Voice assistant awakening method, vehicle-mounted terminal and storage medium

Abstract

The invention provides a voice assistant awakening method, a vehicle-mounted terminal and a storage medium, and relates to the technical field of voice assistants. According to the method, identity information of a user sending suspected wake-up voice data can be recognized, and an initial wake-up threshold value associated with the identity information of the user and historical wake-up habit data of a voice assistant are searched from a preset database; and adjusting the initial wake-up threshold according to the similarity between the real-time wake-up habit data and the historical wake-up habit data and the noise energy to obtain a real-time wake-up threshold. The real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy, so that the reliability of the obtained real-time wake-up threshold is high, and the suspected wake-up voice data is evaluated to obtain the target wake-up value; and when the target wake-up value is greater than or equal to the real-time wake-up threshold value, waking up the voice assistant. Because the reliability of the real-time wake-up threshold is high, the wake-up of the voice assistant at the moment does not belong to false wake-up, and the sensitivity of wake-up of the voice assistant is also high.

Description

Technical Field

[0001] The present application relates to the field of voice assistant technology, and in particular to a voice assistant wake-up method, a vehicle-mounted terminal, and a storage medium. Background Art

[0002] With the rapid development of smart cockpit technology, voice assistants have become a standard feature of intelligent vehicles. Voice assistants need to interact with users through voice after being awakened.

[0003] Typically, the vehicle terminal can evaluate the real-time wake-up value of the voice data when monitoring the user's input voice data. If the real-time wake-up value is greater than the set wake-up threshold, it is determined that the user's input voice data contains the voice wake-up word, and the voice assistant is then woken up.

[0004] However, if the wake-up threshold is set too low, the voice assistant is prone to false wake-ups. Conversely, if the wake-up threshold is set too high, the voice assistant's wake-up sensitivity is low, making it difficult to wake up. Therefore, how to balance the high probability of false wake-ups and the low wake-up sensitivity of the voice assistant is a current problem. Summary of the Invention

[0005] The present application provides a voice assistant wake-up method, a vehicle-mounted terminal, and a storage medium, which are used to solve the problem in the prior art that if the wake-up threshold is set too low, the voice assistant is very likely to be woken up by mistake; conversely, if the wake-up threshold is set too high, the voice assistant has low wake-up sensitivity and is often difficult to wake up.

[0006] In a first aspect, the present application provides a voice assistant wake-up method, which is applied to an on-board terminal of a vehicle, where the operating system of the on-board terminal is installed with a voice assistant. The method provided by the present application includes:

[0007] When suspected wake-up voice data is identified, determining the noise energy of sound data in the vehicle's environment collected a preset time period ago, and collecting real-time wake-up habit data of the person emitting the suspected wake-up voice data;

[0008] Identify the identity of the user who issued the suspected wake-up voice data, and search the preset database for the initial wake-up threshold and historical wake-up habit data associated with the user's identity information;

[0009] Adjusting the initial wake-up threshold based on the similarity between the real-time wake-up habit data and the historical wake-up habit data, as well as the noise energy, to obtain a real-time wake-up threshold, wherein the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy;

[0010] Evaluate the suspected wake-up speech data to obtain the target wake-up value;

[0011] When the target wake-up value is greater than or equal to the real-time wake-up threshold, the voice assistant is woken up.

[0012] In some embodiments, adjusting the initial wake-up threshold according to the similarity between the real-time wake-up habit data and the historical wake-up habit data, and the noise energy, to obtain the real-time wake-up threshold includes:

[0013] Determine the similarity between the real-time wake-up habit data and the historical wake-up habit data;

[0014] Searching a preset database for a first weight corresponding to a similarity between the real-time wake-up habit data and the historical wake-up habit data;

[0015] Searching for a second weight corresponding to the noise energy from a preset database;

[0016] According to the formula H 实时 =k1k2H 初始 , determine the real-time wake-up threshold, where H 初始 is the initial wake-up threshold, H 实时 is the real-time wake-up threshold, k1 is the first weight, and k2 is the second weight.

