Vehicle voice interaction method, server and storage medium

Abstract

The application discloses a vehicle voice interaction method, a server and a storage medium. The vehicle voice interaction method comprises the following steps: receiving a voice request for setting a remark for a point of interest in a navigation map of a vehicle-mounted system by a user forwarded by the vehicle; identifying the voice request to obtain a remark name; establishing a mapping relationship between the remark name and the point of interest; and determining a navigation request for navigating to the remark name after the vehicle-mounted system is woken up by the user next time according to the mapping relationship as a navigation request for navigating to the corresponding point of interest by the user. The application can identify the remark name by identifying the voice request with the remark, establish the mapping relationship between the remark name and the point of interest, and determine the navigation request for navigating to the remark name after the vehicle-mounted system is woken up by the user next time according to the mapping relationship as the navigation request for navigating to the corresponding point of interest by the user, so that the intention of the voice request with the remark can be correctly understood.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent vehicles, in particular to a vehicle voice interaction method, a server and a storage medium. BACKGROUND

[0002] In the current navigation scene, users often use navigation to meet the demand of a certain destination. However, in the current navigation logic, only navigation to a specific address can be realized, for example, a user wants to go to the home of his / her third uncle, and the address of the third uncle's home is "Furongli community in Zhongguancun", and the specific address needs to be completely spoken out in voice. In such a scheme, it is relatively cumbersome to input the specific address in voice every time, and it is not convenient for users to remember. SUMMARY

[0003] The present application provides a vehicle voice interaction method, a server and a storage medium.

[0004] The present application provides a vehicle voice interaction method. The vehicle voice interaction method comprises: receiving a voice request for setting a remark for a point of interest in a navigation map of a vehicle-mounted system by a user forwarded by a vehicle; obtaining a remark name by identifying the voice request; establishing a mapping relationship between the remark name and the point of interest; determining a navigation request for navigating to the remark name after the user next time wakes up the vehicle-mounted system by voice as a navigation request for navigating to the corresponding point of interest according to the mapping relationship.

[0005] In this way, the present application can obtain a remark name by identifying a voice request for setting a remark, and determine a navigation request for navigating to a corresponding point of interest according to a mapping relationship between the remark name and the point of interest after the user next time wakes up the vehicle-mounted system by voice, so as to correctly understand the intention of the voice request with the remark.

[0006] The receiving of the voice request for setting the remark for the point of interest in the navigation map of the vehicle-mounted system by the user forwarded by the vehicle comprises: receiving a current navigation request of the user forwarded by the vehicle; feeding back a navigation search result containing a plurality of the points of interest in the navigation map of the vehicle-mounted system to the user for confirmation according to the current navigation request; and receiving a voice request for setting a remark for the confirmed point of interest by the user forwarded by the vehicle.

[0007] In this way, the server side can receive a voice request for setting a remark for a point of interest in a navigation map of a vehicle-mounted system by a user forwarded by a vehicle, so as to determine the point of interest in the navigation map of the vehicle-mounted system corresponding to the voice request related to the remark.

[0008] After the step of forwarding the voice request of setting a note for a point of interest in a navigation map of a vehicle system by a user of a receiving vehicle, the vehicle voice interaction comprises: slot identification on the voice request; in a case where the slot identification result does not include point of interest slot information, if it is determined by a navigation application program interface of the vehicle that the voice request has a destination field, setting a note for a point of interest corresponding to the destination field.

[0009] In this way, the application can determine whether the voice request of the user interacting with the vehicle contains a destination field by calling the navigation application program interface of the vehicle, so as to determine the point of interest corresponding to the destination field and set a note for the point of interest, thereby facilitating direct navigation to the point of interest corresponding to the destination field in the next voice request of the user containing the same destination field.

[0010] In the case where the slot identification result does not include point of interest slot information, if it is determined by the navigation application program interface of the vehicle that the voice request has a destination field, a note is set for a point of interest corresponding to the destination field, comprising: if it is determined by the navigation application program interface of the vehicle that the voice request has a plurality of destination fields, the user is asked to select a desired destination field through voice broadcast; and a note is set for a point of interest corresponding to the destination field confirmed by the user.

[0011] In this way, the unique destination field can be determined in the case where the voice request contains a plurality of destination fields.

[0012] In the case where the slot identification result does not include point of interest slot information, if it is determined by the navigation application program interface of the vehicle that the voice request has a destination field, a note is set for a point of interest corresponding to the destination field, comprising: if it is determined by the navigation application program interface of the vehicle that the voice request has one destination field, a note is set for a point of interest corresponding to the destination field.

[0013] In this way, when it is determined by the navigation application program interface of the vehicle that the destination field of the voice request does not exist, the user is fed back through voice broadcast that the point of interest is not found, so that the user can be fed back in time and adjust the statement in time until the destination field is determined and the point of interest corresponding to the destination field is determined.

[0014] The vehicle voice interaction method comprises: if it is determined by the navigation application program interface of the vehicle that the destination field of the voice request does not exist, the user is fed back through voice broadcast that the point of interest is not found.

[0015] Thus, when the navigation application interface of the vehicle determines that the destination field of the voice request does not exist, the feedback of not finding the point of interest through voice broadcast can timely feedback to the user and timely adjust the statement until the destination field is determined and the point of interest corresponding to the destination field is determined.

[0016] After the step of receiving the voice request in which the user forwards the remark set for the point of interest in the navigation map of the vehicle-mounted system by the vehicle, the vehicle voice interaction method comprises: slot identification of the voice request; in the case that the result of the slot identification includes point of interest slot information, setting a remark for the point of interest obtained by navigation search on the point of interest slot information.

[0017] Thus, in the case that the voice request includes a remark name and a point of interest, the point of interest slot information in the slot identification result is directly navigated to obtain the point of interest, and a remark is set for the point of interest.

[0018] In the case that the result of the slot identification includes point of interest slot information, setting a remark for the point of interest obtained by navigation search on the point of interest slot information, comprising: if the point of interest slot information navigation search obtains multiple points of interest, asking the user for the desired point of interest through voice broadcast; setting a remark for the point of interest confirmed by the user's reply.

[0019] Thus, in the case that the point of interest slot information navigation search obtains multiple points of interest, the user is asked for the desired point of interest through voice broadcast, and a remark is set for the point of interest confirmed by the user's reply.

[0020] In the case that the result of the slot identification includes point of interest slot information, setting a remark for the point of interest obtained by navigation search on the point of interest slot information, comprising: if the point of interest slot information navigation search obtains one point of interest, setting a remark for the point of interest.

[0021] Thus, in the case that the point of interest slot information navigation search obtains only one point of interest, a remark is directly set for the point of interest.

[0022] In the case that the result of the slot identification includes point of interest slot information, setting a remark for the point of interest obtained by navigation search on the point of interest slot information, comprising: if the point of interest slot information navigation search does not exist search result, feedback of not finding the point of interest through voice broadcast.

[0023] Thus, in the case that the point of interest slot information navigation search does not exist search result, the feedback of not finding the point of interest through voice broadcast can timely feedback to the user and timely adjust the statement until the point of interest is found.

[0024] The application provides a server. The server comprises a processor and a memory, and the memory stores a computer program. When the computer program is executed by the processor, the vehicle voice interaction method of any one of the above embodiments is implemented.

[0025] Therefore, the application can obtain the remark name by recognizing the voice request with the remark, and determine the navigation request of the user to the corresponding interest point by establishing the mapping relationship between the remark name and the interest point, so as to correctly understand the intention of the voice request with the remark.

[0026] The application also provides a nonvolatile computer readable storage medium comprising a computer program. When the computer program is executed by one or more processors, the vehicle voice interaction method of any one of the above embodiments is implemented.

[0027] Therefore, the application can obtain the remark name by recognizing the voice request with the remark, and determine the navigation request of the user to the corresponding interest point by establishing the mapping relationship between the remark name and the interest point, so as to correctly understand the intention of the voice request with the remark.

[0028] Additional aspects and advantages of the application will be described in the following description, become apparent from the following description, or be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0029] The above and / or additional aspects and advantages of the application can become apparent and be easily understood from the description of the embodiments, taken in conjunction with the accompanying drawings, in which:

[0030] Figure 1 is one of the flowcharts of the guiding method of the application;

[0031] Figure 2 is another of the flowcharts of the guiding method of the application;

[0032] Figure 3 is a scene diagram of the guiding method of the application;

[0033] Figure 4 is one of the flowcharts of the vehicle voice interaction method of the application;

[0034] Figure 5 is another of the scene diagrams of the vehicle voice interaction method of the application;

[0035] Figure 6Fig. 3 is a flowchart of a third process of the vehicle voice interaction method of the present application;

[0036] Figure 7 Fig. 4 is a flowchart of a fourth process of the vehicle voice interaction method of the present application;

[0037] Figure 8 Fig. 5 is a flowchart of a fifth process of the vehicle voice interaction method of the present application;

[0038] Figure 9 Fig. 6 is a flowchart of a sixth process of the vehicle voice interaction method of the present application;

[0039] Figure 10 Fig. 7 is a flowchart of a seventh process of the vehicle voice interaction method of the present application;

[0040] Figure 11 Fig. 8 is a flowchart of an eighth process of the vehicle voice interaction method of the present application;

[0041] Figure 12 Fig. 9 is a flowchart of a ninth process of the vehicle voice interaction method of the present application.DETAILED DESCRIPTION

[0042] The present application is described in detail below, with examples of the embodiments shown in the accompanying drawings, wherein the same or similar reference numerals are used throughout to designate the same or similar elements or elements having the same or similar functions. The embodiments described below by reference to the drawings are exemplary and are for the purpose of explanation only, and are not to be understood as limiting the present application.

[0043] The present application is described in detail below, with examples of the embodiments shown in the accompanying drawings, wherein the same or similar reference numerals are used throughout to designate the same or similar elements or elements having the same or similar functions. The embodiments described below by reference to the drawings are exemplary and are for the purpose of explanation only, and are not to be understood as limiting the present application.

[0044] When a user uses voice navigation, there are many points of interest that the user will visit at a fixed frequency. For example, the gym is visited twice a week, the parents' home is visited once every half month, the dentist is visited once a month, the friend's home is often visited, the third uncle's home is occasionally visited, and the son's school is visited every day. These points of interest have special meanings for the user, and the meanings are more concise than the complete description of the points of interest. Therefore, we hope that through the function described in the patent, when the user goes to the above-mentioned places again, the user does not need to say the complete address, but can complete the navigation action by directly saying the nickname. In this way, the complexity of voice navigation is reduced, and the voice assistant appears more human. For example, assume that a user will take the son to the sports center to play badminton every Sunday. For navigation to the point of interest, the user navigates to the place is cumbersome. For example, navigate to XXX district, XXX road, XXX sports center, and XXX door. However, by setting a nickname for the point of interest, the user can navigate to the place more simply and more humanly. For example, navigate to the badminton court.

[0045] In the case where the user does not use the function of "navigating to the nickname" of the vehicle, the user cannot know in time that the vehicle can navigate to the nickname. For example, the user wants to navigate to the third uncle's home every time, and every time the user directly searches for the specific address of the third uncle's home, which affects the user's navigation experience.

[0046] In order to enable the user to say [navigate to the "nickname"] to complete the navigation task, the user needs to give the point of interest a nickname and collect it first. At this time, the user can be provided with a nickname for a point of interest according to the user's navigation frequency and the nickname of other users, so that the user knows that there is a function of navigating to the nickname.

[0047] Please refer to Figure 1 The present application provides a guiding method. The guiding method comprises:

[0048] 011: receiving first report information reported by a vehicle, the first report information being initiated by a user in the vehicle through a vehicle-mounted system and being used for navigation to a point of interest;

[0049] 013: querying, according to the first report information, collection and nickname information of an address of the point of interest by all users in a nickname information set;

[0050] 015: determining, according to a query result and a type of the point of interest, voice broadcast content to be fed back according to a predetermined rule;

[0051] 017: feeding back the voice broadcast content to the vehicle to broadcast to the user in the vehicle, so as to guide the user to add a nickname to the point of interest after the user completes the collection of the point of interest through a map of the vehicle-mounted system.

[0052] The application further provides a server. The server comprises a processor and a memory, and the memory stores a computer program. The processor is configured to receive first report information reported by a vehicle, the first report information being initiated by a user in the vehicle through a vehicle-mounted system and being used for navigation to a point of interest; query, according to the first report information, all collection and comment information of the point of interest by all users in a comment information set; determine, according to a query result and a type of the point of interest, voice broadcast content to be fed back according to a predetermined rule; and send the voice broadcast content to the vehicle to be broadcast to the user in the vehicle, so as to guide the user to add a comment name to the point of interest after the user completes the collection of the point of interest through a map of the vehicle-mounted system.

[0053] The server of the application is used for realizing voice interaction between the vehicle and the user. In the case that the information reported by the vehicle meets a condition, the server can send voice broadcast to the vehicle and the user in the cabin of the vehicle to realize voice interaction. Or the server can receive a voice request forwarded by the vehicle, process the voice request, and send a control instruction and / or voice broadcast according to natural language understanding to complete voice interaction.

[0054] Specifically, for the guiding method of the application, first, the server end receives first report information reported by the vehicle, the first report information being initiated by a user in the vehicle through a vehicle-mounted system and being used for navigation to a point of interest.

[0055] The user in the vehicle initiates navigation to the point of interest through the vehicle-mounted system, which means that the user in the vehicle initiates navigation to a certain place through voice search of a map of the vehicle-mounted system. The point of interest (POI) refers to a specific address in the map of the vehicle-mounted system. In a geographic information system, one point of interest can be a house, a shop, a mailbox, a bus stop, etc.

[0056] It can be understood that the map of the vehicle-mounted system can be a map application installed on the vehicle side. Taking a Gaode map application installed on the vehicle side as an example, the map server corresponding to the Gaode map application has a collection folder associated with a user account, which is used to store points of interest added by the user. The vehicle can realize the functions of positioning, route planning and navigation to the point of interest by communicating with the map server through the Gaode map application.

[0057] When the user in the vehicle initiates navigation to the point of interest through the vehicle-mounted system, the server end of the application can simultaneously receive the first report information reported by the vehicle, the first report information being initiated by the user through the vehicle-mounted system and being used for navigation to the point of interest. The first report information comprises identity information of the user, a message of the user initiating navigation to the point of interest, a name of the point of interest, a current search time of the user to the point of interest, a historical search time and a search frequency of the user before the current search time.

