CN117238281B - Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium - Google Patents
Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium Download PDFInfo
- Publication number
- CN117238281B CN117238281B CN202311483511.9A CN202311483511A CN117238281B CN 117238281 B CN117238281 B CN 117238281B CN 202311483511 A CN202311483511 A CN 202311483511A CN 117238281 B CN117238281 B CN 117238281B
- Authority
- CN
- China
- Prior art keywords
- domain
- vehicle
- current
- dialogue
- field
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000001965 increasing effect Effects 0.000 claims abstract description 17
- 230000006872 improvement Effects 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 7
- 230000003993 interaction Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 description 5
- 238000004378 air conditioning Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000002452 interceptive effect Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000001308 synthesis method Methods 0.000 description 2
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Landscapes
- Navigation (AREA)
Abstract
The invention provides a vehicle-mounted system voice guide word arbitration method, a device, a vehicle-mounted system voice guide word arbitration device and a storage medium. The method comprises the following steps: acquiring all generalized guide words in a second-level module catalog under a first-level domain catalog as candidate guide words; determining initial trigger probability of each candidate guide word according to the total number; acquiring current voice input of a user, a history dialogue record, an application program operated by a vehicle machine system, a state of the application program, a current vehicle state and current external environment information of a vehicle; according to the information, the triggering probability of each candidate guide word in the dialogue field, the application program field, the vehicle state field, the external environment field and the dialogue history field is respectively increased by a first multiple to a fifth multiple, and finally the triggering probabilities corresponding to the fields are summed to obtain the current triggering probability of the candidate guide word; and randomly selecting the guide word from the candidate guide words according to the trigger probability. The invention can intelligently select the most proper guide words and provide personalized guide.
Description
Technical Field
The present invention relates generally to the field of vehicle-to-vehicle voice systems, and more particularly, to a vehicle-to-vehicle voice guidance word arbitration method, device, vehicle-to-vehicle, and storage medium.
Background
With the popularization of car machine systems and the development of car networking technologies, voice control becomes a convenient and safe function in modern automobiles.
The design of the voice guide words in the traditional car machine system is more random, for example, a scheme of random scrolling of the guide words is adopted, so that the guide content of the guide words is untimely, the appearance time of the guide words is inaccurate, a proper guide effect cannot be achieved, and the user experience is poor.
Disclosure of Invention
In order to solve the above-mentioned problems in the prior art, in a first aspect, an embodiment of the present invention provides a method for arbitrating a speech guide word of a machine system, the method comprising: obtaining all generalized guide words in a secondary module catalog under a primary domain catalog as candidate guide words, wherein the primary domain catalog comprises a plurality of domains selected from the navigation domain, the media domain, the vehicle control domain, the message domain, the common problem solving domain, the setting domain, the weather domain and the vehicle information domain; determining initial trigger probability of each candidate guide word according to the total number of the candidate guide words; acquiring current voice input of a user, a history dialogue record, an application program operated by a vehicle machine system, a state of the application program, a current vehicle state and current external environment information of a vehicle; determining a current dialogue context according to the current voice input and the historical dialogue record; determining the dialogue domain corresponding to the current dialogue context in the primary domain catalog; the triggering probability of candidate guide words in all the two-level module catalogues in the dialogue field is improved by a first multiple; determining the application program domain corresponding to the application program in the primary domain catalog according to the application program and the state thereof; the triggering probability of candidate guide words in all the two-level module catalogues in the application program field is increased by a second multiple; determining a vehicle state field corresponding to the current vehicle state in the primary field catalog according to the current vehicle state; the triggering probability of candidate guide words in all the two-level module catalogues in the vehicle state field is improved by a third multiple; according to the external environment information, determining an external environment field corresponding to the external environment information in the primary field catalog; the triggering probability of candidate guide words in all the two-level module catalogues in the external environment field is increased by a fourth multiple; according to the current voice input and the history dialogue record, determining a dialogue history field corresponding to the current round dialogue history in the primary field catalog; the triggering probability of candidate guide words in all the two-level module catalogues in the dialogue history field is improved by a fifth multiple; and randomly selecting the guide words from the candidate guide words according to the triggering probability of all the candidate guide words, and using the guide words as the guide words to be broadcasted by the vehicle-mounted voice system.
In some embodiments, the method further comprises: judging whether the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue according to the current voice input and the history dialogue record; in the case that the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue, the triggering probability of the candidate guide words in all the two-level module catalogues in the same dialogue field as the dialogue field of the current round of dialogue is increased by a sixth multiple.
In some embodiments, the method further comprises: acquiring a current navigation destination under the condition that an application program operated by the vehicle system comprises a navigation application program and the state of the navigation application program is in navigation; determining whether the current navigation destination is a strange place according to the current navigation destination and a previously stored historical navigation destination; and in the case that the current navigation destination is a strange place, increasing the broadcasting frequency of the guide word by a preset number.
In some embodiments, the first multiple is equal to or greater than the second multiple, the second multiple is equal to or greater than the third multiple, the third multiple is equal to or greater than the fourth multiple, and the fourth multiple is equal to or greater than the fifth multiple.