[0017] In some embodiments, the real-time wake-up habit data includes the time when the suspected wake-up voice data is received; the historical wake-up habit data includes the number of times the voice assistant is woken up in each preset time period in history, the time period wake-up ratio of each time period to the total number of times the voice assistant is woken up, and the voice wake-up ratio of the number of times the voice assistant is woken up using a voice wake-up method to the total number of times the voice assistant is woken up; determining the similarity between the real-time wake-up habit data and the historical wake-up habit data includes:

[0018] The similarity between the real-time wake-up habit data and the historical wake-up habit data is determined based on the wake-up ratio by time period and the wake-up ratio by voice.

[0019] In some embodiments, determining the similarity between the real-time wake-up habit data and the historical wake-up habit data based on the time period wake-up ratio and the voice wake-up ratio includes:

[0020] According to the awakening ratio of the time period, searching for a first sub-similarity from a preset database, wherein the first sub-similarity is positively correlated with the awakening ratio of the time period;

[0021] According to the voice wake-up ratio, searching for a second sub-similarity from a preset database, wherein the second sub-similarity is positively correlated with the voice wake-up ratio;

[0022] The first sub-similarity and the second sub-similarity are weighted averaged to obtain the similarity between the real-time wake-up habit data and the historical wake-up habit data.

[0023] In some embodiments, before identifying the presence of suspected wake-up voice data, the method provided by this application further includes:

[0024] When real wake-up voice data is identified, collect the identity information of the user who sent the real wake-up voice data;

[0025] Evaluate the speech quality of real wake-up speech data based on a pre-trained speech quality assessment model. The speech quality assessment model is trained by inputting multiple historical real wake-up speech data and their corresponding historical speech qualities into the network to be trained.

[0026] Searching for an initial wake-up threshold corresponding to the voice quality of the actual wake-up voice data from a preset database;

[0027] The user identity information is associated with the initial wake-up threshold and stored.

[0028] In some implementations, identifying the identity of a user who issues suspected wake-up voice data includes:

[0029] Controlling the vehicle's onboard camera to shoot in the direction of the user who sent the suspected wake-up voice data to obtain an image containing the user's face;

[0030] Determine the identity of the user who issued the suspected wake-up voice data based on an image containing the user's face.

[0031] In some implementations, evaluating the suspected wake-up speech data to obtain a target wake-up value includes:

[0032] Evaluate the speech quality of suspected wake-up speech data based on a pre-trained speech quality assessment model. This model is trained by inputting multiple historical real wake-up speech data and their corresponding historical speech qualities into the network to be trained.

[0033] The target wake-up value corresponding to the voice quality of the suspected wake-up voice data is searched from a preset database.

[0034] In the second aspect, the present application also provides a vehicle-mounted terminal, including a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the vehicle-mounted terminal executes the method provided in the first aspect of the present application.

[0035] In a third aspect, the present application further provides a storage medium storing a computer program. When the computer program is executed by a processor, the computer executes the method provided in the first aspect of the present application.

[0036] In a fourth aspect, the present application also provides a computer program product, including a computer program, which, when executed, enables the vehicle-mounted terminal to execute the method provided in the first aspect of the present application.

[0037] The present application provides a voice assistant wake-up method, a vehicle-mounted terminal, and a storage medium, which can identify the identity information of the user who sends suspected wake-up voice data, and search the preset database for the initial wake-up threshold associated with the user identity information and the historical wake-up habit data for the voice assistant. Then, the initial wake-up threshold is adjusted according to the similarity between the real-time wake-up habit data and the historical wake-up habit data, as well as the noise energy, to obtain the real-time wake-up threshold. Since the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy. In this way, the obtained real-time wake-up threshold has high reliability. Then, the suspected wake-up voice data is evaluated to obtain the target wake-up value; when the target wake-up value is greater than or equal to the real-time wake-up threshold, the voice assistant is awakened. Since the reliability of the real-time wake-up threshold is high, waking up the voice assistant at this time is neither a false wake-up case, nor is the sensitivity of waking up the voice assistant high, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0039] Figure 1 A flowchart of the voice assistant wake-up method provided in an embodiment of the present application;

[0040] Figure 2 This is a functional module block diagram of the voice assistant wake-up device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely illustrative and are not intended to limit the scope of the present disclosure. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0042] The accompanying drawings illustrate various schematic diagrams of structures according to embodiments of the present disclosure. These figures are not drawn to scale, and for the purpose of clarity, certain details are exaggerated and certain details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0043] In the context of the present disclosure, when a layer / element is referred to as being "on" another layer / element, it can be directly on the other layer / element or an intervening layer / element may be present therebetween. In addition, if a layer / element is "on" another layer / element in one orientation, it may be "below" the other layer / element when the orientation is reversed.