[0058] Understandably, before the user initiates the navigation function to the point of interest, the server receives the first report information reported by the vehicle, and needs to judge whether the navigation at this time meets the condition of triggering "voice push". If the condition of triggering "voice push" is met, voice push is performed, that is, the voice broadcast content is fed back to the user.

[0059] Therefore, the server side needs to query the collection and note information of the address of the point of interest of all users in the note information set according to the first report information. Among them, all users refer to all users who have account information on the server side.

[0060] For example, the current user can initiate navigation to the point of interest through voice request: "navigate to XX hotel", at this time the vehicle map searches "XX hotel" according to the voice request, and queries that other user 1 has collected the address of "XX hotel", and the note information is that "XX hotel" is noted as "company", and other user 2 has collected the address of "XX hotel", and the note information is that "XX hotel" is noted as "frequently visited hotel".

[0061] Then, according to the query result and the type of the point of interest, the voice broadcast content to be fed back is determined according to the predetermined rule. Among them, the application can divide the point of interest into the type of point of interest including entertainment place type, company type and home type according to the scenario that the user makes notes for the point of interest, realizing the self-defined classification of the point of interest, so as to guide the user to make notes in combination with voice push. The query result is that the names collected and noted by other user 1 and user 2 for the "XX hotel" respectively include "company" and "frequently visited hotel". The predetermined rule refers to the way of determining the feedback voice broadcast content according to whether the user has collected the point of interest, the point of interest and the query result.

[0062] Understandably, the purpose of distinguishing the type of the noted point of interest is to push different voice broadcast contents for different types of points of interest, such as the voice broadcast content corresponding to the company type point of interest is "try to navigate to the company", and the entertainment place type point of interest is more inclined to guide the user to note as "frequently visited supermarket" and "frequently visited cinema", and the corresponding voice broadcast content is "try to say to me: navigate to frequently visited cinema". Different types of points of interest hope that the user establishes different forms of memory.

[0063] More specifically, for example, the type of the point of interest "XX hotel" belongs to the entertainment place type, combined with the above query result, if the user has completed the collection of the point of interest through the map of the vehicle-mounted system, the voice broadcast content to be fed back can be determined according to the predetermined rule as "there are 5 car owners who have noted the address, the note names are 'frequently visited hotel' and 'company', what do you want to note as".

[0064] Finally, the voice broadcast content is sent to the vehicle to be broadcast to the user in the vehicle to guide the user to add a nickname to the point of interest after the user completes the collection of the point of interest through the map of the vehicle-mounted system.

[0065] In this way, when the user in the vehicle initiates navigation to the point of interest through the vehicle-mounted system, the user can be guided to first complete the collection of the point of interest through the map of the vehicle-mounted system and then add a nickname to the point of interest through voice broadcast, so that the user can know and use the function of the vehicle that can navigate to the nickname in time, which provides convenience for the user and improves the user experience.

[0066] After step 011, the guiding method comprises: processing the name of the point of interest, the search time and the search times of the user for the point of interest to obtain the type of the point of interest recommended for the nickname.

[0067] The processor is configured to process the name of the point of interest, the search time and the search times of the user for the point of interest to obtain the type of the point of interest recommended for the nickname.

[0068] As can be seen from the foregoing, the type of the point of interest can include an entertainment place type, a company type and a home type. Since the first report information includes the name of the point of interest, the current search time and the historical search time of the user for the point of interest and the search times of the user before the current search time, the type of the point of interest recommended for the nickname by the user is determined by the name of the point of interest, the search time and the search times of the user for the point of interest.

[0069] It can be understood that if the user has searched a large number of times within a predetermined time period before the current search time for the point of interest, for example, the search times reach 2 or more than 2 within 1 month before the current time point, it indicates that the point of interest is an address frequently used by the user, and therefore, the user can be recommended to add a nickname to the point of interest, thereby improving the user experience.

[0070] For example, when the name of the point of interest is “XX Hotel”, the user has searched for the point of interest a total of 4 times within 1 month before the current search time, and the type of the point of interest recommended for the nickname by the server to the user can be the entertainment place type.

[0071] In this way, after receiving the first report information of the vehicle-mounted system initiated by the user in the vehicle to navigate to the point of interest reported by the vehicle, the server can process the name of the point of interest, the search time and the search times of the user for the point of interest to obtain the type of the point of interest recommended for the nickname by the user.

[0072] In addition, step 015 comprises: in the case that the user has collected the interest point and no nickname has been added to the interest point, if the user has performed the operation of initiating navigation to the interest point within the first preset time, determining the first voice broadcast content to be fed back according to the query result and the type of the interest point according to a predetermined rule.

[0073] The processor is configured to, in the case that the user has collected the interest point and no nickname has been added to the interest point, if the user has performed the operation of initiating navigation to the interest point within the first preset time, determine the first voice broadcast content to be fed back according to the query result and the type of the interest point according to a predetermined rule.

[0074] The first preset time refers to a preset time before the time point of currently initiating navigation to the interest point, for example, the first preset time is 1 month as described above, or can be other time ranges, which are not limited herein.

[0075] The predetermined rule herein can be to obtain the first voice broadcast content by fusing the query result and the type of the interest point.

[0076] In the case that the user has collected the interest point and no nickname has been added to the interest point, if the user has performed the operation of initiating navigation to the interest point within the first preset time, for example, the first preset time is 1 month, then within 1 month before the time point of currently initiating navigation to the interest point, it is found that the user has performed the operation of initiating navigation to the interest point, the number of times of initiation can be 1 or more than 1, and the query result of querying all the user's collection and nickname information of the address of the interest point in the nickname information set is that 5 car owners collect and note the address as a company, and other users collect and note the name of the address as 'company', 'friend's company', 'husband's company', etc., and the type of the interest point belongs to the company type, then the first voice broadcast content to be fed back is determined according to the query result and the type of the interest point according to a predetermined rule, for example, the content of the first voice broadcast can be: "5 car owners have noted the address as a company, and other users have noted the name of the address as 'company', 'friend's company', 'husband's company', etc., what do you want to note as?", thereby guiding the user to note the interest point.

[0077] In this way, in the case that the user has collected the interest point and no nickname has been added to the interest point, if the user has performed the operation of initiating navigation to the interest point within the first preset time, the first voice broadcast content to be fed back is determined according to the query result and the type of the interest point according to a predetermined rule, thereby guiding the user to note the interest point, so that the user can directly say the nickname of the interest point to initiate navigation next time, and the user experience is improved.

[0078] It can be understood that when it is determined to feed back the first voice broadcast content to the user, the selected nickname can have multiple candidate nicknames, at this time, the server of the present application can select the nickname same as the type of the point of interest and feed it back to the user.

[0079] Specifically, step 015 further comprises: selecting multiple candidate nicknames according to the nickname information same as the type of the point of interest in the query result according to a predetermined rule to determine the first voice broadcast content to be fed back.

[0080] The processor is configured to select multiple candidate nicknames according to the nickname information same as the type of the point of interest in the query result according to a predetermined rule to determine the first voice broadcast content to be fed back.

[0081] The predetermined rule here can refer to a rule determined according to the degree of coincidence of the nickname in the query result with the type of the point of interest, the more the nickname in the query result coincides with the type of the point of interest, the greater the probability of being fed back to the user as a candidate nickname, and correspondingly, the earlier the order of voice broadcasting the nickname.

[0082] For example, the point of interest initiated by the user for navigation is "XX office building", i.e. the type of the point of interest is company type, and the query result of the collection and remark information of all users on the address of the point of interest "XX office building" in the remark information set is that 4 car owners collect and remark the address as a company, and other users collect and remark the name of the address as 'company', 'friend's company', 'husband's company', then according to the type of the point of interest being company type, three candidate nicknames "company", "friend's company", "husband's company" can be selected to recommend to the user, i.e. the first voice broadcast content to be fed back can be "there are 4 car owners who have remarked the address as a company, and other users have remarked the name of the address as 'company', 'friend's company', 'husband's company', etc., what do you want to remark as", thereby guiding the user to make remarks on the point of interest.

[0083] In this way, the user can select multiple candidate nicknames same as the type of the point of interest in the query result and feed them back to the user through voice broadcast.

[0084] It can be understood that when it is determined to feed back the first voice broadcast content to the user, the user can not have collected the point of interest, and can not have made remarks, at this time, the user needs to be reminded to collect the point of interest first, and then reminded to make remarks on the point of interest.

[0085] Specifically, step 015 further comprises: in the case where the user has not collected the point of interest, determining the first voice broadcast content to be fed back for prompting the user to collect the point of interest.

[0086] The processor is configured to determine first voice broadcast content to be fed back for prompting the user to collect the point of interest if the user does not collect the point of interest.

[0087] For example, if the point of interest is not collected, the first voice broadcast content is "XX drivers have collected the address, you can try to say 'collect the address'".

[0088] In this way, if the user has not collected the point of interest, the user can be reminded to collect the point of interest first, and then the user can be reminded to set a remark for the point of interest for the next navigation, so that the user can directly say the remark name of the point of interest to initiate navigation, thereby improving user experience.

[0089] In addition, when the user completes the operation of adding a remark name for the point of interest, the server can receive the information in real time, generate second voice broadcast content according to the remark name determined by the user, and deliver the second voice broadcast content to the vehicle end to feed back to the user in the vehicle by voice broadcast, so as to prompt the user to input a navigation request for navigation to the point of interest after the next voice wake-up of the vehicle-mounted system.

[0090] Please refer to Figure 2 After step 017, the guiding method comprises:

[0091] 0171: receiving second report information reported by the vehicle that a user in the vehicle completes adding a remark name for the point of interest;

[0092] 0172: determining second voice broadcast content to be fed back according to the remark name in the second report information;

[0093] 0173: delivering the second voice broadcast content to the vehicle to broadcast to the user in the vehicle, so as to prompt the user to input a navigation request for navigation to the point of interest after the next voice wake-up of the vehicle-mounted system.

[0094] The processor is configured to receive second report information reported by the vehicle that a user in the vehicle completes adding a remark name for the point of interest; determine second voice broadcast content to be fed back according to the remark name in the second report information; and deliver the second voice broadcast content to the vehicle to broadcast to the user in the vehicle, so as to prompt the user to input a navigation request for navigation to the point of interest after the next voice wake-up of the vehicle-mounted system.

[0095] The second report information comprises identity information of the user, a message that the user completes adding a remark name for the point of interest, an address of the point of interest, and a remark name added by the point of interest.

[0096] For example, if the user adds a note of "my company" to the point of interest, the server can receive the second report information reported by the vehicle in real time, and can further determine the second voice broadcast content to be fed back to the user according to the note name in the second report information. The second voice broadcast content can be "you can try 'navigate to my company' next time".

[0097] Then, the server will issue the second voice broadcast content "you can try 'navigate to my company' next time" to the vehicle to broadcast to the user in the vehicle, prompting the user to input the navigation request of "navigate to my company" next time the voice wakes up the vehicle system.

[0098] Therefore, when the user completes the operation of adding a note to the point of interest, the server can receive the second report information reported by the vehicle in real time, and generate the second voice broadcast content according to the note name determined by the user, and issue it to the vehicle to broadcast to the user in the vehicle, prompting the user to input the navigation request of "navigate to my company" next time the voice wakes up the vehicle system.

[0099] It can be understood that, in addition to the voice broadcast method, the vehicle can also guide the user to collect and add notes to the point of interest through card pushing on the center screen of the vehicle. For example, when the user does not perform navigation, the voice assistant of the vehicle will suddenly broadcast, which will be very strange. Therefore, when the user does not perform navigation, the vehicle can also provide card pushing to guide the user to collect and add notes to the point of interest.

[0100] Specifically, after step 013, the guiding method comprises: determining the push card content to be fed back according to the query result and the type of the point of interest according to a predetermined rule; and

[0101] The push card content is issued to the vehicle to display to the user in the vehicle, so as to guide the user to add a note to the point of interest after collecting the point of interest through the map of the vehicle system.

[0102] The processor is configured to determine the push card content to be fed back according to the query result and the type of the point of interest according to a predetermined rule; and issue the push card content to the vehicle to display to the user in the vehicle, so as to guide the user to add a note to the point of interest after collecting the point of interest through the map of the vehicle system.

[0103] The predetermined rule here can refer to a rule determined according to the degree of coincidence between the nickname in the query result and the type of the point of interest. The more coincident the nickname in the query result is with the type of the point of interest, the greater the probability of feeding back the nickname to the user as a candidate nickname, and correspondingly, the earlier the order of displaying the nickname in the recommendation card of the vehicle center screen, as shown in FIG. 8. The card content shows the nicknames "company", "friend's company" and "husband's company" which are coincident with the type of the point of interest being a company. Figure 3

[0104] According to the query result and the type of the point of interest, the push card content to be fed back is determined according to the predetermined rule, that is, for example, the point of interest for which the user initiates navigation is "XX office building", that is, the type of the point of interest is a company, and the query result of the collection and the note information of all users on the address of the point of interest "XX office building" in the note information set is that 4 car owners collect and note the address as a company, and the names noted by other users on the address are "company", "friend's company" and "husband's company". According to the type of the point of interest being a company, the three candidate nicknames "company", "friend's company" and "husband's company" can be selected to be recommended to the user, that is, the card content to be fed back can be "There are 4 car owners who have noted the address as a company, and the names noted by other users on the address are 'company', 'friend's company', 'husband's company' and the like. You can select one of the above nicknames to set a note for the current point of interest", and the card content can be as shown in FIG. 8. Figure 3

[0105] Then, after the card content is determined on the server side, the server pushes the card content to the vehicle to display to the user in the vehicle, so as to guide the user to add a note for the point of interest after completing the collection of the point of interest through the map of the vehicle-mounted system.

[0106] In this way, the application can also provide a card pushing manner for the user, so as to guide the user to collect and note the point of interest and improve the user experience.

[0107] In addition, since the point of interest selected by the user cannot be obviously determined by the server to which type it is more coincident, at this time, the point of interest selected by the user needs to be compared with each type of point of interest for similarity. Through the similarity, the server can determine to push the card content corresponding to each type of point of interest to the vehicle to display to the user in the vehicle in the order from high to low or in the order from low to high according to the similarity, so as to guide the user to add a note for the point of interest after completing the collection of the point of interest through the map of the vehicle-mounted system.​​

[0108] Specifically, the step 0142 comprises: according to the similarity between the interest point and each type of interest point, sequentially issuing the push card content corresponding to each type of interest point to the user in the vehicle in a preset order within a second preset time to guide the user to add a nickname for the interest point after completing the collection of the interest point through the map of the vehicle-mounted system.