In some embodiments, the first multiple is 5 times, the second multiple is 3 times, the third multiple is 2 times, the fourth multiple is 2 times, and the fifth multiple is 1.5 times.
In some embodiments, determining, in the primary domain directory, an external environment domain corresponding to the external environment information according to the external environment information includes: and determining the domain of the event meeting the triggering condition under the current external environment as the external environment domain according to the external environment information.
In some embodiments, the state of the application includes one or more of the following: whether the application program is started; whether the application program is in the foreground or the background of the vehicle-mounted system; the function currently executed by the application program; a state corresponding to the interaction information of the user and the application program; and the state corresponding to the output information of the application program.
In a second aspect, an embodiment of the present invention provides a voice guidance word arbitration device for a vehicle-mounted system, where the device includes: the system comprises a candidate guide word acquisition module, a first-level domain directory and a second-level domain directory, wherein the candidate guide word acquisition module is configured to acquire all generalized guide words in the second-level module directory under the first-level domain directory as candidate guide words, and the first-level domain directory comprises a plurality of domains selected from the navigation domain, the media domain, the vehicle control domain, the message domain, the common problem solving domain, the setting domain, the weather domain and the vehicle information domain; the initial trigger probability determining module is configured to determine the initial trigger probability of each candidate guide word according to the total number of the candidate guide words; the related information acquisition module is configured to acquire current voice input of a user, a history dialogue record, an application program operated by a vehicle-mounted system and state thereof, current vehicle state and current external environment information of a vehicle; a dialog context determination module configured to determine a current dialog context based on the current speech input and the historical dialog record; a dialogue domain determining module configured to determine a dialogue domain corresponding to the current dialogue context in the primary domain directory; the first probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the dialogue field by a first multiple; an application program domain determining module, configured to determine an application program domain corresponding to the application program in the primary domain directory according to the application program and a state thereof; the second probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the application program field by a second multiple; a vehicle state domain determining module configured to determine a vehicle state domain corresponding to the current vehicle state in the primary domain directory according to the current vehicle state; the third probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the vehicle state field by a third multiple; an external environment domain determining module configured to determine an external environment domain corresponding to the external environment information in the primary domain directory according to the external environment information; the fourth probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the external environment field by a fourth multiple; a conversation history domain determining module configured to determine a conversation history domain corresponding to a current round of conversation history in the primary domain directory according to the current voice input and the history conversation record; a fifth probability improvement module configured to improve the trigger probability of the candidate guide words in all the two-level module catalogues in the dialogue history field by a fifth multiple; the guide word selecting module is configured to randomly select guide words from the candidate guide words according to the trigger probabilities of all the candidate guide words, and the guide words are used as the guide words to be broadcasted by the vehicle-to-vehicle voice system.
In a third aspect, an embodiment of the present invention provides a vehicle machine, where the vehicle machine includes a memory and a processor, and the memory stores a computer program, and when the computer program is executed by the processor, implements the vehicle machine system voice guidance word arbitration method described in any of the foregoing embodiments.
In a fourth aspect, embodiments of the present invention provide a storage medium storing computer readable instructions that, when executed by a processor, perform a vehicle-to-machine system voice guidance word arbitration method described in accordance with any of the embodiments above.
The embodiment of the invention provides arbitration logic of intelligent voice guidance words, which can be used for conjecture the voice guidance words needed by a user by combining information such as a vehicle-mounted machine, user intention, vehicle state, environment state and the like in the current scene, so that the user can conveniently execute related operations. Compared with the existing arbitration method, the method uses a new data analysis technology, a context synthesis method, vehicle state sensing, system application state sensing and the like, thereby realizing more intelligent and more accurate guide word selection. The customized guidance is better provided according to the user intention and the vehicle state, so that the user experience of using voice and the voice operation efficiency are improved.
The guide word arbitration method provided by the embodiment of the invention can be applied and implemented in an actual vehicle-to-machine system, and brings practical value to the field of intelligent transportation.
The voice guidance word arbitration scheme of the vehicle-mounted system provided by the embodiment of the invention comprehensively considers a plurality of factors including voice input of a user, context information and current vehicle state. By using a particular arbitration algorithm, the system can intelligently select the most appropriate guide word to meet the current intent and needs of the user.
The invention calculates the probability of possible occurrence of each guide word by using a calculation rule for a plurality of candidate guide words, and carries out random selection according to the calculated probability, wherein the probability is higher, and the probability of being randomly selected is higher, so that the most suitable guide word in the current state can be ensured to be selected, and the guide word can be prevented from being changed when the vehicle is always kept in a certain state.
Since the arbitration scheme considers the context information of the users and the navigation destination, it is possible to provide personalized guidance according to different situations of each user. This will make the user feel more understandable and careable, enhancing the interactive experience between the vehicle system and the user.
The arbitration scheme also carries out the selection of the guide words according to the current vehicle state, and considers the information such as the vehicle speed, the steering signal and the like. Therefore, the rationality and the practicability of the guide words in the current state of the vehicle can be ensured, the interference in the driving process is reduced, the driving safety is improved, and the vehicle state adaptation is optimized.