[0044] The following describes in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0045] The embodiment of the present application provides a voice assistant wake-up method, which is applied to a vehicle-mounted terminal, and the operating system of the vehicle-mounted terminal is installed with a voice assistant. Figure 1 As shown, the method provided in the embodiment of the present application includes:

[0046] S101: When suspected wake-up voice data is identified, determine the noise energy of sound data in the vehicle environment collected before a preset time, and collect real-time wake-up habit data of the suspected wake-up voice data.

[0047] Exemplarily, the vehicle-mounted terminal is provided with a microphone, which can monitor the sound data in the environment in which the vehicle is located. The vehicle-mounted terminal can perform voice recognition on the sound data (which may include ambient noise and the user's voice data), and when it is identified that the sound data contains suspected wake-up words, it is determined that there is suspected wake-up voice data. Furthermore, the noise energy (such as 40 decibels, 60 decibels) of the sound data (i.e., ambient noise) in the vehicle environment collected before a preset time length (such as 1s or 2s) is determined. It can be understood that the greater the energy of the ambient noise, the greater the degree of interference with the identification of whether there is wake-up voice data in the sound data. Among them, the real-time wake-up habit data of the suspected wake-up voice data may include the time period to which the moment when the suspected wake-up voice data is identified belongs.

[0048] S102: Identify the identity information of the user who sends the suspected wake-up voice data, and search the preset database for the initial wake-up threshold and historical wake-up habit data of the voice assistant associated with the user identity information.

[0049] For example, the vehicle's onboard camera can be controlled to shoot in the direction of the user who issued the suspected wake-up voice data to obtain an image containing the user's face; based on the image containing the user's face, the identity information of the user who issued the suspected wake-up voice data is determined. Among them, the number of times the voice assistant was awakened in each preset time period in history (such as 9:00-10:00; 10:00-11:00, etc. preset time periods), the time period wake-up ratio of each period to the total number of times the voice assistant was awakened (which can represent the user's time period preference for waking up the voice assistant), and the voice wake-up ratio of the number of times the voice assistant was awakened using the voice wake-up method to the total number of times the voice assistant was awakened (which can represent the user's preference for using the voice wake-up method to wake up the voice assistant).

[0050] It should be noted that, before the above S201, the method provided in the embodiment of the present application further includes:

[0051] When real wake-up voice data is identified, identity information of the user who sends the real wake-up voice data is collected.

[0052] The real wake-up voice data may be a recognized real wake-up word (eg, "Hello, my car"). Optionally, the real wake-up voice data may be collected in a quiet environment.

[0053] The voice quality of real wake-up speech data is evaluated based on a pre-trained voice quality assessment model. This model is trained by inputting multiple historical real wake-up speech data and their corresponding historical voice qualities into a network to be trained. The initial wake-up threshold corresponding to the voice quality of the real wake-up speech data is searched from a pre-set database. The user identity information is then associated with the initial wake-up threshold and stored. This allows for pre-configuration of the user's initial wake-up threshold.

[0054] It is understandable that the real wake-up voice data input by different users are different, for example, different accents (some may have dialects), different tones, and different sound intensities, etc. Therefore, it is necessary to evaluate the voice quality of the user's real wake-up voice data. In this way, the initial wake-up threshold determined based on the voice quality of the user's real wake-up voice data has high reliability.

[0055] S103: Adjusting the initial wake-up threshold according to the similarity between the real-time wake-up habit data and the historical wake-up habit data, and the noise energy, to obtain a real-time wake-up threshold, wherein the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy.

[0056] S103 can be specifically implemented as follows: determining the similarity between the real-time wake-up habit data and the historical wake-up habit data; searching a preset database for a first weight corresponding to the similarity between the real-time wake-up habit data and the historical wake-up habit data; searching a preset database for a second weight corresponding to the noise energy; and calculating the noise energy according to the formula H. 实时 =k1k2H 初始 , determine the real-time wake-up threshold, where H 初始 is the initial wake-up threshold, H 实时 is the real-time wake-up threshold, k1 is the first weight, and k2 is the second weight. In this way, the real-time wake-up threshold can be accurately obtained.