[0109] The processor is configured to sequentially issue the push card content corresponding to each type of interest point to the user in the vehicle in a preset order within a second preset time according to the similarity between the interest point and each type of interest point, to guide the user to add a nickname for the interest point after completing the collection of the interest point through the map of the vehicle-mounted system.

[0110] The similarity between the interest point and each type of interest point can be matched according to a similarity comparison model to obtain the similarity.

[0111] The preset order comprises an order from high to low similarity or an order from low to high similarity.

[0112] The second preset time can refer to a time after the server side determines the push card content corresponding to the type of the interest point, for example, the second preset time can be 1s, 1.2s, 2s, 2.4s, 3s, 4s, 5s, 6s, 7s, and 8s, etc., which is not limited herein.

[0113] For example, if the similarity matching result of the user-selected interest point and each type of interest point obtained according to the interest point classification model is: company type: 90%, indicating that the similarity between the interest point and the company type interest point is 90%; home type: 50%, indicating that the similarity between the interest point and the home type interest point is 50%; entertainment place type: 70%, indicating that the similarity between the interest point and the entertainment place type interest point is 70%.

[0114] Taking the order from high to low similarity as an example, it can be known from the above similarity matching result that the user-selected interest point is most consistent with the company type interest point, the user-selected interest point is relatively consistent with the entertainment place type interest point, and the user-selected interest point has a smaller probability of being consistent with the home type interest point.

[0115] Therefore, the server can preferentially push the company type card to the user, and the content of the company type card can be: other vehicle owners have already noted “XXX company” as: friend's company, my company, try to select a nickname to make a note.

[0116] After pushing the company type card, the entertainment place type card is pushed in the second preset time, and the content of the entertainment place card can be that other vehicle owners have noted "XXX market" as "frequently visited cinema", "favorite cinema", and the like, and a note name is selected to make a note.

[0117] After pushing the entertainment place type card, the home type card is pushed in the second preset time, and the content of the home type card can be that other vehicle owners have noted "XXX road XXX number" as "friend's home", "my home", "third uncle's home", and the like, and a note name is selected to make a note.

[0118] In this way, the server can sequentially issue the push card content corresponding to the type of each interest point to the vehicle, such as the user in the vehicle, in the order from high to low of the similarity or in the order from low to high of the similarity, to guide the user to add a note name to the interest point after completing the collection of the interest point through the map of the vehicle-mounted system.

[0119] In addition, in the case that the user has collected a certain address of the map but has not made a note, at this time, the user can be reminded to make a note on the point according to the user's navigation to the interest point in the collection without a note, so that the user understands the function of navigating to the interest point.

[0120] Specifically, please refer to Figure 4 The application provides a vehicle voice interaction method. The vehicle voice interaction method comprises the following steps:

[0121] 021: receiving a current navigation request forwarded by a vehicle; wherein the current navigation request is a navigation request initiated by a user in a vehicle cabin through voice to navigate to an interest point in a collection associated with a user account in a map server;

[0122] 022: generating a voice broadcast indicating that a navigation route has been selected according to a route confirmed by the user and issuing the voice broadcast to the vehicle;

[0123] 023: after determining that the interest point of the current navigation request does not contain a note name, generating a voice broadcast guiding the user to add a note name to the interest point and issuing the voice broadcast to the vehicle, so that the user adds a note name to the interest point;

[0124] 024: when the user in the vehicle cabin adds a note name to the interest point, adding attribute information containing the note name to the corresponding interest point in the collection associated with the user account in the map server;

[0125] 025: receiving a next navigation request initiated by the user through voice to navigate to the note name and forwarded by the vehicle;

[0126] 026: Intention recognition and slot recognition are performed on the next navigation request, and after confirming that the intention recognition result is a navigation intention, it is determined whether the interest point in the slot recognition result is completely matched with the attribute information containing the nickname;

[0127] 027: The matching result is sent to the map server to complete the navigation in response to the next navigation request.

[0128] The processor is configured to receive a current navigation request forwarded by the vehicle; the current navigation request is a navigation request initiated by a user in the vehicle cabin to an interest point in a collection associated with a user account in the map server through voice; generate a voice broadcast indicating that the navigation route has been selected according to the route confirmed by the user and send it to the vehicle; after determining that the interest point of the current navigation request does not contain a nickname, generate a voice broadcast guiding the user to add a nickname to the interest point and send it to the vehicle, so that the user adds a nickname to the interest point; after the user in the vehicle cabin adds a nickname to the interest point, add attribute information containing the nickname to the corresponding interest point in the collection associated with the user account in the map server; receive a next navigation request initiated by the user through voice to the nickname and forwarded by the vehicle; perform intention recognition and slot recognition on the next navigation request, and after confirming that the intention recognition result is a navigation intention, determine whether the interest point in the slot recognition result is completely matched with the attribute information containing the nickname; send the matching result to the map server to complete the navigation in response to the next navigation request.

[0129] Specifically, first, a current navigation request forwarded by the vehicle is received. The current navigation request is a navigation request initiated by a user in the vehicle cabin to an interest point in a collection associated with a user account in the map server through voice. For example, the interest point in the collection associated with the user account in the map server is "XXX District XXX Road XXX Sports Center XXX Gate", and the current navigation request initiated by the user through voice is "navigate to XXX District XXX Road XXX Sports Center XXX Gate". At this time, the map on the vehicle side will search for one or more routes according to the current navigation request, and can pop up a pop-up window of the route list.

[0130] Then, a voice broadcast indicating that the navigation route has been selected is generated according to the route confirmed by the user and is sent to the vehicle. That is, for example, when there are route 1, route 2 and route 3 in the route list, the pop-up window can be as shown in FIG. 1. Figure 5 When the user clicks route 1 as the navigation route, the server side will generate a voice broadcast indicating that the navigation route has been selected according to the route 1 confirmed by the user, which can be "route 1 has been selected for you", and send the voice broadcast "route 1 has been selected for you" to the vehicle, so that the vehicle side initiates navigation according to the route 1.

[0131] Then, after the vehicle end receives the voice broadcast indicating that the navigation route has been selected according to the user's confirmation of the route generation, the voice server end can trigger an additional voice broadcast to guide the user to add a nickname to the point of interest.

[0132] That is, after the server end determines that the point of interest of the current navigation request does not contain a nickname, the server end generates a voice broadcast guiding the user to add a nickname to the point of interest and sends it to the vehicle, so that the user can add a nickname to the point of interest. For example, after the server end determines that the point of interest "XXX District XXX Road XXX Sports Center XXX Gate" in the current navigation request "navigate to XXX District XXX Road XXX Sports Center XXX Gate" does not contain a nickname, the server end generates a voice broadcast guiding the user to add a nickname to the point of interest, for example:

[0133] "By the way, I can also help you give the destination a name. For example, you can tell me [give the destination the nickname Old Place], and I will remember it."

[0134] Or:

[0135] "Oh, by the way, you can give this destination a nickname. Try saying [give the destination the nickname Old Place] and unlock my learning skills."

[0136] Or:

[0137] "By the way, you can tell me [give the destination the nickname Old Place], and next time you can just say [navigate to Old Place]. Remember, it's [give the destination the nickname Old Place]."

[0138] After the user in the vehicle cabin adds a nickname to the point of interest, the map server adds attribute information containing the nickname to the corresponding point of interest in the user's account associated with the collection, so that the user can complete navigation according to the attribute information of the point of interest and the contained nickname the next time the user initiates navigation to the nickname function. It can be understood that each point of interest of the present application can have multiple attribute information, including specific address, user account, and nickname attribute information. For example, after the user in the vehicle cabin adds the nickname "Badminton Hall" to the point of interest "navigate to XXX District XXX Road XXX Sports Center XXX Gate", the map server can add attribute information containing the nickname "Badminton Hall" to the corresponding point of interest "XXX District XXX Road XXX Sports Center XXX Gate" in the user's account associated with the collection.

[0139] Then, the next time the user needs to initiate navigation to the nickname, the vehicle end can receive the next navigation request initiated by the user through voice navigation to the nickname forwarded by the vehicle, perform intent recognition and slot recognition on the next navigation request, and compare whether the interest point in the slot recognition result is completely matched with the attribute information containing the nickname after confirming that the intent recognition result is a navigation intent. Then the vehicle end sends the matching result to the map server to complete the navigation in response to the next navigation request.

[0140] In this way, the application can trigger the voice server to add a voice broadcast to guide the user to add a nickname to the interest point after the vehicle end receives the voice broadcast indicating that the navigation route has been selected according to the user's confirmed route generation instruction. Thus, the user can use the navigation function of the vehicle to add a nickname to the interest point next time.

[0141] It should be noted that the frequency of the voice broadcast guiding the user to add a nickname to the interest point can be the frequency set by default by the vehicle-mounted system, for example, it can be once a day or other frequencies, which are not limited herein.

[0142] In addition, the frequency of the voice broadcast can also be limited, for example, for the same type of navigation request, the voice broadcast to guide the user to add a nickname to the interest point can be triggered at most once every 14 days.

[0143] For the next navigation request of the user to navigate to a single destination, for example, the next navigation request is "navigate to the badminton hall", the interest point in the slot recognition result and the attribute information containing the nickname can be completely matched. Correspondingly, step 027 comprises: sending the matching result of the complete match between the interest point in the slot recognition result and the attribute information containing the nickname to the map server, and the map server and the map communication of the vehicle-mounted system complete the navigation to the interest point in the collection folder corresponding to the attribute information containing the nickname.

[0144] The processor is configured to send the matching result of the complete match between the interest point in the slot recognition result and the attribute information containing the nickname to the map server, and the map server and the map communication of the vehicle-mounted system complete the navigation to the interest point in the collection folder corresponding to the attribute information containing the nickname.

[0145] When the next navigation request initiated by the user through voice to navigate to the nickname is "navigate to the badminton hall", the intent recognition result obtained by performing intent recognition on the navigation request is a navigation intent, and the interest point in the slot recognition result obtained by performing slot recognition is "badminton hall", the interest point in the slot recognition result is completely matched with the attribute information containing the nickname "badminton hall", and the matching result can be sent from the voice server to the map server.

[0146] The map server and the map of the vehicle system communicate to navigate to the point of interest in the collection corresponding to the attribute information containing the nickname, i.e., the map server makes the map of the vehicle system navigate to the point of interest "xx District xx Road xx Sports Center xx Gate" in the collection corresponding to the "badminton court".

[0147] Thus, when the user next issues a navigation request by voice to navigate to the nickname, the voice server can perform intent recognition and slot recognition on the navigation request, and when the intent recognition result is a navigation intent, the matching result in which the point of interest in the slot recognition result and the attribute information containing the nickname are completely matched is sent to the map server, and the map server completes the process of navigating the map of the vehicle system to the point of interest in the collection.

[0148] In addition, step 027 includes: sending the matching result in which the point of interest in the slot recognition result and the attribute information containing the nickname are not completely matched to the map server, and the map server and the map of the vehicle system communicate to navigate to the point of interest selected by the user, and the point of interest selected by the user is determined by searching the result according to the point of interest in the slot recognition result.

[0149] The processor is configured to send the matching result in which the point of interest in the slot recognition result and the attribute information containing the nickname are not completely matched to the map server, and the map server and the map of the vehicle system communicate to navigate to the point of interest selected by the user, and the point of interest selected by the user is determined by searching the result according to the point of interest in the slot recognition result.

[0150] That is, when the matching result in which the point of interest in the slot recognition result and the attribute information containing the nickname are not completely matched is sent to the map server, the map server can search for one or more points of interest in the map collection according to the point of interest in the slot recognition result, and the one or more points of interest can be presented in the form of a list on the map of the vehicle system and selected by the user.

[0151] For example, the next navigation request initiated by the user is "navigate to the frequently-visited badminton court", and when the point of interest "frequently-visited badminton court" in the slot recognition result does not completely match the attribute information containing the nickname "badminton court", the map server can search for one or more sports center addresses that have been collected by the user in the collection corresponding to the map according to the point of interest "frequently-visited badminton court" in the slot recognition result, and the one or more sports center addresses are one or more points of interest in the collection corresponding to the map.

[0152] After the user selects one of the points of interest in the collection corresponding to the map, the map server and the map of the vehicle system communicate to navigate to the point of interest selected by the user in the collection corresponding to the map.

[0153] Thus, when the interest point in the slot recognition result in the next navigation request initiated by the user and the attribute information containing the nickname do not completely match, the map server can perform navigation search on the map end of the vehicle-mounted system according to the interest point in the slot recognition result to obtain one or more interest points in the map collection, and complete navigation according to the user's selection of an interest point in the map collection.

[0154] In addition, for a multi-purpose navigation request issued by the user, for example, the next navigation request is "first navigate to the badminton court and then go home", that is, when the navigation request of the user contains multiple interest points, the matching result of complete matching of at least one of the multiple interest points in the slot recognition result and the attribute information containing the nickname can be sent to the map server, so as to complete navigation to the interest point in the collection corresponding to the attribute information containing the nickname.

[0155] Specifically, step 027 comprises: sending the matching result of complete matching of at least one of the multiple interest points in the slot recognition result and the attribute information containing the nickname to the map server, and the map server and the map of the vehicle-mounted system communicate to complete navigation to the interest point in the collection corresponding to the attribute information containing the nickname.

[0156] The processor is configured to send the matching result of complete matching of at least one of the multiple interest points in the slot recognition result and the attribute information containing the nickname to the map server, and the map server and the map of the vehicle-mounted system communicate to complete navigation to the interest point in the collection corresponding to the attribute information containing the nickname.

[0157] For example, the next navigation request is "first navigate to the badminton court and then go home", and the slot recognition result includes two interest points, "badminton court" and "home".

[0158] When the interest point in the collection of the map of the vehicle-mounted system includes the attribute information "badminton court" containing the nickname, but does not include the attribute information "home", the matching result in the voice server can be: one of the two interest points in the slot recognition result "badminton court" and the attribute information containing the nickname "badminton court" completely match. Therefore, the map server and the map of the vehicle-mounted system communicate to complete navigation to the interest point in the collection corresponding to the attribute information containing the nickname "badminton court".