By providing accurate guide words, the scheme encourages users to use voice operations more frequently, reducing user operation errors and unnecessary repetition. The method and the device can increase the dependence and satisfaction degree of the user on the vehicle-mounted system, improve the accuracy and efficiency of voice operation and improve the operation frequency and accuracy of the user.
The arbitration scheme is applicable to various vehicles including private cars, commercial vehicles, public transportation vehicles and the like. Whether in daily driving or commercial use, the scheme can provide more intelligent, convenient and safe vehicle-mounted voice control functions for drivers and passengers. The method can also be used in other fields requiring voice interaction and guidance word display, and has wide application coverage.
Drawings
The above, as well as additional purposes, features, and advantages of embodiments of the present invention will become apparent in the following detailed written description and claims upon reference to the accompanying drawings. Several embodiments of the present invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which:
FIG. 1 illustrates a flow chart of a method of vehicle-to-machine system voice guidance word arbitration according to an embodiment of the invention;
fig. 2 shows a block diagram of a voice guidance word arbitration device of a car machine system according to an embodiment of the present invention.
In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts.
Detailed Description
The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are presented merely to enable those skilled in the art to better understand and practice the invention and are not intended to limit the scope of the invention in any way.
In order to improve the frequency and prediction accuracy of voice operation and enable each guide word to help solve the current problem of a user, the embodiment of the invention provides a brand-new method for generating and arbitrating voice guide words of a vehicle-computer system.
In one aspect, the embodiment of the invention provides a method for arbitrating voice guide words of a vehicle-mounted system. Referring to fig. 1, a flow chart of a vehicle-to-machine system voice guidance word arbitration method 100 is shown, according to an embodiment of the invention. The method 100 includes steps S101-S115.
In step S101, all generalized guide words in a secondary module directory under a primary domain (domain) directory are obtained as candidate guide words, wherein the primary domain directory includes a plurality of domains selected from a navigation domain, a media domain, a vehicle control domain, a message domain, a common problem answer (FAQ) domain, a setting domain, a weather domain, and a vehicle information domain. The embodiment of the invention uses the domain as a primary catalog to divide the guide words preliminarily.
On the basis of the above primary catalog, each module is subdivided for different functional points: taking navigation as an example, the guide words corresponding to navigation may include the following functional points: map control (zoom-in and zoom-out of a map, north orientation of a map, etc.), map setting (switch icon, switch map theme, switch detail Jian Bobao, etc.), destination search (navigation to a place, peripheral search, along-road search, etc.), information query in navigation (remaining examples, remaining time, road conditions ahead, etc.), etc.
For the subdivision of each module in the second stage, there are a plurality of generalized instruction words under each subdivision module so as to improve the diversity of the instruction words.
In step S102, an initial trigger probability for each candidate guide word is determined based on the total number of candidate guide words. For example, the initial trigger probability n=1/(the number of all secondary guide words) for each candidate guide word.
In step S103, the current voice input of the user, the history dialogue record, the application program and the state thereof run by the vehicle system, the current vehicle state, and the current external environment information of the vehicle are acquired.
In step S104, the current dialog context, i.e. the intent and requirements of the voice interaction, is determined from the current voice input and the history dialog records. For example, the user may ask for a navigation route, media play, or vehicle status, etc. For example, the user's voice input is first converted to text form and the dialog context is determined through semantic understanding and dialog management.
In step S105, the dialog domain corresponding to the current dialog context is determined in the primary domain directory.
In step S106, the trigger probability of the candidate guide words in all the two-level module directories in the dialogue domain is increased by a first factor.
In step S107, an application domain corresponding to the application is determined in the primary domain directory according to the application and its state. Applications such as navigation, music, weather, messaging center, etc.
As one embodiment of the invention, the state of the application includes one or more of the following: whether the application program is started; whether the application program is in the foreground or the background of the vehicle-mounted system; the function currently executed by the application program; a state corresponding to the interaction information of the user and the application program; the state corresponding to the output information of the application program.
In step S108, the trigger probability of the candidate guide words in all the two-level module directories in the application domain is increased by a second factor.
In step S109, a vehicle state field corresponding to the current vehicle state is determined in the primary field directory according to the current vehicle state. Current vehicle status information such as vehicle speed, gear, seat heating switch, window switch status, vehicle fault information, air conditioning switch status, lights, etc.
In step S110, the trigger probability of the candidate guide words in all the two-level module catalogues in the vehicle state field is increased by a third multiple.
In step S111, according to the external environment information, an external environment area corresponding to the external environment information is determined in the primary area directory. Current environmental conditions such as light sensor information, wind window sensors, air quality information, outdoor temperature, noise figure, etc.
As one embodiment of the present invention, determining, in the primary domain directory, an external environment domain corresponding to external environment information according to the external environment information includes: according to the external environment information, determining the domain of the event meeting the triggering condition under the current external environment as the external environment domain.