[0057] Exemplarily, the above-mentioned similarity can be determined in the following manner: the real-time wake-up habit data includes the moment when the suspected wake-up voice data is received; the historical wake-up habit data includes the number of times the voice assistant is woken up in each preset time period in history, the time period wake-up ratio to the total number of times the voice assistant is woken up, and the voice wake-up ratio of the number of times the voice assistant is woken up using the voice wake-up method to the total number of times the voice assistant is woken up; based on the time period wake-up ratio and the voice wake-up ratio, the similarity between the real-time wake-up habit data and the historical wake-up habit data is determined.

[0058] Furthermore, a first sub-similarity can be searched from a preset database based on the wake-up ratio by time period, where the first sub-similarity is positively correlated with the wake-up ratio by time period. A second sub-similarity can be searched from a preset database based on the wake-up ratio by voice, where the second sub-similarity is positively correlated with the wake-up ratio by voice. The first and second sub-similarity are weighted averaged to obtain the similarity between the real-time wake-up habit data and the historical wake-up habit data. It can be understood that the similarity determined in this manner is highly reliable.

[0059] S104: Evaluate the suspected wake-up voice data to obtain a target wake-up value.

[0060] Specifically, based on a pre-trained speech quality assessment model, the speech quality of the suspected wake-up speech data is evaluated, wherein the speech quality assessment model is obtained by inputting multiple historical real wake-up speech data and the corresponding historical speech quality as a training sample set into the network to be trained; from a preset database, the target wake-up value corresponding to the speech quality of the suspected wake-up speech data is searched.

[0061] S105: Determine whether the target wake-up value is less than the real-time wake-up threshold. If yes, execute S106.

[0062] S106: Wake up the voice assistant.

[0063] In summary, the embodiment of the present application provides a voice assistant wake-up method that can identify the identity information of the user who sends suspected wake-up voice data, and search the preset database for the initial wake-up threshold and the historical wake-up habit data of the voice assistant associated with the user identity information. Then, the initial wake-up threshold is adjusted according to the similarity between the real-time wake-up habit data and the historical wake-up habit data, as well as the noise energy, to obtain the real-time wake-up threshold. Since the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy. In this way, the obtained real-time wake-up threshold has high reliability. Then, the suspected wake-up voice data is evaluated to obtain the target wake-up value; when the target wake-up value is greater than or equal to the real-time wake-up threshold, the voice assistant is awakened. Since the reliability of the real-time wake-up threshold is high, waking up the voice assistant at this time is neither a false wake-up case, nor is the sensitivity of waking up the voice assistant high, thereby improving the user experience.

[0064] See also Figure 2 , the embodiment of the present application also provides a voice assistant wake-up device, which is configured in a vehicle-mounted terminal, and the operating system of the vehicle-mounted terminal is installed with a voice assistant. It should be noted that the basic principle and technical effects of the voice assistant wake-up device provided in the embodiment of the present application are the same as those in the above embodiment. For the sake of brief description, for parts not mentioned in the embodiment of the present application, please refer to the corresponding content in the above embodiment. Figure 2 As shown, the device provided in the embodiment of the present application includes a noise energy determination unit, a data acquisition unit, an identity recognition unit, a data search unit, a wake-up threshold determination unit, a wake-up value determination unit, and a voice wake-up unit, wherein:

[0065] a noise energy determination unit, configured to determine, when suspected wake-up voice data is identified, the noise energy of sound data in the vehicle environment collected before a preset time;

[0066] A data collection unit, configured to collect real-time wake-up habit data of the user emitting suspected wake-up voice data;

[0067] An identity recognition unit, used to identify the identity information of the user who sends the suspected wake-up voice data;

[0068] A data search unit, configured to search a preset database for an initial wake-up threshold value associated with user identity information and historical wake-up habit data for the voice assistant;

[0069] a wake-up threshold determination unit, configured to adjust an initial wake-up threshold according to a similarity between the real-time wake-up habit data and the historical wake-up habit data, and noise energy, to obtain a real-time wake-up threshold, wherein the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy;

[0070] a wake-up value determination unit, configured to evaluate the suspected wake-up speech data to obtain a target wake-up value;

[0071] The voice wake-up unit is used to wake up the voice assistant when the target wake-up value is greater than or equal to the real-time wake-up threshold.