[0159] When the interest points in the collection of the map of the vehicle-mounted system include the attribute information "badminton hall" of the nickname and the attribute information "home" of the nickname, the matching result in the voice server can be that the two interest points of the slot recognition result are respectively "badminton hall" and "home", which are completely matched with the attribute information "badminton hall" and "home" of the nickname respectively. Therefore, the map communication between the map server and the vehicle-mounted system can be completed in combination with semantic understanding to navigate to the interest point in the collection corresponding to the attribute information "badminton hall" of the nickname first, and then navigate to the interest point in the collection corresponding to the attribute information "home" of the nickname.

[0160] Therefore, the present application can realize the navigation request with multiple interest points, and when at least one of the multiple interest points of the slot recognition result is completely matched with the attribute information of the nickname, the map communication between the map server and the vehicle-mounted system can be completed to navigate to the interest point in the collection corresponding to the attribute information of the nickname, and one or more interest points can be navigated to.

[0161] After determining that the interest point of the current navigation request does not include the nickname, a voice broadcast guiding the user to add the nickname to the interest point is generated and sent to the vehicle, so that the user can add the nickname to the interest point. The following describes how the user adds the nickname to the interest point.

[0162] Specifically, after step 022, the vehicle voice interaction method comprises: receiving the voice request of the user adding the nickname to the interest point forwarded by the vehicle; and

[0163] The voice request is recognized to obtain the nickname of the interest point added by the user.

[0164] The processor is configured to receive the voice request of the user adding the nickname to the interest point forwarded by the vehicle, and recognize the voice request to obtain the nickname of the interest point added by the user.

[0165] The voice server end first receives the voice request of the user adding the nickname to the interest point forwarded by the vehicle. For example, the voice request of the user adding the nickname "badminton hall" to the interest point "XXX district XXX road XXX sports center XXX door" can be "remark the XXX district XXX road XXX sports center XXX door as badminton hall". The voice request can also be other sentences indicating the nickname, which is not limited here.

[0166] Then, the voice server can recognize the voice request to obtain the nickname of the interest point added by the user. When the voice request is "remark the XXX district XXX road XXX sports center XXX door as badminton hall", the recognition of the voice request can be that the nickname of the interest point added by the user is "badminton hall".

[0167] In this way, the user can add a nickname to the point of interest through voice.

[0168] After the user adds a nickname to the point of interest through voice, the map server can add attribute information containing the nickname to the corresponding point of interest in the collection associated with the user account. How to add attribute information containing the nickname to the corresponding point of interest in the collection associated with the user account in the map server is described below.

[0169] Specifically, step 024 includes: when the user in the vehicle cabin adds a nickname to the point of interest, according to the recognized nickname of the point of interest added by the user, the map server adds attribute information containing the nickname to the corresponding point of interest in the collection associated with the user account.

[0170] The processor is configured to: when the user in the vehicle cabin adds a nickname to the point of interest, according to the recognized nickname of the point of interest added by the user, the map server adds attribute information containing the nickname to the corresponding point of interest in the collection associated with the user account.

[0171] For example, when the user in the vehicle cabin adds a nickname "badminton hall" to the point of interest "XXX District XXX Road XXX Sports Center XXX Gate", the voice server can add attribute information containing the nickname "badminton hall" to the corresponding point of interest "XXX District XXX Road XXX Sports Center XXX Gate" in the collection associated with the user account in the map server according to the recognized nickname "badminton hall" of the point of interest added by the user, so that the user can directly issue a voice request to navigate to the nickname next time to initiate navigation.

[0172] After step 024, the vehicle voice interaction method includes: issuing a voice broadcast guiding the user to initiate a navigation request to the nickname through voice next time according to the added attribute information containing the nickname, so that the user can initiate a navigation request to the nickname through voice next time.

[0173] The processor is configured to: issue a voice broadcast guiding the user to initiate a navigation request to the nickname through voice next time according to the added attribute information containing the nickname, so that the user can initiate a navigation request to the nickname through voice next time.

[0174] That is, the server can issue a voice broadcast guiding the user to initiate a navigation request to the nickname through voice next time according to the added attribute information containing the nickname, so that the user can initiate a navigation request to the nickname through voice next time.

[0175] For example, when the user adds the attribute information including the remark name "badminton hall" to the interest point "XXX District XXX Road XXX Sports Center XXX Gate", the voice broadcast content can be "OK, the XXX District XXX Road XXX Sports Center XXX Gate is remarked as the badminton hall, and next time you can directly say [navigate to the badminton hall] to achieve the navigation request.

[0176] Therefore, the present application can issue a voice broadcast to guide the user to initiate the navigation request to the remark name through voice next time according to the added attribute information including the remark name, so that the user can initiate the navigation request to the remark name through voice next time.

[0177] As can be seen from the foregoing, for the current navigation request without the remark name, the map of the vehicle end can search one or more routes according to the current navigation request, and a pop-up window of the route list can be popped up, and the user can also select a route through voice confirmation at this time, so as to determine the route confirmed by the user according to the voice issued by the user.

[0178] Specifically, before step 022, the vehicle voice interaction method comprises: receiving a voice request of the user for selecting the alternative route provided by the vehicle for the current navigation request; and

[0179] According to the voice request, the route confirmed by the user is determined.

[0180] The processor is configured to receive a voice request of the user for selecting the alternative route provided by the vehicle for the current navigation request; and determine the route confirmed by the user according to the voice request.

[0181] For example, when the route list includes route 1, route 2 and route 3 respectively, the pop-up window can be as shown in Figure 5 The voice request of the user for selecting the alternative route provided by the vehicle for the current navigation request can be "select route 2 for navigation", and the route confirmed by the user according to the voice request is route 2.

[0182] Therefore, the user does not need to manually click the center control screen of the vehicle-mounted system, but can select the route through voice, so that the user can interact with the vehicle end through voice during driving, which is safer.

[0183] In addition, after the user understands the remark function of the navigation destination, the user can realize the collection and remark of the interest point through the voice request of actively setting the remark name of the interest point through voice, so that the user can directly say "navigate to the remark name" when using the function of navigating to the remark name next time, without saying "navigate to the corresponding interest point" as in the traditional voice navigation mode.

[0184] Specifically, please refer to Figure 6 The present application also provides a vehicle voice interaction method. The vehicle voice interaction method comprises:

[0185] 031: receiving the current voice request forwarded by the vehicle, wherein the current voice request is a voice request for setting a nickname for a point of interest in a map of a vehicle-mounted system by a user in a vehicle cabin through voice;

[0186] 032: setting the nickname for the point of interest after completing the collection of the point of interest according to the current voice request and synchronizing to a map server;

[0187] 033: receiving a next navigation request initiated by the user through voice for navigation to the nickname forwarded by the vehicle;

[0188] 034: performing intent recognition and slot recognition on the next navigation request, and confirming that the intent recognition result is a navigation intent, and then comparing whether the point of interest of the slot recognition result is completely matched with the nickname;

[0189] 035: sending the matching result to the map server to complete the navigation in response to the next navigation request.

[0190] The processor is configured to receive the current voice request forwarded by the vehicle, wherein the current voice request is a voice request for setting a nickname for a point of interest in a map of a vehicle-mounted system by a user in a vehicle cabin through voice; set the nickname for the point of interest after completing the collection of the point of interest according to the current voice request and synchronizing to a map server; receive a next navigation request initiated by the user through voice for navigation to the nickname forwarded by the vehicle; perform intent recognition and slot recognition on the next navigation request, and confirm that the intent recognition result is a navigation intent, and then compare whether the point of interest of the slot recognition result is completely matched with the nickname; and send the matching result to the map server to complete the navigation in response to the next navigation request.

[0191] First, the voice server end receives the current voice request forwarded by the vehicle. Wherein the current voice request is a voice request for setting a nickname for a point of interest in a map of a vehicle-mounted system by a user in a vehicle cabin through voice. For example, the current voice request issued by the user is “remark my son's school position”, wherein my son's school position is a point of interest in a map of a vehicle-mounted system by the user.

[0192] Next, the voice server, based on the voice request, saves the point of interest (POI), sets a nickname for it, and synchronizes it to the map server. For example, if the voice request is "Note the location of my son's school," and the POI is "the location of my son's school," the voice server can save the address of the corresponding POI "the location of my son's school" from the in-vehicle map, set a nickname for the saved POI, and synchronize it to the map server. For example, this nickname and POI can be synchronously stored on the map server corresponding to the Gaode Map application, so that when the user sends a navigation request to that nickname again, the navigation request can be successfully responded to and completed.

[0193] Then, the voice server can receive the user's next navigation request, initiated via voice, to navigate to the nickname. It then performs intent recognition and slot recognition on the next navigation request. After confirming that the intent recognition result is a navigation intent, it compares the point of interest in the slot recognition result with the nickname to see if they completely match. The matching result is then sent to the map server to respond to the next navigation request and complete the navigation. For example, the user's next navigation request could be "Navigate to my son's school."

[0194] Thus, the present invention can also, after the user knows that there is a note function, receive the user's voice request to actively set a note name for the point of interest, and then collect and note the point of interest. This allows the user to directly say "Navigation to note name" to navigate to the point of interest the next time they use the navigation to go to the note name function.

[0195] Prior to step 031, the vehicle voice interaction method includes: configuring a voice template for users to set nicknames for points of interest, the voice template being used to match and recognize voice requests from users in the vehicle cabin to set nicknames for points of interest in the vehicle system's map.

[0196] The processor is used to configure voice templates for users to set nicknames for points of interest. The voice templates are used to match and recognize voice requests from users in the vehicle cabin to set nicknames for points of interest in the in-vehicle system's map.

[0197] Understandably, before a user actively makes a voice request to add a name to a point of interest, a voice template for the user to set a name for the point of interest needs to be configured on the server side. This allows the voice template to be matched and recognized so that the user in the vehicle cabin can set a name for the point of interest in the in-vehicle system map via voice.

[0198] The voice template includes "[Put][Take][This|This one](destination|end point|place to go|navigation end point|navigation destination)(Set|Configure|Remark|Mark)(as|into|call)#Remark name#". For example, for this voice template, the voice requests that the user can issue can be "Set the end point to my uncle's house" or "Remark this address as the old place".

[0199] The voice template also includes "[Car name][You][You tell me](Remember|Write down|Make a note of|Remark|Mark)[This|This one][You][You tell me](destination|end point|place to go|navigation end point|navigation destination)(is called|is named|has the remark name of|has the alias of|name is|remark as|mark as)#Remark name#". For example, for this voice template, the voice requests that the user can issue can be "Xiaop, make a note that this place is my gym" or "You remember that this destination is my second uncle's house".

[0200] The voice template also includes "[Put][Take]#POI(Set|Configure|Remark|Mark)(as|into)[My|My]#Remark name#". For example, for this voice template, the voice requests that the user can issue can be "Set No. 28 Nanjing Road as my son's school".

[0201] The voice template also includes "(Set|Remark|Mark)#Remark name#[of](address|location|positioning)". For example, for this voice template, the voice requests that the user can issue can be "Set the location of my son's school".

[0202] The voice template also includes "(Set|Remark|Mark)[This|This one](place|address|location|positioning|here)(as|into|call)#Remark name#". For example, for this voice template, the voice requests that the user can issue can be "Set this place as my son's school".

[0203] In this way, by pre-configuring the voice template for the user to set the remark name for the point of interest on the server side, when the user issues the corresponding voice request according to this voice template, the server can understand according to the template that the user hopes to make a remark for the point of interest and the corresponding remark name.

[0204] It can be understood that when the server identifies through this voice request and sets the remark name for the point of interest, it is necessary to first judge whether the map of the in-vehicle system is bound to the user account. If the map of the in-vehicle system is not bound to the user account, the vehicle side can issue a voice reminder to the user that the map of the in-vehicle system needs to be bound to the user account first before the remark name can be set, and the map account binding page can be opened at the same time. For example, the voice to remind the user can be "You need to bind the Amap account first before you can set the remark name".

[0205] Specifically, the step 032 comprises: in the case that the map of the vehicle-mounted system is bound with the user account, performing slot recognition on the current voice request to obtain a recognition result; and

[0206] After completing the collection of the point of interest according to the recognition result, a nickname is set for the point of interest and is synchronized to the map server.

[0207] The processor is configured to, in the case that the map of the vehicle-mounted system is bound with the user account, perform slot recognition on the current voice request to obtain a recognition result; and after completing the collection of the point of interest according to the recognition result, set a nickname for the point of interest and synchronize the nickname to the map server.

[0208] The user account comprises identity information of the user, so that whether the current voice request is a voice request issued by the user can be identified through the user account.

[0209] It can be understood that the slot recognition on the current voice request can be performed only after it is determined that the voice request is issued by the user corresponding to the user account, so that the point of interest corresponding to the current voice request of the user can be found in the user account of the map of the corresponding vehicle-mounted system, and the collection of the point of interest is completed according to the recognition result, and a nickname is set for the point of interest and is synchronized to the map server.

[0210] In this way, it can be ensured that the current voice request issued by the user can find the corresponding point of interest in the user account of the map of the corresponding vehicle-mounted system, so that the nickname can be set for the point of interest and is synchronized to the map server.

[0211] More specifically, the application can further determine whether the current voice request issued by the user has a point of interest slot, and when the current voice request issued by the user has a point of interest slot, the point of interest searched according to the point of interest slot can be found, and then the nickname can be set for the point of interest.

[0212] In detail, the step 0322 comprises: if the slot recognition result is that the current voice request comprises a point of interest slot and a nickname, the point of interest searched according to the point of interest slot is found, and a voice broadcast is issued to the vehicle to inquire about the expected point of interest of the user; and

[0213] After completing the collection of the point of interest according to the point of interest confirmed by the user, a nickname is set for the point of interest and is synchronized to the map server.

[0214] The processor is configured to, if the slot recognition result is that the current voice request comprises a point of interest slot and a nickname, find the point of interest searched according to the point of interest slot, issue a voice broadcast to the vehicle to inquire about the expected point of interest of the user, and after completing the collection of the point of interest according to the point of interest confirmed by the user, set a nickname for the point of interest and synchronize the nickname to the map server.

[0215] For example, the current voice request issued by the user is "set No. 28, Nanjing Road as son's school", the slot recognition result of the server for the voice request is that the current voice request includes the interest point slot "No. 28, Nanjing Road" and the nickname "son's school", and the navigation search according to the interest point slot "No. 28, Nanjing Road" can obtain one or more interest points to form an interest point list, for example, the search obtains interest points including "1, No. 28, Nanjing Road north gate", "2, No. 28, Nanjing Road south gate", and "3, No. 28, Nanjing Road east gate".