In step S112, the triggering probability of the candidate guide words in all the two-level module directories in the external environment field is increased by a fourth factor.
In step S113, a dialogue history field corresponding to the current turn dialogue history is determined in the primary field catalog based on the current voice input and the history dialogue record.
In step S114, the trigger probability of the candidate guide words in all the two-level module catalogues in the dialogue history field is increased by a fifth multiple.
For example, the calculation formula of the second-level guide word probability can be expressed as follows:
P(tip)= N tip * (1+m1 (whether the domain is the same as the current dialog domain) +m2 (whether the application is in the foreground \whether the application is currently executing the corresponding operation) +m3 (whether a certain vehicle state is met) +m4 (whether a triggered environmental condition is met) +m5 (whether the domain appears in the current round))
Wherein, tip represents a candidate guide word, P (tip) represents the trigger probability of the candidate guide word, N tip Representing the initial trigger probability of the candidate guide word, M1-M5 represent the first to fifth times respectively, if the content in the brackets is judged to be yes, the value in the brackets is 1, otherwise, the value in the brackets is 0.
As an embodiment of the present invention, the first multiple is equal to or greater than the second multiple, the second multiple is equal to or greater than the third multiple, the third multiple is equal to or greater than the fourth multiple, and the fourth multiple is equal to or greater than the fifth multiple. In this embodiment, the priority of probability boosting is firstly the domain of the current round of dialogue with the vehicle system, secondly the domain in which the application is located, and then the vehicle state, external environment and dialogue history in that order. In other embodiments, the priority may be flexibly configured according to the actual application scenario or the user requirement.
As just one specific example, the first multiple is 5 times, the second multiple is 3 times, the third multiple is 2 times, the fourth multiple is 2 times, and the fifth multiple is 1.5 times. In other words, in this example, the probability of the guidance word trigger for the corresponding domain is raised to five times based on the domain of the current dialog; judging according to the state of each application, and improving the guide words in the corresponding field by three times; the triggering probability of the guide word in the corresponding field is doubled according to the vehicle state information; and improving the triggering probability of the guide word in the corresponding field to be twice according to the state of the current external environment. The probability of triggering the guide words in the corresponding domain is improved by 1.5 times according to the domain related to the current round dialog history.
It should be noted that the specific value of the multiple may be flexibly configured according to the application scenario or the user requirement.
It should be further noted that the above-described execution order of the dialog field, application field, vehicle state field, external environment field, determination of dialog history field, and multiplication of probabilities of respective guidance words is merely an example, and the determination of respective fields and the probability improvement may be performed in parallel or in other order, and the present invention is not limited in this respect.
In step S115, according to the trigger probabilities of all the candidate guide words, the guide words are randomly selected from the candidate guide words, and the selected guide words are used as the guide words to be broadcasted by the vehicle-mounted voice system. The corresponding fields mentioned in the embodiment of the invention are subdivided into two-level modules, and for different guide words contained in the two-level modules, one of the guide words is selected by using a random method for displaying.
The steps S101 to S115 described above calculate the probability of possible occurrence of each guide word using a calculation rule for a plurality of candidate guide words, and randomly select according to the calculated probability, the higher the probability of being randomly selected, so that it can be ensured that the guide word most suitable for the current state is selected, and it can be prevented that the guide word does not change when the vehicle is kept in a certain state all the time.
By calculating the data such as the vehicle state, the application data, the environment information and the context content, the system can sort the priorities of the guide words according to the actual situation, and the current guide requirements of the user can be better met. Such adjustment can make the guide word more intelligent and accurate under different circumstances, promotes user experience and voice operation's efficiency. It should be noted that the specific state weight allocation can be adjusted according to the actual situation and the user feedback, so as to gradually optimize and meet different requirements of the user.
The voice guidance word arbitration method of the vehicle-mounted system provided by the embodiment of the invention combines comprehensive analysis of context information such as user dialogue history, navigation destination, front and rear voice instructions and the like in voice guidance, and avoids inaccurate or invalid guidance caused by misunderstanding of user intention by the vehicle-mounted system. The method also combines the vehicle state, and solves the problem that the traditional vehicle-mounted system does not fully sense and analyze the current vehicle state, so that the guiding time and the guiding content are unsuitable.
Vehicle systems often integrate a variety of applications such as navigation, entertainment, communications, etc. The conventional method does not comprehensively consider the current state of each application, which may lead to booting when not appropriate, interfering with the operation currently being performed by the user. Embodiments of the present invention address this problem in the prior art by considering applications and their state in the context of a bootstrap arbitration.
In addition, the internal and external environments of the vehicle, such as noise in the vehicle, external traffic conditions, etc., are also taken into consideration to ensure clear communication of the guidance language.
As an embodiment of the present invention, the method may further include: judging whether the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue according to the current voice input and the history dialogue record; in the case that the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue, the triggering probability of the candidate guide words in all the two-level module catalogues in the same dialogue field is increased by a sixth multiple.