[0072] In some embodiments, the wake-up threshold determination unit is specifically configured to determine the similarity between the real-time wake-up habit data and the historical wake-up habit data; search a preset database for a first weight corresponding to the similarity between the real-time wake-up habit data and the historical wake-up habit data; search a preset database for a second weight corresponding to the noise energy; and calculate the value of the wake-up threshold according to the formula H. 实时 =k1k2H 初始 , determine the real-time wake-up threshold, where H 初始 is the initial wake-up threshold, H 实时 is the real-time wake-up threshold, k1 is the first weight, and k2 is the second weight.

[0073] In some embodiments, the real-time wake-up habit data includes the moment when the suspected wake-up voice data is received; the historical wake-up habit data includes the number of times the voice assistant is woken up in each preset time period in history, the time period wake-up ratio to the total number of times the voice assistant is woken up, and the number of times the voice assistant is woken up using the voice wake-up method to the voice wake-up ratio to the total number of times the voice assistant is woken up; the wake-up threshold determination unit is further used to determine the similarity between the real-time wake-up habit data and the historical wake-up habit data based on the time period wake-up ratio and the voice wake-up ratio.

[0074] In some embodiments, the wake-up threshold determination unit is further used to search a first sub-similarity from a preset database based on the time period wake-up ratio, wherein the first sub-similarity is positively correlated with the time period wake-up ratio; search a second sub-similarity from a preset database based on the voice wake-up ratio, wherein the second sub-similarity is positively correlated with the voice wake-up ratio; and perform a weighted average of the first sub-similarity and the second sub-similarity to obtain the similarity between the real-time wake-up habit data and the historical wake-up habit data.

[0075] In some embodiments, the identity recognition unit is further configured to, upon recognizing the real wake-up voice data, collect the identity information of the user who sent the real wake-up voice data;

[0076] The apparatus provided in the embodiment of the present application further includes: a speech quality assessment unit, configured to assess the speech quality of the real wake-up speech data based on a pre-trained speech quality assessment model, wherein the speech quality assessment model is trained by inputting a plurality of historical real wake-up speech data and the corresponding historical speech qualities as a training sample set into a network to be trained;

[0077] The data search unit is further configured to search a preset database for an initial wake-up threshold value corresponding to the voice quality of the actual wake-up voice data;

[0078] The device provided in the embodiment of the present application also includes: a relationship establishment unit, which is used to establish and store an association relationship between the user identity information and the initial wake-up threshold.

[0079] In some embodiments, the identity recognition unit is specifically used to control the vehicle's on-board camera to shoot in the direction of the user who sent the suspected wake-up voice data, so as to obtain an image containing the user's face; and determine the identity information of the user who sent the suspected wake-up voice data based on the image containing the user's face.

[0080] In some embodiments, the wake-up value determination unit is specifically used to evaluate the voice quality of suspected wake-up voice data based on a pre-trained voice quality assessment model, wherein the voice quality assessment model is obtained by inputting multiple historical real wake-up voice data and corresponding historical voice qualities as a training sample set into the network to be trained; and searching for the target wake-up value corresponding to the voice quality of the suspected wake-up voice data from a preset database.

[0081] In addition, an embodiment of the present application also provides a vehicle-mounted terminal, including a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the vehicle-mounted terminal executes the method provided in the above embodiment of the present application.

[0082] In addition, an embodiment of the present application further provides a storage medium, which stores a computer program. When the computer program is executed by a processor, the computer executes the method provided in the above embodiment of the present application.

[0083] In addition, an embodiment of the present application also provides a computer program product, including a computer program, which, when executed, enables the vehicle terminal to execute the method provided in the above embodiment of the present application.

[0084] While the above description does not provide detailed technical details regarding the patterning of each layer, those skilled in the art will appreciate that various technical means can be employed to form layers, regions, and the like in desired shapes. Furthermore, those skilled in the art may devise methods that differ from those described above to achieve the same structure. Furthermore, while each embodiment has been described separately, this does not mean that the measures in each embodiment cannot be advantageously combined.