[0216] At this time, the server can issue a voice broadcast to the vehicle to ask the user about the expected interest point, for example, the content of the voice broadcast is "which one is it". If the user replies that it is the second one, the server sets the nickname for the interest point after completing the collection of the interest point according to the interest point confirmed by the user, and synchronizes to the map server. At this time, the server can also reply to the user that the nickname has been set through the voice broadcast again, and at this time, the content of the voice broadcast is "OK, No. 28, Nanjing Road south gate has been set as son's school, next time you can directly say [navigate to son's school]".

[0217] In this way, the application can also judge whether the current voice request issued by the user has an interest point slot, when the current voice request issued by the user has an interest point slot, the interest point obtained by navigation search according to the interest point slot can be set with a nickname.

[0218] Step 0322 includes: if the slot recognition result is that the current voice request includes a destination slot and a nickname, then according to the destination slot and the nickname, the collection of the interest point is completed, the nickname is set for the interest point, and the map server is synchronized.

[0219] The processor is configured to, if the slot recognition result is that the current voice request includes a destination slot and a nickname, then according to the destination slot and the nickname, the collection of the interest point is completed, the nickname is set for the interest point, and the map server is synchronized.

[0220] The user can set a nickname for the interest point of the destination after initiating voice navigation. That is, the user can perform navigation search through voice, select a destination, and set a nickname for the destination through voice request.

[0221] For example, the user can search for "No. 29, Nanjing Road" through voice navigation, the server identifies the corresponding point of interest, and then performs navigation search to obtain "No. 1, No. 29, Nanjing Road", "No. 2, No. 29, Nanjing Road", and "No. 3, No. 29, Nanjing Road", and asks the user which destination the user wants to navigate to. After the user confirms that the destination is "No. 1, No. 29, Nanjing Road", the server provides route selection for the user, and initiates navigation after the user selects a route. At this time, if the current voice request issued by the user is "set the end point as Uncle Third's house", the slot recognition result of the server performing slot recognition on the voice request is that the current voice request includes a destination slot "end point" and a nickname "Uncle Third's house". According to the destination slot "end point" and the nickname "Uncle Third's house", the "end point" is collected in the point of interest "No. 1, No. 29, Nanjing Road" in the map, and the nickname "Uncle Third's house" is set for the point of interest. The point of interest and the nickname are synchronized to the map server.

[0222] In addition, after the collection of the point of interest is completed, the nickname is set for the point of interest and is synchronized to the map server. At this time, the server can also issue a voice broadcast to reply to the user that the nickname has been set. For the current voice request "set the end point as Uncle Third's house" issued by the user, the content of the voice broadcast at this time is "OK, the end point has been set as Uncle Third's house. Next time, you can directly say [navigate to Uncle Third's house]".

[0223] In this way, the application can also determine whether the current voice request issued by the user has a destination slot. When the current voice request issued by the user has a destination slot, the collection of the point of interest corresponding to the destination can be completed according to the destination slot and the nickname, and then the nickname can be set for the point of interest.

[0224] Step 0322 includes: if the slot recognition result is that the current voice request includes an end point slot and a nickname, issuing a voice broadcast to the vehicle according to the end point slot to ask the user about the desired point of interest; and

[0225] After the collection of the point of interest confirmed by the user is completed, the nickname is set for the point of interest and is synchronized to the map server.

[0226] The processor is configured to: if the slot recognition result is that the current voice request includes an end point slot and a nickname, issue a voice broadcast to the vehicle according to the point of interest slot to ask the user about the desired point of interest; and after the collection of the point of interest confirmed by the user is completed, set the nickname for the point of interest and synchronize the point of interest and the nickname to the map server.

[0227] It can be understood that the user can make a voice request to want to initiate navigation and set a nickname for the end point of the navigation, but the user does not select a destination, the server can issue a voice broadcast to the vehicle to ask the user about the desired point of interest, and after completing the collection of the point of interest confirmed by the user's reply, set a nickname for the point of interest and synchronize to the map server. For example, after the user inputs "navigate to Nanjing Road" by voice, without selecting a specific address of Nanjing Road as the destination, the user's voice request for this time is "set the end point as Uncle Third's house", the slot recognition result of the server for this voice request is that this voice request includes an end point slot and a nickname "Uncle Third's house", and under the condition that the "Nanjing Road" corresponding to the end point slot is not a point of interest with a specific address, the server issues a voice broadcast to the vehicle to ask the user about the desired point of interest, for example, the content of the voice broadcast is "OK, please say the specific address". At this time, the user can answer the specific address of the point of interest corresponding to the end point, for example, "29, Nanjing Road", the server can obtain one or more points of interest according to "29, Nanjing Road" under the condition of navigation search, and can pop up a pop-up window of the point of interest list in the center screen of the vehicle, and then issue a voice broadcast to the vehicle to ask the user about the desired point of interest.

[0228] For example, when the points of interest obtained by navigation search have three, the point of interest list includes "1, 29, Nanjing Road, Building 1", "2, 29, Nanjing Road, Building 2", and "3, 29, Nanjing Road, Building 3", and the server issues a voice broadcast to the vehicle to ask the user about the desired point of interest, which is "please tell me which one".

[0229] If the user's reply is the second one, the server can set a nickname for the point of interest after completing the collection of the point of interest confirmed by the user's reply, and synchronize to the map server, at this time, the server can also reply to the user that the nickname has been set by issuing a voice broadcast again, at this time, the content of the voice broadcast is "OK, 29, Nanjing Road, Building 2 has been set as Uncle Third's house, next time you can directly say [navigate to Uncle Third's house]".

[0230] In this way, the application can also determine whether the user's voice request for this time has an end point slot, when the user's voice request for this time has an end point slot, the confirmed point of interest is obtained by asking according to the end point slot, and the user's confirmed point of interest is collected and a nickname is set for the point of interest.

[0231] In addition, when the voice request for this time does not include a point of interest slot, a destination slot or an end point slot, the server needs to issue a voice broadcast to the vehicle to prompt the user to input the point of interest by voice, so as to complete the collection of the point of interest and the nickname of the point of interest according to the point of interest input by the user by voice.

[0232] That is, step 0322 comprises: if the slot identification result is that the current voice request does not include the interest point slot, the destination slot or the terminal point slot, issuing a voice broadcast to the vehicle prompting the user to input the interest point by voice; and

[0233] After the user completes the collection of the interest point by inputting the interest point by voice, a nickname is set for the interest point and is synchronized to the map server.

[0234] The processor is configured to, if the slot identification result is that the current voice request does not include the interest point slot, the destination slot or the terminal point slot, issue a voice broadcast to the vehicle prompting the user to input the interest point by voice; and after the user completes the collection of the interest point by inputting the interest point by voice, set a nickname for the interest point and synchronize it to the map server.

[0235] It can be understood that, before the current voice request, the user has not initiated a navigation request by voice, and the current voice request does not include the destination slot or the terminal point slot. If the current voice request also does not include the interest point slot, it cannot be determined which interest point the user wants to nickname. At this time, a voice broadcast needs to be issued to the vehicle to prompt the user to input the interest point by voice, so as to assist the user to complete the collection and nickname of the interest point through voice interaction.

[0236] For example, the current voice request issued by the user is “nickname as the third uncle's house”. The slot identification result of the server performing slot identification on the voice request is that the voice request does not include the interest point slot, the destination slot or the terminal point slot. At this time, the server can issue a voice broadcast to the vehicle prompting the user to input the interest point by voice. The prompt content of the voice broadcast can be “which address do you want to nickname as the third uncle's house”. If the user replies by voice “No. 30, Nanjing Road”, the interest point “No. 30, Nanjing Road” on the map of the vehicle-mounted system can be collected, and the interest point “No. 30, Nanjing Road” can be nicknamed as “the third uncle's house”.

[0237] In this way, when the current voice request issued by the user does not include the interest point slot, the destination slot or the terminal point slot, the server can issue a voice broadcast to the vehicle prompting the user to input the interest point by voice, thereby guiding the user to input the interest point by voice, and completing the collection and nickname of the interest point.

[0238] More specifically, step 0322 comprises: if a plurality of interest points are obtained by navigation search according to the interest point input by the user by voice, issuing a voice broadcast to the vehicle to ask the user about the expected interest point; and

[0239] After the user replies to confirm the interest point, the collection of the interest point is completed, and a nickname is set for the interest point and is synchronized to the map server.

[0240] The processor is configured to: if a plurality of interest points are obtained by performing navigation search according to the interest point input by the user through voice, issue a voice broadcast to the vehicle to inquire the interest point expected by the user; and set a nickname for the interest point after the collection of the interest point is completed according to the interest point confirmed by the user, and synchronize the nickname to the map server.

[0241] For example, the current voice request input by the user is "set a nickname as the house of the third uncle", the slot recognition result of the server performing slot recognition on the voice request is that the current voice request does not include an interest point slot, a destination slot or a terminal point slot. At this time, the server can issue a voice broadcast to the vehicle to prompt the user to input an interest point through voice, and the prompt content of the voice broadcast can be "which address do you want to set as the house of the third uncle". If the user replies through voice "No. 30, Nanjing Road", the "No. 30, Nanjing Road" can also be searched, and in the case that there are a plurality of interest points in the navigation search result of "No. 30, Nanjing Road", the user is inquired about the interest point to be set as a nickname, and then the nickname is set for the interest point confirmed by the user. For example, if there are 3 interest points obtained by navigation search, the interest point list includes "1, No. 30, Nanjing Road, A Building", "2, No. 30, Nanjing Road, B Building" and "3, No. 30, Nanjing Road, C Building", the server issues a voice broadcast to the vehicle to inquire the interest point expected by the user, and the voice broadcast is "which one do you want?". If the user replies that it is the second one, the server completes the collection of the interest point according to the interest point confirmed by the user, sets a nickname for the interest point and synchronizes the nickname to the map server, and the server can also reply to the user that the nickname has been set by issuing a voice broadcast again, and the content of the voice broadcast is "OK, No. 30, Nanjing Road, B Building has been set as the house of the third uncle, next time you can directly say [navigate to the house of the third uncle]".

[0242] In this way, in the case that the current voice request does not include an interest point slot, a destination slot or a terminal point slot, the server issues a voice broadcast to prompt the user to input an interest point through voice, and in the case that a plurality of interest points are obtained by performing navigation search according to the interest point input by the user through voice, the server can trigger the inquiry of the interest point expected by the user again, so as to determine a unique interest point, collect the interest point and set a nickname for the interest point.

[0243] The above description provides a plurality of scenarios of setting a nickname for an interest point and corresponding processing manners, in addition to which, the active setting of a nickname for an interest point by a user is completed by establishing a mapping relationship between the interest point and the nickname, which will be described in detail below.

[0244] Specifically, referring to Figure 7 The application provides a vehicle voice interaction method. The vehicle voice interaction method comprises the following steps:

[0245] 041: receive a voice request of setting a remark of a point of interest in a navigation map of a vehicle-mounted system forwarded by a vehicle;

[0246] 043: recognize the voice request to obtain a remark name;

[0247] 045: establish a mapping relationship between the remark name and the point of interest;

[0248] 047: according to the mapping relationship, determine a navigation request of navigating to the remark name after the vehicle-mounted system is woken up by a voice of the user next time as a navigation request of navigating to the corresponding point of interest by the user.

[0249] The processor is configured to receive a voice request of setting a remark of a point of interest in a navigation map of a vehicle-mounted system forwarded by a vehicle, recognize the voice request to obtain a remark name, establish a mapping relationship between the remark name and the point of interest, and according to the mapping relationship, determine a navigation request of navigating to the remark name after the vehicle-mounted system is woken up by a voice of the user next time as a navigation request of navigating to the corresponding point of interest by the user.

[0250] The voice request of setting a remark of a point of interest in a navigation map of a vehicle-mounted system forwarded by a vehicle is a voice request with a remark name, for example, the voice request is "set point of interest 1 as uncle's house", wherein the remark name set for point of interest 1 in the navigation map of the vehicle-mounted system is "uncle's house", and the voice request is "point of interest 2 is my favorite Japanese restaurant", wherein the remark name set for point of interest 2 in the navigation map of the vehicle-mounted system is "my favorite Japanese restaurant".

[0251] The server end first receives a voice request of setting a remark of a point of interest in a navigation map of a vehicle-mounted system forwarded by a vehicle, and then recognizes the voice request to obtain a remark name. The method of recognizing the voice request can be a natural language understanding (NLU) method to obtain the remark name.

[0252] Then, the server end can establish a mapping relationship between the remark name and the point of interest and store it. For example, the remark name "uncle's house" is mapped to point of interest 1 in the navigation map, and the remark name "my favorite store" is mapped to point of interest 2 in the navigation map.

[0253] Finally, the server end can determine a navigation request of navigating to the remark name after the vehicle-mounted system is woken up by a voice of the user next time as a navigation request of navigating to the corresponding point of interest by the user according to the mapping relationship, so that the voice request with the remark name by the user next time can be correctly understood, and directly navigate to the corresponding point of interest, improving the user experience.

[0254] For example, the next time the user voice wakes up the vehicle system, the navigation request to the note name is "navigate to the uncle's house", and the navigation request to the note name "uncle's house" corresponding to the interest point 1 can be determined directly according to the navigation request, so as to directly initiate navigation to the interest point 1, and realize correct understanding of the intention of the voice request with the note of the user.

[0255] Therefore, the present application can obtain the note name by recognizing the voice request with the note, and determine the navigation request to the corresponding interest point according to the mapping relationship between the note name and the interest point, so as to realize correct understanding of the intention of the voice request with the note.

[0256] More specifically, please refer to Figure 8 , step 041 includes:

[0257] 0411: receiving the current navigation request of the user forwarded by the vehicle;

[0258] 0412: feeding back the navigation search result containing multiple interest points in the navigation map of the vehicle system to the user for confirmation according to the current navigation request;

[0259] 0413: receiving the voice request of the user setting the note for the confirmed interest point forwarded by the vehicle.

[0260] The processor is configured to receive the current navigation request of the user forwarded by the vehicle; feed back the navigation search result containing multiple interest points in the navigation map of the vehicle system to the user for confirmation according to the current navigation request; and receive the voice request of the user setting the note for the confirmed interest point forwarded by the vehicle.