As an embodiment of the present invention, the method may further include: acquiring a current navigation destination under the condition that an application program operated by the vehicle system comprises a navigation application program and the state of the navigation application program is in navigation; determining whether the current navigation destination is a strange place according to the current navigation destination and the previously stored historical navigation destination; and in the case that the current navigation destination is a strange place, increasing the broadcasting frequency of the guide words by a preset number. More frequent navigation indications may be required if the user is traveling to a strange location.
Specific applications of embodiments of the present invention are described below by taking several specific scenarios as examples.
First, let the average probability of the guide word of all the initial states be n.
Scene one (reference context): during the current dialogue of the user, the presented guide words should be guided most preferably according to the current user's phonetic NLU (natural language understanding) result and dialogue history. If the dialog intention of the current round of the user is "turn on the air conditioner", the triggering probability of the guide word (for example, the air conditioner temperature is set to 24 degrees) in the corresponding air conditioner operation field is increased by 5 times (5 n). Meanwhile, as the current air-conditioning state is changed into on, the triggering probability of all relevant guide words in the corresponding air-conditioning field is improved by 2 times.
Scene two (listening to vehicle state): judging whether the vehicle window is closed or not when the current vehicle speed exceeds 50km/h, and if the vehicle window is not closed, improving the probability of triggering the guide word corresponding to the closed window scene in the field of the vehicle window by 2 times; if the vehicle has a fault and other prompts, the probability of the corresponding FAQ voice guide words is improved according to the current fault information, the user is guided to acquire a fault solution through voice, and user anxiety is reduced. And then, a scene of combining the vehicle state and the current wake-up information is detailed in a third scene.
Scene three (listening environment information): the current vehicle machine acquires the temperature outside the vehicle exceeding 26 ℃, judges the current air conditioner state, increases the probability of opening the guide word corresponding to the air conditioner by two times if the air conditioner is not opened, and restores the probability of opening the guide word corresponding to the air conditioner to n and increases the probability of the guide word in the relevant field of air conditioner control by two times if the air conditioner is opened; if the current noise outside the vehicle is larger, or the current noise inside the vehicle is larger and the vehicle speed is higher, judging whether the vehicle window state is open, if so, guiding a user to close the vehicle window through voice.
Scene four (listening application state): the current navigation application is in a road calculation state or a navigation state, the probability of guiding words corresponding to the fields of navigation peripheral searching, road searching, destination arrival time, map interface control and the like is improved by 2 times; if the current navigation state is judged to be the navigation state and the current road condition information is obtained to be the congestion, the guiding language corresponding to the multimedia playing field is correspondingly improved by 2 times, and the user is helped to enter a relaxed state; in addition, the probability of the guide language also changes according to whether the current application is displayed in the foreground, if the current application is in the foreground, the probability of the guide language corresponding to the relevant voice operation of the application is improved by 2 times, and when the application is in the background, the probability of the guide language corresponding to the relevant voice operation is set as n.
In another aspect, the embodiment of the invention provides a voice guidance word arbitration device of a vehicle-mounted system. Referring to fig. 2, a block diagram of a voice guidance word arbitration device of a car system according to an embodiment of the present invention is shown. The apparatus includes modules 201-215.
The candidate guide term obtaining module 201 may be configured to obtain, as the candidate guide term, all the generalized guide terms in the second-level module catalog under the first-level domain catalog, where the first-level domain catalog includes a plurality of domains selected from the navigation domain, the media domain, the vehicle control domain, the message domain, the common problem solving domain, the setting domain, the weather domain, and the vehicle information domain.
The initial trigger probability determination module 202 may be configured to determine an initial trigger probability for each candidate guide word based on the total number of candidate guide words.
The related information acquisition module 203 may be configured to acquire a current voice input of a user, a history dialogue record, an application program operated by a car system and its state, a current vehicle state, and current external environment information of a vehicle.
The dialog context determination module 204 may be configured to determine a current dialog context based on the current speech input and the historical dialog records.
The dialog domain determination module 205 may be configured to determine a dialog domain corresponding to a current dialog context in a primary domain directory.
The first probability improvement module 206 may be configured to improve the trigger probabilities of candidate guide words in all secondary module catalogues in the dialog domain by a first factor.
The application domain determining module 207 may be configured to determine an application domain corresponding to an application in a primary domain directory according to the application and its state.
The second probability improvement module 208 may be configured to improve the trigger probabilities of candidate guide words in all secondary module catalogs in the application domain by a second factor.
The vehicle state domain determining module 209 may be configured to determine a vehicle state domain corresponding to the current vehicle state in the primary domain catalog based on the current vehicle state.
The third probability improvement module 210 may be configured to improve the trigger probabilities of the candidate guide words in all of the two-level module inventory in the vehicle state field by a third factor.
The external environment domain determining module 211 may be configured to determine an external environment domain corresponding to the external environment information in the primary domain directory according to the external environment information.
The fourth probability improvement module 212 may be configured to improve the trigger probabilities of candidate guide words in all secondary module catalogues in the external environment domain by a fourth factor.
The dialog history domain determination module 213 may be configured to determine a dialog history domain corresponding to the current turn dialog history in the primary domain directory based on the current speech input and the history dialog record.