[0085] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0086] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A voice assistant wake-up method, characterized in that: The method is applied to an in-vehicle terminal of a vehicle, wherein the operating system of the in-vehicle terminal is installed with a voice assistant, and comprises: When suspected wake-up voice data is identified, determining the noise energy of sound data in the vehicle environment collected a preset time period ago, and collecting real-time wake-up habit data of the person emitting the suspected wake-up voice data; Identify the identity information of the user who issued the suspected wake-up voice data, and search the preset database for the initial wake-up threshold and historical wake-up habit data of the voice assistant associated with the user identity information; adjusting the initial wake-up threshold according to the similarity between the real-time wake-up habit data and the historical wake-up habit data, and the noise energy, to obtain a real-time wake-up threshold, wherein the real-time wake-up threshold is negatively correlated with the similarity and positively correlated with the noise energy; Evaluating the suspected wake-up voice data to obtain a target wake-up value; When the target wake-up value is greater than or equal to the real-time wake-up threshold, the voice assistant is woken up.

2. The method according to claim 1, characterized in that The adjusting the initial wake-up threshold according to the similarity between the real-time wake-up habit data and the historical wake-up habit data and the noise energy to obtain the real-time wake-up threshold includes: determining a similarity between the real-time wake-up habit data and the historical wake-up habit data; searching a preset database for a first weight corresponding to a similarity between the real-time wake-up habit data and the historical wake-up habit data; Searching a preset database for a second weight corresponding to the noise energy; According to the formula H 实时 =k1k2H 初始 , determine the real-time wake-up threshold, where H 初始 is the initial wake-up threshold, H 实时 is the real-time wake-up threshold, k1 is the first weight, and k2 is the second weight.

3. The method according to claim 2, characterized in that The real-time wake-up habit data includes the time when the suspected wake-up voice data is received; the historical wake-up habit data includes the number of times the voice assistant is woken up in each preset time period in history, the wake-up ratio of each time period to the total number of times the voice assistant is woken up, and the voice wake-up ratio of the number of times the voice assistant is woken up using a voice wake-up method to the total number of times the voice assistant is woken up; The determining the similarity between the real-time wake-up habit data and the historical wake-up habit data includes: The similarity between the real-time wake-up habit data and the historical wake-up habit data is determined according to the time period wake-up ratio and the voice wake-up ratio.

4. The method according to claim 3, characterized in that The determining, based on the time period wake-up ratio and the voice wake-up ratio, the similarity between the real-time wake-up habit data and the historical wake-up habit data includes: According to the awakening ratio of the time period, searching a first sub-similarity from a preset database, wherein the first sub-similarity is positively correlated with the awakening ratio of the time period; According to the voice wake-up ratio, searching for a second sub-similarity from a preset database, wherein the second sub-similarity is positively correlated with the voice wake-up ratio; A weighted average of the first sub-similarity and the second sub-similarity is performed to obtain a similarity between the real-time wake-up habit data and the historical wake-up habit data.

5. The method according to claim 1, wherein Before identifying the presence of suspected wake-up voice data, the method further includes: When the real wake-up voice data is identified, collecting the identity information of the user who sent the real wake-up voice data; Evaluate the voice quality of the real wake-up voice data based on a pre-trained voice quality assessment model, wherein the voice quality assessment model is trained by inputting a plurality of historical real wake-up voice data and corresponding historical voice qualities as a training sample set into a network to be trained; Searching a preset database for an initial wake-up threshold corresponding to the voice quality of the real wake-up voice data; The user identity information is associated with the initial wake-up threshold and stored.

6. The method according to any one of claims 1 to 5, characterized in that: The identifying the identity information of the user who sent the suspected wake-up voice data includes: Controlling the vehicle's onboard camera to shoot toward the direction of the user who sent the suspected wake-up voice data to obtain an image containing the user's face; The identity information of the user who sent the suspected wake-up voice data is determined based on the image containing the user's face.

7. The method according to any one of claims 1 to 5, characterized in that: The evaluating the suspected wake-up voice data to obtain a target wake-up value includes: Evaluate the voice quality of the suspected wake-up voice data based on a pre-trained voice quality assessment model, wherein the voice quality assessment model is trained by inputting a plurality of historical real wake-up voice data and corresponding historical voice qualities as a training sample set into a to-be-trained network; The target wake-up value corresponding to the voice quality of the suspected wake-up voice data is searched from a preset database.

8. A vehicle-mounted terminal comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the vehicle-mounted terminal executes the method according to any one of claims 1 to 7.

9. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the computer is caused to perform the method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed, the vehicle-mounted terminal executes the method as claimed in any one of claims 1 to 7.

Images