[0261] The current navigation request of the user forwarded by the vehicle, for example, the navigation request is "go to the Furongli community", that is, the vehicle forwards the navigation request "go to the Furongli community" to the server end.

[0262] Then, the server end feeds back the navigation search result containing multiple interest points in the navigation map of the vehicle system to the user for confirmation according to the current navigation request. For example, according to the navigation request "go to the Furongli community", 3 interest points are obtained by navigation search in the navigation map of the vehicle system, for example, the multiple interest points are interest point 1, interest point 2 and interest point 3, and the multiple interest points are fed back to the user as the navigation search result.

[0263] If the user selects the confirmed interest point as interest point 1, it is determined that the destination of the user using the navigation is interest point 1.

[0264] Then, if the server receives a voice request for setting a note for a confirmed point of interest by the user forwarded by the vehicle, for example, "set as uncle's house", the server can set a note for the confirmed point of interest 1 as "uncle's house" for the user. That is to say, in the case that the user has set the point of interest 1 as a navigation destination through a voice request before the server receives a voice request for setting a note for the point of interest, the voice request for setting a note "uncle's house" for the point of interest this time can directly inherit the point of interest 1 confirmed by voice during navigation as the point of interest to be noted, thereby establishing a mapping relationship between the point of interest 1 and "uncle's house", and at this time, the content of the voice request for setting a note "uncle's house" for the point of interest this time can not include the point of interest 1.

[0265] In this way, the server can receive a voice request for setting a note for a point of interest in a navigation map of a vehicle-mounted system by the user forwarded by the vehicle, thereby determining the point of interest in the navigation map of the vehicle-mounted system corresponding to the voice request related to the note.

[0266] For the case that the user does not set a navigation destination through voice, after step 041, the vehicle voice interaction includes: slot recognition on the voice request; and

[0267] In the case that the result of slot recognition does not include point of interest slot information, if it is determined by a navigation application program interface of the vehicle that the voice request has a destination field, a note is set for the point of interest corresponding to the destination field.

[0268] The processor is configured to perform slot recognition on the voice request, and in the case that the result of slot recognition does not include point of interest slot information, if it is determined by a navigation application program interface of the vehicle that the voice request has a destination field, a note is set for the point of interest corresponding to the destination field.

[0269] In the slot recognition, the slot recognition can be performed by a pre-trained slot extraction model, and slot information is obtained.

[0270] It should be noted that, in order to meet the demand for setting a note for a longer address in a voice request of the user, first, a sentence pattern of a voice request corresponding to an intention of setting a note can be formulated, for example, the sentence pattern of the voice request corresponding to the intention of setting a note includes "note as XXX", "set as XXX", and "remember XXX as XXX", and a large number of voice requests containing note names are collected to train a slot extraction model, so that the voice requests containing note names can correctly recognize the intention of setting a note.

[0271] After identifying that the user voice request has the intention of setting a nickname, a large number of voice requests containing nicknames can be used for continuous training of the slot extraction model, such as voice requests including "set address A as my uncle's house", "set address B as my child's school", and "set address C as my favorite store". When the slot extraction model performs slot recognition, it can accurately extract "uncle's house", "child's school", and "favorite store" in the voice request and identify the corresponding slot name as "cust_name", and the slot value corresponding to the nickname as "uncle's house", "child's school", and "favorite store", that is, the corresponding nickname slot information can be accurately identified. At the same time, when the slot extraction model performs slot recognition, it can accurately extract "address A", "address B", and "address C" in the voice request and identify the corresponding slot name as "poi_name", and the slot value corresponding to the point of interest as "address A", "address B", and "address C", that is, the corresponding point of interest slot information can be accurately identified.

[0272] The slot recognition result can include point of interest slot information or not.

[0273] It can be understood that after the server receives the voice request forwarded by the vehicle, the server can first perform slot recognition on the voice request. For example, after the server receives the voice request "go to uncle's house" forwarded by the vehicle, slot recognition is performed to obtain a slot name of the slot recognition result as "cust_name", that is, a nickname slot, and a slot value as "uncle's house". At this time, the slot recognition result does not include point of interest slot information. In this case, the server can call the navigation application program interface of the vehicle to determine whether the voice request contains a destination field to determine which point of interest the user wants to set a nickname for. If it is determined that the voice request contains a destination field by using the navigation application program interface of the vehicle, the point of interest corresponding to the destination field is set a nickname.

[0274] It can be understood that by calling the navigation application program interface of the vehicle, it can be determined whether the current voice request interacted by the user and the vehicle contains a destination field, that is, a field indicating going to a certain destination, so as to determine the point of interest corresponding to the destination field of the current voice request of the user and set a nickname for the point of interest.

[0275] For example, the voice request of the user is "set the end point to the company", the result of slot identification of the voice request is that the slot name is "cust_name" and the slot value is "company", and the interest point slot information is not included in the identification result. At this time, the navigation application interface of the vehicle can be called to determine whether the current voice request contains a destination field, and the result is that the voice request contains a destination field, and the destination field is "Zhongguancun No. 2". Then, the interest point corresponding to the destination field "Zhongguancun No. 2" can be set with a note, so that the next time the user inputs the voice request "start to Zhongguancun No. 2", the interest point corresponding to the destination field "Zhongguancun No. 2" can be directly navigated to.

[0276] Therefore, by calling the application interface of the map application on the vehicle-mounted system of the vehicle, it can be determined whether the voice request of the user interacting with the vehicle contains a destination field, so as to determine the interest point corresponding to the destination field and set a note for the interest point, so that the next time the voice request of the user containing the same destination field is input, the interest point corresponding to the destination field can be directly navigated to.

[0277] It can be understood that in the case where the navigation application interface of the vehicle is used to determine that the voice request contains a destination field, there can be multiple destination fields. At this time, the user needs to be actively asked and guided to confirm a unique destination field, and the interest point corresponding to the confirmed unique destination field is set with a note.

[0278] Specifically, step 0422 includes: if the navigation application interface of the vehicle is used to determine that the voice request contains multiple destination fields, the user is asked to query the desired destination field through voice broadcast; and

[0279] The interest point corresponding to the destination field confirmed by the user is set with a note.

[0280] The processor is configured to ask the user to query the desired destination field through voice broadcast if the navigation application interface of the vehicle is used to determine that the voice request contains multiple destination fields, and to set the interest point corresponding to the destination field confirmed by the user with a note.

[0281] For example, the voice request of the user is "set the destination as the company", and the navigation application interface of the vehicle is called to determine that the destination field in the voice request of the user interacting with the vehicle includes address 1: Zhongguancun No. 1, address 2: Zhongguancun No. 2, and address 3: Zhongguancun No. 3. At this time, the three destination field search lists above can be sequentially displayed on the center screen of the vehicle in the form of a pop-up window, and the vehicle can voice broadcast to ask the user for the desired destination field, for example, the voice content of the inquiry is "which one do you want". If the user replies that the second destination field is confirmed, it means that the destination field is "Zhongguancun No. 2", and the point of interest corresponding to the destination field "Zhongguancun No. 2" is noted as a company.

[0282] In this way, in the case that there are multiple destination fields in the voice request, a unique destination field can be determined.

[0283] It can be understood that when the navigation application interface of the vehicle is used to determine that there is a unique destination field in the voice request, the point of interest corresponding to the unique destination field can be directly set as a note.

[0284] In addition, step 0422 further includes setting a note for the point of interest corresponding to the destination field if it is determined by the navigation application interface of the vehicle that there is one destination field in the voice request.

[0285] The processor is configured to set a note for the point of interest corresponding to the destination field if it is determined by the navigation application interface of the vehicle that there is one destination field in the voice request.

[0286] For example, the voice request of the user is "set the destination as the company", and the navigation application interface of the vehicle is called to determine that there is one destination field in the current voice request, and the destination field is "Zhongguancun No. 3". The point of interest corresponding to the destination field "Zhongguancun No. 3" can be directly noted as a company, so that the next time the user inputs the voice request "go to the company", the point of interest corresponding to the destination field "Zhongguancun No. 3" can be directly navigated to.

[0287] In addition, in the case that the navigation application interface of the vehicle is used to determine that the destination field of the voice request does not exist, the application can feed back to the user that the point of interest is not found.

[0288] Specifically, the vehicle voice interaction method comprises: if it is determined by the navigation application interface of the vehicle that the destination field of the voice request does not exist, the point of interest is not found by voice broadcast.

[0289] The processor is configured to feed back that the point of interest is not found by voice broadcast if it is determined by the navigation application interface of the vehicle that the destination field of the voice request does not exist.

[0290] It can be understood that when the user's voice request does not plan a destination, the navigation application interface of the vehicle is used to determine the destination field of the voice request, and then it is determined that there is no corresponding destination field result.

[0291] At this time, the vehicle-mounted system needs to make a voice broadcast prompt to guide the user to change the way of speaking. For example, the feedback content of the voice broadcast is "I can't find the place you said. Try another way of speaking."

[0292] Therefore, when the navigation application interface of the vehicle is used to determine that there is no result of the destination field of the voice request, the voice broadcast feedback that the point of interest is not found can be used to timely feedback to the user and timely adjust the way of speaking until the destination field is determined and the point of interest corresponding to the destination field is determined.

[0293] In addition, the foregoing describes that in the case that the result of the slot recognition does not include the point of interest slot information, the navigation application interface of the vehicle can be called to determine whether the voice request contains the destination field, so as to determine the point of interest corresponding to the destination field and set a note for the point of interest.

[0294] The following describes how to set a note for the point of interest according to the point of interest slot information in the case that the result of the slot recognition includes the point of interest slot information. That is, for the voice request that includes the note name slot information and the point of interest slot information, the point of interest corresponding to the point of interest slot information is searched by using the navigation application interface of the vehicle, and a note is set for the point of interest.

[0295] Specifically, after step 041, the vehicle voice interaction method includes: performing slot recognition on the voice request; and

[0296] In the case that the result of the slot recognition includes the point of interest slot information, a note is set for the point of interest searched by using the navigation application interface of the point of interest slot information.

[0297] The processor is configured to perform slot recognition on the voice request, and in the case that the result of the slot recognition includes the point of interest slot information, set a note for the point of interest searched by using the navigation application interface of the point of interest slot information.

[0298] At this time, the voice request is a voice type that includes the note name slot information and the point of interest slot information, for example, the voice request is "set Zhongguancun as the place where dad works", wherein the result of the slot recognition includes the slot name "poi_name", the point of interest corresponding slot value is "Zhongguancun", the slot name is "cust_name", and the note name corresponding slot value is "the place where dad works", that is, the point of interest slot information and the note name slot information are recognized. Therefore, the point of interest searched by using the navigation application interface of the point of interest slot information can be obtained, and a note is set for the point of interest. Specifically, the point of interest is searched by using the navigation application interface of the point of interest corresponding slot value "Zhongguancun" in the slot information, and then a note is set for the corresponding point of interest.

[0299] Thus, in the case that the voice request includes the nickname information and the point of interest information, the point of interest is directly obtained by performing the navigation search on the point of interest slot information in the slot recognition result, and the point of interest is set with the nickname.

[0300] Step 0424 includes: if the point of interest slot information navigation search obtains multiple points of interest, the user is asked to expect the point of interest through voice broadcast; and

[0301] The point of interest confirmed by the user is set with the nickname.

[0302] The processor is configured to: if the point of interest slot information navigation search obtains multiple points of interest, the user is asked to expect the point of interest through voice broadcast; and the point of interest confirmed by the user is set with the nickname.

[0303] If the point of interest slot information navigation search obtains multiple points of interest, the user can also be asked to expect the point of interest through voice broadcast, and the point of interest confirmed by the user is set with the nickname.

[0304] For example, the voice request of the user is "set Zhongguancun as the place where dad works", since the address range of Zhongguancun indicated in the navigation map is large, the point of interest slot information navigation search obtains multiple points of interest, which are Zhongguancun No. 4, Zhongguancun No. 5 and Zhongguancun No. 6, respectively. At this time, the above three point of interest search lists can be sequentially displayed on the pop-up window of the vehicle's center control screen, and the vehicle can also ask the user to expect the point of interest through voice broadcast, for example, the voice content of the inquiry is "please tell me the number". If the user confirms the point of interest as the second one, i.e. the point of interest is "Zhongguancun No. 5", the point of interest "Zhongguancun No. 5" can be set as the company.

[0305] Thus, in the case that the point of interest slot information navigation search obtains multiple points of interest, the user is asked to expect the point of interest through voice broadcast, and the point of interest confirmed by the user is set with the nickname.

[0306] In addition, step 0424 includes: if the point of interest slot information navigation search obtains one point of interest, the point of interest is set with the nickname.

[0307] The processor is configured to: if the point of interest slot information navigation search obtains one point of interest, the point of interest is set with the nickname.

[0308] It can be understood that if the point of interest slot information navigation search obtains only one point of interest, the point of interest can be directly set with the nickname.

[0309] For example, the voice request of the user is "set Zhongguancun No. 7 as the place where dad works", and according to the navigation search of the point of interest slot information, only one point of interest "Zhongguancun No. 7" is obtained, so the point of interest "Zhongguancun No. 7" can be directly noted as a company.

[0310] In this way, in the case where only one point of interest is obtained by the navigation search of the point of interest slot information, the point of interest is directly set with a note.

[0311] In addition, step 0424 further includes: if the navigation search of the point of interest slot information does not exist a search result, the point of interest is not found through voice broadcast feedback.

[0312] The processor is configured to, if the navigation search of the point of interest slot information does not exist a search result, the point of interest is not found through voice broadcast feedback.

[0313] It can be understood that when the voice request of the user explicitly indicates the point of interest, the navigation application interface of the vehicle is used to determine the point of interest of the voice request, and it is determined that there is no corresponding search result.

[0314] At this time, the vehicle-mounted system needs to perform voice broadcast prompting to guide the user to change the way of saying. For example, the feedback content of the voice broadcast is "no location is found as you said, change the way of saying".

[0315] In this way, in the case where the navigation search of the point of interest slot information does not exist a search result, the point of interest is not found through voice broadcast feedback, which can timely feedback to the user and timely adjust the way of saying until the point of interest is found.

[0316] In the case where the function of navigating to the note name of the point of interest has been learned and the point of interest has been collected and noted, how the server specifically performs intent and slot recognition according to the voice request of the user for navigating to the note name of the point of interest and issues natural language understanding to the vehicle to complete navigation will be described in detail in the following description.