The fifth probability improvement module 214 may be configured to improve the trigger probabilities of candidate guide words in all secondary module catalogues in the conversation history area by a fifth factor.
The guidance word selection module 215 may be configured to randomly select a guidance word from the candidate guidance words according to the trigger probabilities of all the candidate guidance words, as the guidance word to be broadcasted by the vehicle voice system.
It should be noted that, the functions implemented by each module in the voice guidance word arbitration device for a vehicle system according to the embodiment of the present invention are in one-to-one correspondence with each step of the voice guidance word arbitration method for a vehicle system described above, and specific embodiments, examples and beneficial effects thereof refer to the description of the method above.
In another aspect, an embodiment of the present invention provides a vehicle machine, where the vehicle machine includes a memory and a processor, and the memory stores a computer program, and when the computer program is executed by the processor, implements the method for arbitrating the voice guidance word of the vehicle machine system described in any of the foregoing embodiments.
In yet another aspect, embodiments of the present invention provide a storage medium storing computer readable instructions that, when executed by a processor, perform a vehicle-to-machine system voice guidance word arbitration method described in accordance with any of the above embodiments.
The embodiment of the invention provides arbitration logic of intelligent voice guidance words, which can be used for conjecture the voice guidance words needed by a user by combining information such as a vehicle-mounted machine, user intention, vehicle state, environment state and the like in the current scene, so that the user can conveniently execute related operations. Compared with the existing arbitration method, the method uses a new data analysis technology, a context synthesis method, vehicle state sensing, system application state sensing and the like, thereby realizing more intelligent and more accurate guide word selection. The customized guidance is better provided according to the user intention and the vehicle state, so that the user experience of using voice and the voice operation efficiency are improved.
The guide word arbitration method provided by the embodiment of the invention can be applied and implemented in an actual vehicle-to-machine system, and brings practical value to the field of intelligent transportation.
The voice guidance word arbitration scheme of the vehicle-mounted system provided by the embodiment of the invention comprehensively considers a plurality of factors including voice input of a user, context information and current vehicle state. By using a particular arbitration algorithm, the system can intelligently select the most appropriate guide word to meet the current intent and needs of the user.
The invention calculates the probability of possible occurrence of each guide word by using a calculation rule for a plurality of candidate guide words, and carries out random selection according to the calculated probability, wherein the probability is higher, and the probability of being randomly selected is higher, so that the most suitable guide word in the current state can be ensured to be selected, and the guide word can be prevented from being changed when the vehicle is always kept in a certain state.
Since the arbitration scheme considers the context information of the users and the navigation destination, it is possible to provide personalized guidance according to different situations of each user. This will make the user feel more understandable and careable, enhancing the interactive experience between the vehicle system and the user.
The arbitration scheme also carries out the selection of the guide words according to the current vehicle state, and considers the information such as the vehicle speed, the steering signal and the like. Therefore, the rationality and the practicability of the guide words in the current state of the vehicle can be ensured, the interference in the driving process is reduced, the driving safety is improved, and the vehicle state adaptation is optimized.
By providing accurate guide words, the scheme encourages users to use voice operations more frequently, reducing user operation errors and unnecessary repetition. The method and the device can increase the dependence and satisfaction degree of the user on the vehicle-mounted system, improve the accuracy and efficiency of voice operation and improve the operation frequency and accuracy of the user.
The arbitration scheme is applicable to various vehicles including private cars, commercial vehicles, public transportation vehicles and the like. Whether in daily driving or commercial use, the scheme can provide more intelligent, convenient and safe vehicle-mounted voice control functions for drivers and passengers. The method can also be used in other fields requiring voice interaction and guidance word display, and has wide application coverage.
The foregoing description of embodiments of the invention have been presented for the purpose of illustration and is not intended to be exhaustive or to limit the invention to the precise form disclosed. It will be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed as the best mode contemplated for carrying out this invention, but that the invention will include all embodiments falling within the scope of the appended claims.
Claims (10)
1. A method for arbitrating voice guidance words of a vehicle-mounted system, which is characterized by comprising the following steps:
obtaining all generalized guide words in a secondary module catalog under a primary domain catalog as candidate guide words, wherein the primary domain catalog comprises a plurality of domains selected from the navigation domain, the media domain, the vehicle control domain, the message domain, the common problem solving domain, the setting domain, the weather domain and the vehicle information domain;
determining initial trigger probability of each candidate guide word according to the total number of the candidate guide words;
acquiring current voice input of a user, a history dialogue record, an application program operated by a vehicle machine system, a state of the application program, a current vehicle state and current external environment information of a vehicle;
determining a current dialogue context according to the current voice input and the historical dialogue record;
determining the dialogue domain corresponding to the current dialogue context in the primary domain catalog;
the triggering probability of candidate guide words in all the two-level module catalogues in the dialogue field is improved by a first multiple;
determining the application program domain corresponding to the application program in the primary domain catalog according to the application program and the state thereof;
the triggering probability of candidate guide words in all the two-level module catalogues in the application program field is increased by a second multiple;
determining a vehicle state field corresponding to the current vehicle state in the primary field catalog according to the current vehicle state;
the triggering probability of candidate guide words in all the two-level module catalogues in the vehicle state field is improved by a third multiple;
according to the external environment information, determining an external environment field corresponding to the external environment information in the primary field catalog;
the triggering probability of candidate guide words in all the two-level module catalogues in the external environment field is increased by a fourth multiple;
according to the current voice input and the history dialogue record, determining a dialogue history field corresponding to the current round dialogue history in the primary field catalog;
the triggering probability of candidate guide words in all the two-level module catalogues in the dialogue history field is improved by a fifth multiple;
and randomly selecting the guide words from the candidate guide words according to the triggering probability of all the candidate guide words, and using the guide words as the guide words to be broadcasted by the vehicle-mounted voice system.
