described instruction and is updated.
11. device according to any one of claims 7 to 9, which is characterized in that described device further include:
Communication connection establishment unit is communicated to connect for establishing in advance with server-side;
Vocal print feature receiving unit, for it is special to receive first vocal print that the server-side is sent by the communication connection
Sign.
12. device according to claim 7, which is characterized in that described after the target vehicle has replaced owning user
Device further include:
Vocal print feature deletes unit, for deleting pre-stored first vocal print feature.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811626195.5A CN109637523A (en) | 2018-12-28 | 2018-12-28 | A kind of voice-based door lock for vehicle control method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811626195.5A CN109637523A (en) | 2018-12-28 | 2018-12-28 | A kind of voice-based door lock for vehicle control method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109637523A true CN109637523A (en) | 2019-04-16 |
Family
ID=66078890
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811626195.5A Pending CN109637523A (en) | 2018-12-28 | 2018-12-28 | A kind of voice-based door lock for vehicle control method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109637523A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111783892A (en) * | 2020-07-06 | 2020-10-16 | 广东工业大学 | Robot instruction identification method and device, electronic equipment and storage medium |
CN112053678A (en) * | 2019-06-06 | 2020-12-08 | 北京快松果科技有限公司 | Lock opening and closing method and system based on voice recognition, lock opening and closing body and shared vehicle |
CN112298104A (en) * | 2019-07-31 | 2021-02-02 | 比亚迪股份有限公司 | Vehicle control method and device, storage medium, electronic equipment and vehicle |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105976812A (en) * | 2016-04-28 | 2016-09-28 | 腾讯科技(深圳)有限公司 | Voice identification method and equipment thereof |
CN106920303A (en) * | 2017-01-21 | 2017-07-04 | 云丁网络技术(北京)有限公司 | A kind of method for unlocking and its intelligent door lock system based on speech recognition |
CN108281137A (en) * | 2017-01-03 | 2018-07-13 | 中国科学院声学研究所 | A kind of universal phonetic under whole tone element frame wakes up recognition methods and system |
CN108922521A (en) * | 2018-08-15 | 2018-11-30 | 合肥讯飞数码科技有限公司 | A kind of voice keyword retrieval method, apparatus, equipment and storage medium |
-
2018
- 2018-12-28 CN CN201811626195.5A patent/CN109637523A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105976812A (en) * | 2016-04-28 | 2016-09-28 | 腾讯科技(深圳)有限公司 | Voice identification method and equipment thereof |
CN108281137A (en) * | 2017-01-03 | 2018-07-13 | 中国科学院声学研究所 | A kind of universal phonetic under whole tone element frame wakes up recognition methods and system |
CN106920303A (en) * | 2017-01-21 | 2017-07-04 | 云丁网络技术(北京)有限公司 | A kind of method for unlocking and its intelligent door lock system based on speech recognition |
CN108922521A (en) * | 2018-08-15 | 2018-11-30 | 合肥讯飞数码科技有限公司 | A kind of voice keyword retrieval method, apparatus, equipment and storage medium |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112053678A (en) * | 2019-06-06 | 2020-12-08 | 北京快松果科技有限公司 | Lock opening and closing method and system based on voice recognition, lock opening and closing body and shared vehicle |
CN112053678B (en) * | 2019-06-06 | 2023-10-03 | 北京快松果科技有限公司 | Switch lock method and system based on voice recognition, switch lock body and sharing vehicle |
CN112298104A (en) * | 2019-07-31 | 2021-02-02 | 比亚迪股份有限公司 | Vehicle control method and device, storage medium, electronic equipment and vehicle |
CN111783892A (en) * | 2020-07-06 | 2020-10-16 | 广东工业大学 | Robot instruction identification method and device, electronic equipment and storage medium |
CN111783892B (en) * | 2020-07-06 | 2021-10-01 | 广东工业大学 | Robot instruction identification method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109637523A (en) | A kind of voice-based door lock for vehicle control method and device | |
CN107886957A (en) | Voice wake-up method and device combined with voiceprint recognition | |
CN104036774B (en) | Tibetan dialect recognition methods and system | |
CN101467204B (en) | Method and system for bio-metric voice print authentication | |
US9542938B2 (en) | Scene recognition method, device and mobile terminal based on ambient sound | |
CN110459210A (en) | Answering method, device, equipment and storage medium based on speech analysis | |
CN103730114A (en) | Mobile equipment voiceprint recognition method based on joint factor analysis model | |
CN106920303A (en) | A kind of method for unlocking and its intelligent door lock system based on speech recognition | |
CN109493847B (en) | Sound recognition system and voice recognition device | |
CN105931638A (en) | Intelligent-robot-oriented dialog system data processing method and device | |
CN108766441A (en) | A kind of sound control method and device based on offline Application on Voiceprint Recognition and speech recognition | |
CN108389578A (en) | Smart classroom speech control system | |
CN105096121A (en) | Voiceprint authentication method and device | |
CN107393541A (en) | Information Authentication method and apparatus | |
CN102413100A (en) | Voice-print authentication system having voice-print password picture prompting function and realization method thereof | |
CN106709804A (en) | Interactive wealth planning consulting robot system | |
CN109670166A (en) | Collection householder method, device, equipment and storage medium based on speech recognition | |
CN108074575A (en) | Identity verification method and device based on recurrent neural network | |
CN110490592A (en) | Interior consumption and payment method and cloud server based on recognition of face | |
CN105895082A (en) | Acoustic model training method and device as well as speech recognition method and device | |
CN105118509A (en) | Security authentication method based on voiceprint two-dimensional code | |
CN106898355A (en) | A kind of method for distinguishing speek person based on two modelings | |
CN109243466A (en) | A kind of vocal print authentication training method and system | |
CN109273011A (en) | A kind of the operator's identification system and method for automatically updated model | |
CN1403953A (en) | Palm acoustic-print verifying system |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190416 |
|
RJ01 | Rejection of invention patent application after publication |