[0317] Please refer to Figure 9 The present application provides a vehicle voice interaction method. The vehicle voice interaction method comprises:

[0318] 051: receiving the voice request of the user for navigating to the note name of the point of interest forwarded by the vehicle;

[0319] 052: performing domain classification on the voice request;

[0320] 053: after determining that the classification result of the domain classification is a navigation domain, processing the voice request according to a preset sentence pattern set;

[0321] 054: according to the processing result, parsing the note name of the point of interest and obtaining the intent of navigating to the note name of the point of interest;

[0322] 055: match the analysis result with a preset key-value pair in the navigation domain, wherein the preset key-value pair includes a unique key name composed of a user identifier and a field corresponding to the interest point remark name;

[0323] 056: according to the intention of navigating to the interest point remark name, perform slot extraction to obtain a first slot extraction result after the analysis result matches the corresponding field of the preset key-value pair;

[0324] 057: output the first natural language understanding result to a map of a vehicle-mounted system in combination with the first slot extraction result, and the map completes navigation responding to the voice request according to the first natural language understanding result.

[0325] The application further provides a server. The server includes a processor and a memory, and the memory stores a computer program. The processor is configured to receive a voice request of a user for navigating to an interest point remark name forwarded by a vehicle, perform domain classification on the voice request, process the voice request according to a preset sentence pattern set after determining that the classification result of the domain classification is a navigation domain, parse the interest point remark name and obtain the intention of navigating to the interest point remark name according to the processing result, match the analysis result with a preset key-value pair in the navigation domain, wherein the preset key-value pair includes a unique key name composed of a user identifier and a field corresponding to the interest point remark name, perform slot extraction to obtain a first slot extraction result according to the intention of navigating to the interest point remark name after the analysis result matches the corresponding field of the preset key-value pair, and output the first natural language understanding result to a map of a vehicle-mounted system in combination with the first slot extraction result, and the map completes navigation responding to the voice request according to the first natural language understanding result.

[0326] First, a voice request of a user for navigating to an interest point remark name forwarded by a vehicle is received. For example, the user has previously remarked an interest point “Zhongguancun” as a remark name “company”, and the voice request of the user for navigating to the interest point remark name can be “navigate to the company”, wherein “company” is the interest point remark name.

[0327] Then, the voice request is classified in a domain. It can be understood that the voice request issued by the user can be classified into different domains, for example, a music domain, a weather domain, a vehicle control domain and a navigation domain.

[0328] Then, after determining that the classification result of the domain classification is the navigation domain, the voice request is processed according to the preset sentence pattern set. Understandably, since the application is to realize the correct understanding of the voice request of the user navigation to the interest point nickname, the application processes the voice request according to the preset sentence pattern set to obtain whether the intent of the user navigation is to navigate to the interest point nickname, that is, whether it is the intent of navigation to the interest point nickname "navigation_cust" after determining that the classification result of the domain classification of the voice request is the navigation domain. The processing method is to match the voice request with the preset sentence pattern set to determine whether the sentence pattern of the voice request conforms to a certain sentence pattern in the preset sentence pattern set. That is, if it is determined that the domain of the voice request of the user is not the navigation domain, but the voice domain or other domains, the voice request issued by the user is not processed according to the preset sentence pattern combination, and no subsequent related operation is required.

[0329] Among them, the format of the preset sentence pattern set is "sentence pattern: intent", including "navigation to XXX", "navigation to my XXX" and "go to XXX" and the like. Among them, "XXX" can include the interest point nickname or a specific place name, and the above-mentioned sentence patterns can preliminarily determine that the intent of the user is the intent of navigation to the interest point nickname, for example, the sentence pattern "navigation to my company" preliminarily determines that the corresponding intent is the intent of navigation to the interest point nickname.

[0330] In addition, the vehicle sometimes cannot correctly recognize the sentence pattern of "navigation to the interest point nickname", and cannot meet the needs of the user to navigate to the interest point nickname. Therefore, the application pre-stores the preset sentence pattern set on the server side, and when the sentence pattern of the voice request issued by the user matches a certain sentence pattern in the preset sentence pattern set, it is determined that the intent of the current voice request of the user is the intent of obtaining navigation to the interest point nickname. Among them, the analysis method can be sentence pattern matching or regular matching.

[0331] Of course, the analysis of the corresponding interest point nickname and the obtaining of the intent of navigation to the interest point nickname can also be realized by training the corresponding machine learning model.

[0332] Then, the analysis result needs to be verified, whether the analysis result is the interest point nickname stored in advance, by matching the analysis result with the preset key-value pair of the navigation domain, and the analysis result matches the corresponding field of the preset key-value pair, which means that the verification is passed, and the voice request contains the interest point nickname, that is, the intent of the voice request of navigation to the interest point nickname is correct.

[0333] The preset key-value pair in the navigation field is a key-value pair pre-stored in the server side, and the preset key-value pair includes a unique key name composed of a user identifier and a field corresponding to a point of interest note name. For example, the preset key-value pair in the navigation field is "Key: navigation_id, Value: [Sanji home, Grandma's home, my favorite place, children's school]", where Key represents a unique key name composed of a user identifier, used to identify the identity of the user; Value represents a field corresponding to a point of interest note name, including all point of interest note names noted by the user in the past. The point of interest note names "Sanji home", "Grandma's home", "my favorite place", and "children's school" are all note names noted by the user in the map of the vehicle-mounted system.

[0334] After verifying that the voice request of the user is correct in the intent of navigating to the point of interest note name, the voice request can be slot extracted according to the intent of navigating to the point of interest note name, and a first slot extraction result is obtained. For example, the voice request of the user is "navigate to Sanji home", and after verifying that the intent of the voice request is the intent of navigating to the point of interest note name, the voice request is slot extracted, and the first slot extraction result includes a slot name "cust_name" and a corresponding slot value "Sanji home".

[0335] Finally, the first natural language understanding result is output to the map of the vehicle-mounted system in combination with the first slot extraction result, and the map responds to the voice request of navigation according to the first natural language understanding result. For example, the server side performs natural language understanding on the voice request "navigate to Sanji home" in combination with the first slot extraction result "Sanji home" described above, and outputs the first natural language understanding result as "navigate to Sanji home", domain: navigation, intent: navigation_cust, slot: { "cust_name": "Sanji home"}. Then, the map end of the vehicle-mounted system receives the first natural language understanding result, requests a point of interest pre-stored in the map according to the point of interest note name in the first natural language understanding result, extracts the correct point of interest according to the point of interest note name, and performs navigation.

[0336] In this way, the vehicle voice interaction method of the present application can accurately identify voice requests containing point of interest note names, and meet the navigation needs of users who want to navigate to voice requests containing point of interest note names.

[0337] It can be understood that when the user's voice request meets a certain sentence pattern in the preset sentence pattern set but does not include the interest point nickname, and the user does not modify the interest point nickname in the preset key-value pair, since the vehicle cannot distinguish whether the voice request is a voice request with an interest point nickname, if the vehicle continues to identify the intent of the voice request as the intent of navigating to the interest point nickname, since the interest point nickname is not included in the voice request, the analysis result cannot be matched with the preset key-value pair to obtain the corresponding result, and the voice request cannot be slot extracted according to the intent of navigating to the interest point nickname, so that the voice interaction process cannot be successfully completed.

[0338] Therefore, please refer to Figure 10 After step 055, the vehicle voice interaction method comprises:

[0339] 0551: In the case that the field corresponding to the interest point nickname in the preset key-value pair matched with the analysis result is empty, the intent of navigating to the interest point nickname is rewritten as the intent of navigating to the interest point;

[0340] 0552: According to the intent of navigating to the interest point, the slot extraction result is obtained by slot extracting the voice request;

[0341] 0553: The second natural language understanding result is output to the map in combination with the second slot extraction result, and the map completes the navigation responding to the voice request according to the second natural language understanding result.

[0342] The processor is configured to, in the case that the field corresponding to the interest point nickname in the preset key-value pair matched with the analysis result is empty, rewrite the intent of navigating to the interest point nickname as the intent of navigating to the interest point; according to the intent of navigating to the interest point, the second slot extraction result is obtained by slot extracting the voice request; and the second natural language understanding result is output to the map in combination with the second slot extraction result, and the map completes the navigation responding to the voice request according to the second natural language understanding result.

[0343] After matching the analysis result with the preset key-value pairs in the navigation field, in the case that the user does not modify the interest point remark name in the preset key-value pairs, and in the matching preset key-value pairs, the field corresponding to the matched interest point remark name is empty, it indicates that there is no interest point remark name in the analysis result that matches the preset key-value pairs. At this time, the intent of navigating to the interest point remark name is rewritten as the intent of navigating to the interest point (navigation_poi). For example, the user voice request is "navigate to Zhongguancun", the analysis result is "Zhongguancun", and the field Value corresponding to the interest point remark name in the preset key-value pairs is empty. At this time, the intent of navigating to the interest point remark name can be directly rewritten as the intent of navigating to the interest point "Zhongguancun". In this way, it can be realized to distinguish whether the voice request issued by the user is a voice request containing an interest point remark name.

[0344] Then, according to the intent of navigating to the interest point, slot extraction is performed on the voice request to obtain a second slot extraction result. That is, after the intent is rewritten as navigating to the interest point, slot extraction is performed on the voice request "navigate to Zhongguancun" in the above example to obtain a second slot extraction result "Zhongguancun".

[0345] After that, the second natural language understanding result is output to the map in combination with the second slot extraction result, and the map completes navigation in response to the voice request according to the second natural language understanding result. For example, the server side performs natural language understanding on the voice request "navigate to Zhongguancun" in combination with the slot name "poi_name" and the corresponding layer slot value "Zhongguancun" included in the above-mentioned second slot extraction result, and outputs the second natural language understanding result as: "navigate to Zhongguancun", domain: navigation, intent: navigation_poi, slot: { "poi_name": "Zhongguancun"}. Then, the map end of the vehicle-mounted system receives the second natural language understanding result, and navigates according to the interest point in the first natural language understanding result.

[0346] In this way, when the voice request of the user meets a certain sentence pattern in the preset sentence pattern set but does not include an interest point remark name, and in the matching preset key-value pairs, the field corresponding to the interest point remark name is empty, through rewriting the intent of the voice request and slot extraction, the correct intent recognition and slot extraction of the voice request can also be realized, and then the voice interaction is completed.

[0347] It can be understood that the voice request of the user meets a certain sentence in the preset sentence set and includes the interest point nickname, but the user modifies the interest point nickname in the preset key-value pair. Therefore, if the interest point nickname in the analysis result is the interest point nickname before modification in the preset key-value pair, the interest point nickname cannot be matched to the interest point nickname expected by the user. At this time, if the vehicle continues to identify the intent of the voice request as the intent of navigating to the interest point nickname, the analysis result cannot be matched to the corresponding result in the preset key-value pair, and the voice request cannot be slot extracted according to the intent of navigating to the interest point nickname, so that the voice interaction process cannot be successfully completed.

[0348] Therefore, please refer to Figure 11 After step 055, the vehicle voice interaction method includes:

[0349] 0554: In the case that the matching of the analysis result and the preset key-value pair does not match the corresponding field of the preset key-value pair, the intent of navigating to the interest point nickname is rewritten as the intent of navigating to the interest point;

[0350] 0555: According to the intent of navigating to the interest point, the slot extraction of the voice request is performed to obtain a second slot extraction result;

[0351] 0556: The second natural language understanding result is output to the map in combination with the second slot extraction result, and the map completes the navigation responding to the voice request according to the second natural language understanding result.

[0352] The processor is configured to rewrite the intent of navigating to the interest point nickname as the intent of navigating to the interest point in the case that the matching of the analysis result and the preset key-value pair does not match the corresponding field of the preset key-value pair, to perform slot extraction of the voice request according to the intent of navigating to the interest point to obtain a second slot extraction result, and to output a second natural language understanding result in combination with the second slot extraction result to the map, and the map completes the navigation responding to the voice request according to the second natural language understanding result.

[0353] After matching the analysis result and the preset key-value pair in the navigation field, in the case that the user modifies the interest point remark name in the preset key-value pair, the result of matching the analysis result and the preset key-value pair is that the interest point remark name in the analysis result is matched to the corresponding field of the preset key-value pair, and then the intent of navigating to the interest point remark name is rewritten as the intent of navigating to the interest point. For example, the voice request is "navigate to the house of the third uncle", the analysis result is that the interest point remark name is "the house of the third uncle", and the corresponding field of the interest point remark name in the preset key-value pair is "Value: [company, grandmother's house, my favorite place, children's school]". The interest point remark name in the analysis result is not matched to the corresponding field of the interest point remark name before the modification of the preset key-value pair. At this time, the intent of navigating to the interest point remark name is rewritten as the intent of navigating to the interest point.

[0354] Then, according to the intent of navigating to the interest point, slot extraction is performed on the voice request to obtain a second slot extraction result. That is, after the intent is rewritten as the intent of navigating to the interest point, slot extraction is performed on the voice request "navigate to the house of the third uncle" in the above example to obtain a second slot extraction result including a slot name "poi_name" and a corresponding layer slot value "the house of the third uncle".

[0355] Then, the second natural language understanding result is output to the map in combination with the second slot extraction result, and the map completes navigation in response to the voice request according to the second natural language understanding result. For example, the server side performs natural language understanding on the voice request "navigate to the house of the third uncle" in combination with the slot name "poi_name" and the corresponding layer slot value "the house of the third uncle" included in the above-mentioned second slot extraction result, and outputs a second natural language understanding result as "navigate to the house of the third uncle", domain: navigation, intent: navigation_poi, slot: { "poi_name": "the house of the third uncle"}. Then, the map end of the vehicle-mounted system receives the second natural language understanding result, and navigates according to the interest point in the first natural language understanding result.

[0356] In this way, in the case that the voice request of the user meets a certain sentence pattern in the preset sentence pattern set and includes an interest point remark name, if the interest point remark name in the analysis result is the interest point remark name before the modification of the preset key-value pair, the interest point remark name in the analysis result cannot be matched to the interest point remark name expected by the user, that is, the analysis result and the preset key-value pair are not matched to the corresponding field of the preset key-value pair. Through the rewriting of the intent and the slot extraction of the voice request, correct intent recognition and slot extraction of the voice request can also be realized, and then voice interaction can be completed.