2. The method according to claim 1, wherein the method further comprises:
judging whether the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue according to the current voice input and the history dialogue record;
in the case that the dialogue field of the current round of dialogue is the same as the dialogue field of the previous round of dialogue, the triggering probability of the candidate guide words in all the two-level module catalogues in the same dialogue field as the dialogue field of the current round of dialogue is increased by a sixth multiple.
3. The method according to claim 1, wherein the method further comprises:
acquiring a current navigation destination under the condition that an application program operated by the vehicle system comprises a navigation application program and the state of the navigation application program is in navigation;
determining whether the current navigation destination is a strange place according to the current navigation destination and a previously stored historical navigation destination;
and in the case that the current navigation destination is a strange place, increasing the broadcasting frequency of the guide word by a preset number.
4. The method of claim 1, wherein the first multiple is equal to or greater than the second multiple, the second multiple is equal to or greater than the third multiple, the third multiple is equal to or greater than the fourth multiple, and the fourth multiple is equal to or greater than the fifth multiple.
5. The method of claim 4, wherein the first multiple is 5 times, the second multiple is 3 times, the third multiple is 2 times, the fourth multiple is 2 times, and the fifth multiple is 1.5 times.
6. The method of claim 1, wherein determining, in the primary domain directory, an external environment domain corresponding to the external environment information based on the external environment information comprises:
and determining the domain of the event meeting the triggering condition under the current external environment as the external environment domain according to the external environment information.
7. The method of claim 1, wherein the state of the application comprises one or more of:
whether the application program is started;
whether the application program is in the foreground or the background of the vehicle-mounted system;
the function currently executed by the application program;
a state corresponding to the interaction information of the user and the application program;
and the state corresponding to the output information of the application program.
8. A vehicle-to-machine system voice guidance word arbitration device, the device comprising:
the system comprises a candidate guide word acquisition module, a first-level domain directory and a second-level domain directory, wherein the candidate guide word acquisition module is configured to acquire all generalized guide words in the second-level module directory under the first-level domain directory as candidate guide words, and the first-level domain directory comprises a plurality of domains selected from the navigation domain, the media domain, the vehicle control domain, the message domain, the common problem solving domain, the setting domain, the weather domain and the vehicle information domain;
the initial trigger probability determining module is configured to determine the initial trigger probability of each candidate guide word according to the total number of the candidate guide words;
the related information acquisition module is configured to acquire current voice input of a user, a history dialogue record, an application program operated by a vehicle-mounted system and state thereof, current vehicle state and current external environment information of a vehicle;
a dialog context determination module configured to determine a current dialog context based on the current speech input and the historical dialog record;
a dialogue domain determining module configured to determine a dialogue domain corresponding to the current dialogue context in the primary domain directory;
the first probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the dialogue field by a first multiple;
an application program domain determining module, configured to determine an application program domain corresponding to the application program in the primary domain directory according to the application program and a state thereof;
the second probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the application program field by a second multiple;
a vehicle state domain determining module configured to determine a vehicle state domain corresponding to the current vehicle state in the primary domain directory according to the current vehicle state;
the third probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the vehicle state field by a third multiple;
an external environment domain determining module configured to determine an external environment domain corresponding to the external environment information in the primary domain directory according to the external environment information;
the fourth probability improvement module is configured to improve the triggering probability of the candidate guide words in all the two-level module catalogues in the external environment field by a fourth multiple;
a conversation history domain determining module configured to determine a conversation history domain corresponding to a current round of conversation history in the primary domain directory according to the current voice input and the history conversation record;
a fifth probability improvement module configured to improve the trigger probability of the candidate guide words in all the two-level module catalogues in the dialogue history field by a fifth multiple;
the guide word selecting module is configured to randomly select guide words from the candidate guide words according to the trigger probabilities of all the candidate guide words, and the guide words are used as the guide words to be broadcasted by the vehicle-to-vehicle voice system.
9. A vehicle comprising a memory and a processor, the memory having stored thereon a computer program which, when executed by the processor, implements the vehicle system voice guidance word arbitration method of any one of claims 1 to 7.