[0357] In addition, since different nicknames correspond to different point of interest addresses or destination addresses for each user, for example, the best restaurant is XXXX barbecue for user Zhang San, and the best restaurant is XXXX Japanese food for user Li Si, the navigation demand of the user cannot be correctly recognized to realize the navigation demand of the user to the navigation demand containing the nickname. Therefore, the preset key-value pair of the application includes different user identifiers for identifying different users, thereby distinguishing the voice requests of different users and meeting the navigation demands of different users.

[0358] Specifically, please refer to Figure 12 , step 055 includes:

[0359] 05501: obtaining the current user identifier corresponding to the voice request;

[0360] 05502: determining the target key-value pair corresponding to the unique key name and the current user identifier in the preset key-value pair in the navigation field;

[0361] 05503: matching the point of interest nickname in the analysis result and the field corresponding to the point of interest nickname in the target key-value pair.

[0362] The processor is configured to obtain the current user identifier corresponding to the voice request; determine the target key-value pair corresponding to the unique key name and the current user identifier in the preset key-value pair in the navigation field; and match the point of interest nickname in the analysis result and the field corresponding to the point of interest nickname in the target key-value pair.

[0363] When matching the analysis result and the preset key-value pair in the navigation field, the current user identifier corresponding to the voice request can be obtained first, for example, the current user identifier can be identified according to the tone of the user, that is, the navigation_id1 of the current user is obtained.

[0364] Then, the target key-value pair corresponding to the unique key name and the current user identifier in the preset key-value pair in the navigation field is determined. For example, the preset key-value pair of the navigation field stored in the server side includes: “Key: navigation_id1, Value: [company, home, my favorite restaurant, children's school]” and “Key: navigation_id2, Value: [cousin's home, grandmother's home, my favorite place, children's school]”, and the target key-value pair corresponding to the current user identifier navigation_id1 is determined as: “Key: navigation_id1, Value: [company, home, my favorite restaurant, children's school]”.

[0365] Therefore, at this time, the interest point nickname in the analysis result and the field corresponding to the interest point nickname in the target key-value pair can be matched. For example, the current user's voice request is "navigate to my favorite restaurant", and the interest point nickname "my favorite restaurant" in the analysis result and the field "Value: [company, home, my favorite restaurant, children's school]" corresponding to the interest point nickname in the target key-value pair are matched.

[0366] In this way, the preset key-value pair of the present application includes different user identifiers for identifying different users. By obtaining the current user identifier corresponding to the voice request, the target key-value pair corresponding to the unique key name and the current user identifier in the preset key-value pair in the navigation field is determined, and the interest point nickname in the analysis result and the field corresponding to the interest point nickname in the target key-value pair are matched, so that the voice requests issued by different users can be distinguished, and the different navigation needs of different users can be met.

[0367] In addition, since users have different nickname needs at different times, the user will need to update the preset key-value pair pre-stored on the server side to meet the navigation needs of the user at different times. The present application will introduce the scheme for updating the preset key-value pair pre-stored on the server side.

[0368] Specifically, when it is necessary to add a nickname to the preset key-value pair on the server side to update the preset key-value pair, the vehicle voice interaction method comprises:

[0369] 0581: receiving the nickname information of the user adding the nickname forwarded by the vehicle;

[0370] 0582: updating the preset key-value pair according to the nickname information.

[0371] The processor is configured to receive the nickname information of the user adding the nickname forwarded by the vehicle, and update the preset key-value pair according to the nickname information.

[0372] When it is necessary to add a nickname to the preset key-value pair on the server side, the user can add the nickname on the map side of the vehicle. At this time, the server side can call an application programming interface to synchronously receive the nickname information of the user adding the nickname forwarded by the vehicle, and update the preset key-value pair according to the nickname information.

[0373] For example, the user can issue a voice request "set interest point 1 as my mom's home" on the map side of the vehicle to add the nickname "my mom's home". At this time, the server side calls an application programming interface to synchronously receive the added nickname "my mom's home", so as to add the nickname "my mom's home" to the field corresponding to the interest point nickname in the preset key-value pair.

[0374] Thus, when a nickname needs to be added in the preset key-value pair on the server side, the server side of the application can receive the nickname information added by the user through the vehicle in real time by calling the application programming interface, and update the preset key-value pair according to the nickname information.

[0375] More specifically, step 0582 includes: determining the current user identifier corresponding to the nickname information; and

[0376] Adding the nickname of the nickname information to the field corresponding to the interest point nickname in the preset key-value pair with the unique key name matching the current user identifier.

[0377] The processor is configured to determine the current user identifier corresponding to the nickname information, and add the nickname of the nickname information to the field corresponding to the interest point nickname in the preset key-value pair with the unique key name matching the current user identifier.

[0378] When adding the nickname according to the nickname information, the current user identifier corresponding to the nickname information can be determined first, and the preset key-value pair with the unique key name matching the current user identifier can be determined according to the current user identifier, so that the nickname of the nickname information can be added to the field corresponding to the interest point nickname in the matched preset key-value pair, thereby distinguishing the voice request of adding the nickname issued by different users and meeting the adding demand of different users.

[0379] For example, if the current user identifier corresponding to the nickname information is determined to be "navigation_id2", the preset key-value pair with the unique key name matching the current user identifier is "Key: navigation_id2, Value: [Three Uncle's House, Grandma's House, My Favorite Place, Children's School]", and therefore, the nickname of the nickname information can be accurately added to the field corresponding to the interest point nickname in the preset key-value pair. For example, if the nickname of the nickname information is "the nearest hospital from my home", the preset key-value pair after adding the nickname becomes "Key: navigation_id2, Value: [Three Uncle's House, Grandma's House, My Favorite Place, Children's School, the nearest hospital from my home]", thereby completing the process of adding the nickname in the preset key-value pair according to the voice request of adding the nickname issued by the user.

[0380] Thus, when updating the preset key-value pair according to the nickname information, the current user identifier corresponding to the nickname information can be determined first, and the nickname of the nickname information can be added to the field corresponding to the interest point nickname in the preset key-value pair with the unique key name matching the current user identifier, thereby meeting the demand of adding the nickname in the preset key-value pair according to the voice request of adding the nickname issued by different users.

[0381] In addition, when the nickname needs to be modified in the preset key-value pair on the server side to update the preset key-value pair, the vehicle voice interaction method comprises:

[0382] The nickname information modified by the user in the vehicle is obtained in a polling manner; and

[0383] The preset key-value pair is updated according to the nickname information.

[0384] The processor is configured to obtain the nickname information modified by the user in the vehicle in a polling manner, and update the preset key-value pair according to the nickname information.

[0385] Polling is a way for a processor (CPU) to decide how to provide services for peripheral devices, also known as "programmed I / O". The concept of polling is that the CPU sends inquiries at regular intervals to each peripheral device to determine whether it needs services, and if so, provides services. After the services are completed, the next peripheral device is asked, and the process is repeated continuously.

[0386] That is, when the nickname needs to be modified in the preset key-value pair on the server side, the server side can obtain the nickname information modified by the user in the vehicle in a polling manner. The server side can send inquiries to the vehicle side at regular intervals to determine whether the nickname needs to be modified. When the server side receives a service request for modifying the nickname from the vehicle side, the server side obtains the nickname information modified by the user in the vehicle, and modifies the preset key-value pair according to the nickname information to update the preset key-value pair.

[0387] For example, the server side sends inquiries to the vehicle side every 1 minute to determine whether the nickname needs to be modified. When the user modifies the nickname, the server side can obtain the nickname information modified by the user in the vehicle, for example, the modified nickname information is to modify "my uncle's house" to "my aunt's house". Thus, the field corresponding to the interest point nickname in the preset key-value pair can be modified from "my uncle's house" to "my aunt's house", and the modification of the preset key-value pair is completed.

[0388] In this way, when the nickname needs to be modified in the preset key-value pair on the server side, the server side of the present application can obtain the nickname information modified by the user in the vehicle in a polling manner, and update the preset key-value pair according to the nickname information.

[0389] Step 0592 comprises: determining a current user identifier corresponding to the nickname information; and

[0390] The field corresponding to the interest point nickname in the preset key-value pair is modified according to the nickname information.

[0391] The processor is configured to determine a current user identifier corresponding to the nickname information, and modify a field corresponding to the interest point nickname in a preset key-value pair with a unique key name matching the current user identifier according to the nickname of the nickname information.

[0392] In the process of modifying the preset key-value pair according to the nickname information, the current user identifier corresponding to the nickname information can be determined first, and the preset key-value pair with the unique key name matching the current user identifier can be determined according to the current user identifier, so that the field corresponding to the interest point nickname in the matched preset key-value pair can be modified according to the nickname of the nickname information, thereby distinguishing the voice request for modifying the nickname issued by different users and meeting the modification requirements of different users.

[0393] For example, it is determined that the current user identifier corresponding to the nickname information is "navigation_id2", and the preset key-value pair with the unique key name matching the current user identifier is "Key: navigation_id2, Value: [Third Uncle's House, Grandmother's House, My Favorite Place, Children's School]". Therefore, the field corresponding to the interest point nickname in the matched preset key-value pair can be modified according to the nickname information, for example, the user modifies the nickname "Third Uncle's House" to "Auntie's House", and the preset key-value pair after adding the nickname becomes "Key: navigation_id2, Value: [Auntie's House, Grandmother's House, My Favorite Place, Children's School]", thereby completing the process of modifying the nickname in the preset key-value pair for the voice request for modifying the nickname issued by the user.

[0394] Therefore, in the process of updating the preset key-value pair according to the nickname information, the current user identifier corresponding to the nickname information can be determined first, the field corresponding to the interest point nickname in the preset key-value pair with the unique key name matching the current user identifier can be modified according to the nickname of the nickname information, thereby meeting the demand of modifying the nickname in the preset key-value pair for the voice request for modifying the nickname issued by different users.

[0395] The application further provides a nonvolatile computer readable storage medium containing a computer program. When the computer program is executed by one or more processors, the vehicle voice interaction method described in any of the above embodiments is implemented.

[0396] It can be understood that the computer program includes computer program code. The computer program code can be in the form of source code, object code, executable files or some intermediate forms, etc. The computer readable storage medium can include any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution medium, etc.

Claims

1. A vehicle voice interaction method, characterized in that, include: The user receives a voice request forwarded by the vehicle to add notes to points of interest in the in-vehicle navigation map. The voice request is identified to obtain a note name; Establish a mapping relationship between the annotation names and the points of interest; Based on the mapping relationship, the navigation request to the noted name after the user wakes up the vehicle system by voice next time will be determined as the navigation request to the corresponding point of interest. The method further includes: Receive the first report from the user initiating navigation to the point of interest through the vehicle system; Based on the first reported information, query all users' collections and notes on the addresses of the points of interest within the notes information set; Based on the query results and the type of the points of interest, determine the voice broadcast content to be fed back according to predetermined rules; The voice broadcast content is sent to the vehicle and broadcast to the user inside the vehicle, so as to guide the user to add a note to the point of interest after completing the collection of the point of interest through the map of the in-vehicle system.

2. The vehicle voice interaction method according to claim 1, characterized in that, The voice request forwarded by the receiving vehicle, in which the user sets annotations for points of interest in the vehicle system's navigation map, includes: Receive the user's current navigation request forwarded by the vehicle; Based on the current navigation request, the system will provide the user with navigation search results containing multiple points of interest in the in-vehicle navigation map for confirmation. The user who receives the voice request forwarded by the vehicle sets a note for the confirmed point of interest.

3. The vehicle voice interaction method according to claim 1, characterized in that, Following the step of receiving a voice request forwarded by the vehicle from a user to set annotations for points of interest in the in-vehicle navigation map, the vehicle voice interaction includes: Slot identification is performed on the voice request; If the result of the slot identification does not include point of interest slot information, and if the navigation application interface of the vehicle determines that the voice request has a destination field, then a note is set for the point of interest corresponding to the destination field.

4. The vehicle voice interaction method according to claim 3, characterized in that, If the result of the slot recognition does not include point of interest (POI) slot information, and if the vehicle's navigation application programming interface determines that the voice request has a destination field, then a note is set for the POI corresponding to the destination field, including: If the navigation application interface of the vehicle determines that the voice request has multiple destination fields, then the user is asked the desired destination field by voice broadcast. Add notes to the points of interest corresponding to the destination field in the user's confirmation reply.

5. The vehicle voice interaction method according to claim 3, characterized in that, If the result of the slot recognition does not include point of interest (POI) slot information, and if the vehicle's navigation application programming interface determines that the voice request has a destination field, then a note is set for the POI corresponding to the destination field, including: If the navigation application interface of the vehicle determines that the voice request contains a destination field, then a note is set for the point of interest corresponding to the destination field.

6. The vehicle voice interaction method according to claim 3, characterized in that, The vehicle voice interaction method includes: If the vehicle's navigation application interface determines that the destination field of the voice request does not have a result, then the system will issue a voice broadcast indicating that the point of interest was not found.

7. The vehicle voice interaction method according to claim 1, characterized in that, Following the step of receiving a voice request forwarded by the vehicle from a user to set annotations for points of interest in the in-vehicle navigation map, the vehicle voice interaction method includes: Slot identification is performed on the voice request; If the result of the slot identification includes point of interest (POI) slot information, add notes to the POIs obtained by navigation search based on the POI slot information.

8. The vehicle voice interaction method according to claim 7, characterized in that, When the result of the slot identification includes point of interest (POI) slot information, the POIs obtained from the navigation search of the POI slot information are annotated, including: If the navigation search for the point of interest slot information yields multiple points of interest, the user is asked about their desired point of interest via voice broadcast. Add notes to the points of interest that users confirm in their replies.

9. The vehicle voice interaction method according to claim 7, characterized in that, When the result of the slot identification includes point of interest (POI) slot information, the POIs obtained from the navigation search of the POI slot information are annotated, including: If the navigation search for the point of interest slot information yields a point of interest, then a note is set for the point of interest.

10. The vehicle voice interaction method according to claim 7, characterized in that, When the result of the slot identification includes point of interest (POI) slot information, the POIs obtained from the navigation search of the POI slot information are annotated, including: If no search results are found for the point of interest slot information navigation search, the system will provide voice feedback indicating that the point of interest was not found.

11. A server, characterized in that, The server includes a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the vehicle voice interaction method according to any one of claims 1-10.

12. A non-volatile computer-readable storage medium containing a computer program, characterized in that, When the computer program is executed by one or more processors, it implements the vehicle voice interaction method according to any one of claims 1-10.

Images