10. A storage medium having stored thereon computer readable instructions which, when executed by a processor, perform the method of vehicle-to-vehicle system voice guidance word arbitration of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311483511.9A CN117238281B (en) | 2023-11-09 | 2023-11-09 | Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311483511.9A CN117238281B (en) | 2023-11-09 | 2023-11-09 | Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117238281A CN117238281A (en) | 2023-12-15 |
CN117238281B true CN117238281B (en) | 2024-03-15 |
Family
ID=89088304
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311483511.9A Active CN117238281B (en) | 2023-11-09 | 2023-11-09 | Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117238281B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107704946A (en) * | 2017-08-29 | 2018-02-16 | 平安科技(深圳)有限公司 | Electronic installation, Voice Navigation needing forecasting method and storage medium |
WO2019174318A1 (en) * | 2018-03-14 | 2019-09-19 | 北京三快在线科技有限公司 | Guide word recommendation |
CN111145721A (en) * | 2019-12-12 | 2020-05-12 | 科大讯飞股份有限公司 | Personalized prompt language generation method, device and equipment |
CN112052318A (en) * | 2020-08-18 | 2020-12-08 | 腾讯科技(深圳)有限公司 | Semantic recognition method and device, computer equipment and storage medium |
CN113378057A (en) * | 2021-06-29 | 2021-09-10 | 珠海必要工业科技股份有限公司 | Information prompting method and device, computer equipment and storage medium |
CN114547266A (en) * | 2022-02-21 | 2022-05-27 | 北京百度网讯科技有限公司 | Training method of information generation model, method, device and equipment for generating information |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080071533A1 (en) * | 2006-09-14 | 2008-03-20 | Intervoice Limited Partnership | Automatic generation of statistical language models for interactive voice response applications |
WO2017112813A1 (en) * | 2015-12-22 | 2017-06-29 | Sri International | Multi-lingual virtual personal assistant |
-
2023
- 2023-11-09 CN CN202311483511.9A patent/CN117238281B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107704946A (en) * | 2017-08-29 | 2018-02-16 | 平安科技(深圳)有限公司 | Electronic installation, Voice Navigation needing forecasting method and storage medium |
WO2019174318A1 (en) * | 2018-03-14 | 2019-09-19 | 北京三快在线科技有限公司 | Guide word recommendation |
CN111145721A (en) * | 2019-12-12 | 2020-05-12 | 科大讯飞股份有限公司 | Personalized prompt language generation method, device and equipment |
CN112052318A (en) * | 2020-08-18 | 2020-12-08 | 腾讯科技(深圳)有限公司 | Semantic recognition method and device, computer equipment and storage medium |
CN113378057A (en) * | 2021-06-29 | 2021-09-10 | 珠海必要工业科技股份有限公司 | Information prompting method and device, computer equipment and storage medium |
CN114547266A (en) * | 2022-02-21 | 2022-05-27 | 北京百度网讯科技有限公司 | Training method of information generation model, method, device and equipment for generating information |
Also Published As
Publication number | Publication date |
---|---|
CN117238281A (en) | 2023-12-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7683767B2 (en) | Control device for controlling in-vehicle unit | |
CN110874202B (en) | Interaction method, device, medium and operating system | |
US10423292B2 (en) | Managing messages in vehicles | |
US9667742B2 (en) | System and method of conversational assistance in an interactive information system | |
CN105824494B (en) | Personalized display system for integrally changing vehicle content and vehicle content management method | |
US20140300494A1 (en) | Location based feature usage prediction for contextual hmi | |
US20140303839A1 (en) | Usage prediction for contextual interface | |
RU2685998C2 (en) | Situational vehicle interface | |
CN102145695A (en) | Emotive advisory system including time agent | |
CN110309277B (en) | Man-machine conversation semantic analysis method and system, vehicle-mounted man-machine conversation method and system, controller and storage medium | |
JP3322140B2 (en) | Voice guidance device for vehicles | |
US20220128373A1 (en) | Vehicle and control method thereof | |
CN110203154B (en) | Recommendation method and device for vehicle functions, electronic equipment and computer storage medium | |
US8498991B2 (en) | Neighborhood guide for semantic search system and method to support local POI discovery | |
EP1690706A1 (en) | Air conditioner control device and air conditioner control method | |
US20200309547A1 (en) | Aligning content playback with vehicle travel | |
US9188455B2 (en) | Navigation system and method for computing the overall costs of a route | |
US20100245068A1 (en) | Vehicle operation diagnosis device, vehicle operation diagnosis method and computer program | |
CN117238281B (en) | Voice guide word arbitration method and device for vehicle-mounted system, vehicle-mounted system and storage medium | |
Torkkola et al. | Sensor Sequence Modeling for Driving. | |
US11333518B2 (en) | Vehicle virtual assistant systems and methods for storing and utilizing data associated with vehicle stops | |
US20230317072A1 (en) | Method of processing dialogue, user terminal, and dialogue system | |
Rogers et al. | Adaptive user interfaces for automotive environments | |
US20220321694A1 (en) | Proactive automotive assistant | |
CN115503639A (en) | Voice processing method, voice interaction method, server and storage medium |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |