A vehicle-mounted audio recommendation method, device and vehicle

Abstract

The present invention discloses a method for in-vehicle audio recommendation, which includes obtaining current vehicle environment parameters and calculating a current comprehensive vehicle environment score; querying, according to the current comprehensive vehicle environment score, a current audio type that matches the current comprehensive vehicle environment score from a preset mapping relationship between the comprehensive vehicle environment score and audio types; obtaining the face recognition information of the driver collected and the current user ID that matches the face recognition information; calculating, by using a latent factor algorithm, a first scoring value of each user for each audio in the current audio type, and selecting therefrom a second scoring value of the current user ID for each audio in the current audio type, and sorting each audio in the current audio type corresponding to the second scoring value in descending order of the second scoring value to obtain an audio stream. The embodiment of the present invention can automatically recommend audio to reduce the audio switching operation of the user.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control, and particularly to an in-vehicle audio recommendation method, device and vehicle. Background Art

[0002] The intelligence of automobiles has been increasingly emphasized. Users not only have high requirements for driving safety performance, but also attach great importance to the intelligent and user-friendly experience during driving. For existing in-vehicle audio playback, users usually need to manually operate to select and play audio. At this time, the user's line of sight will move out of the driving field of view, which is likely to cause traffic accidents. Therefore, it is urgent to study an in-vehicle audio recommendation method that can automatically recommend audio to reduce the user's audio switching operation. Summary of the Invention

[0003] The present invention provides an in-vehicle audio recommendation method, device and vehicle, which can automatically recommend audio to reduce the user's audio switching operation, and at the same time can relieve the driver's mood and improve the comfort and safety of driving listening.

[0004] To achieve the above object, an embodiment of the present invention provides an in-vehicle audio recommendation method, including:

[0005] Obtain the current vehicle environment parameters, and calculate the current vehicle environment comprehensive score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle and the current vehicle speed;

[0006] According to the current vehicle environment comprehensive score, query the current audio type that matches the current vehicle environment comprehensive score from the preset mapping relationship between the vehicle environment comprehensive score and the audio type;

[0007] Obtain the face recognition information of the driver collected, and query the current user ID that matches the face recognition information from the preset user identity database;

[0008] Use the latent factor algorithm to calculate a number of first score values of each user for each audio in the current audio type, and select a number of second score values of the current user ID for each audio in the current audio type from them. Arrange each audio in the current audio type corresponding to the second score value in descending order of the second score value to obtain an audio stream, so that the in-vehicle audio device automatically starts playing audio according to the audio stream.

[0009] As an improvement of the above solution, the calculating the current vehicle environment comprehensive score according to the current vehicle environment parameters includes:

[0010] Obtain the score corresponding to the current vehicle environment parameters according to the current vehicle environment parameters;

[0011] Calculate the comprehensive score of the current vehicle environment according to the following formula:

[0012] Y = (0.5t + 0.3x + 0.2v)α

[0013] In the formula, t is the score corresponding to the current temperature inside the vehicle, x is the score corresponding to the current number of passengers inside the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient.

[0014] As an improvement to the above solution, the mapping relationship between the preset comprehensive score of the vehicle environment and the audio type includes:

[0015] When the comprehensive score of the vehicle environment is greater than 0.5 and less than or equal to 1, the audio type is the first audio type;

[0016] When the comprehensive score of the vehicle environment is greater than 0.2 and less than 0.5, the audio type is the second audio type;

[0017] When the comprehensive score of the vehicle environment is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

[0018] As an improvement to the above solution, the calculation of obtaining several first scoring scores of each audio in the current audio type for each user by using the latent factor algorithm includes:

[0019] Obtain the actual scoring score matrix R of each audio in the current audio type for each user;

[0020] Use the UV decomposition of the matrix to decompose the actual scoring score matrix R into two low-dimensional matrices P and Q, and use the gradient descent method to solve the objective function Obtain the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type; where, r ui represents the vector of the actual scoring score of user u for each audio i in the current audio type, q i represents the vector of the user audio type preference of matrix P, p u represents the vector of the elements of each audio in the current audio type of matrix Q, and T represents the transpose;

[0021] Multiply the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain several first scoring scores of each audio in the current audio type for each user.

[0022] Multiply the user audio type preference matrix Q by the element matrix P of each audio in the current audio type to obtain a number of first scoring values for each user for each audio in the current audio type.

[0023] To achieve the above object, an embodiment of the present invention further provides a vehicle audio recommendation device, including:

[0024] A current vehicle environment comprehensive score calculation module, configured to obtain current vehicle environment parameters and calculate a current vehicle environment comprehensive score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed;

[0025] A current audio type acquisition module, configured to query, according to the current vehicle environment comprehensive score, a current audio type that matches the current vehicle environment comprehensive score from a preset mapping relationship between the vehicle environment comprehensive score and the audio type;

[0026] A current user ID acquisition module, configured to acquire the face recognition information of the driver collected, and query a current user ID that matches the face recognition information from a preset user identity database;

[0027] An audio recommendation module, configured to use a latent factor algorithm to calculate a number of first scoring values for each user for each audio in the current audio type, and select a number of second scoring values for each audio in the current audio type by the current user ID from them, and sort each audio in the current audio type corresponding to the second scoring values in descending order of the second scoring values to obtain an audio stream, so that the vehicle audio device automatically starts playing audio according to the audio stream.

[0028] As an improvement of the above solution, the calculating the current vehicle environment comprehensive score according to the current vehicle environment parameters includes:

[0029] According to the current vehicle environment parameters, obtain the scores corresponding to the current vehicle environment parameters;

[0030] Calculate the current vehicle environment comprehensive score according to the following formula:

[0031] Y=(0.5t + 0.3x + 0.2v)α

[0032] In the formula, t is the score corresponding to the current in-vehicle temperature, x is the score corresponding to the current number of passengers in the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient.

[0033] As an improvement of the above solution, the preset mapping relationship between the vehicle environment comprehensive score and the audio type includes:

[0034] When the comprehensive vehicle environment score is greater than 0.5 and less than or equal to 1, the audio type is the first audio type;

[0035] When the comprehensive vehicle environment score is greater than 0.2 and less than 0.5, the audio type is the second audio type;

[0036] When the comprehensive vehicle environment score is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

[0037] As an improvement to the above solution, calculating, by using a latent factor algorithm, a plurality of first scoring scores of each user for each audio in the current audio type includes:

[0038] Obtaining an actual scoring score matrix R of each user for each audio in the current audio type;

[0039] Using the UV decomposition of the matrix to decompose the actual scoring score matrix R into two low-dimensional matrices P and Q, and using the gradient descent method to solve the objective function Obtaining a user audio type preference matrix Q' and an element matrix P' of each audio in the current audio type; where r ui represents a vector of the actual scoring scores of user u for each audio i in the current audio type, q i represents a vector of the user audio type preference of matrix P, p u represents a vector of the elements of each audio in the current audio type of matrix Q, and T represents the transpose;

[0040] Multiplying the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain a plurality of first scoring scores of each user for each audio in the current audio type.

[0041] To achieve the above object, an embodiment of the present invention further provides a vehicle, including: a vehicle body and the in-vehicle audio recommendation device as described above.

[0042] Compared with the prior art, an in-vehicle audio recommendation method provided by an embodiment of the present invention obtains current vehicle environment parameters and calculates a current vehicle environment comprehensive score; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed; according to the current vehicle environment comprehensive score, obtains a current audio type that matches the current vehicle environment comprehensive score; obtains the face recognition information of the driver collected and the current user ID that matches the face recognition information; uses the latent factor algorithm to calculate a number of first score values of each user for each audio in the current audio type, and selects a number of second score values of the current user ID for each audio in the current audio type from them, and sorts each audio in the current audio type corresponding to the second score values in descending order of the second score values to obtain an audio stream, so that the in-vehicle device automatically starts playing the audio according to the audio stream, realizing the active recommendation of in-vehicle audio. It can be seen that the embodiment of the present invention combines the in-vehicle temperature, the number of passengers in the vehicle, and the vehicle speed to bring a better audio experience to the user in the current vehicle environment, and uses the latent factor algorithm to provide the audio that meets the current user's preferences to the user. The embodiment of the present invention can actively recommend audio to the user, reduce the user's audio switching operation, and reduce the boredom and anxiety of the user due to listening to a single content, thereby improving driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flowchart of an in-vehicle audio recommendation method provided by an embodiment of the present invention;

[0044] Figure 2 is a structural block diagram of an in-vehicle audio recommendation device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0046] See Figure 1 , Figure 1 is a flowchart of an in-vehicle audio recommendation method provided by an embodiment of the present invention. The in-vehicle audio recommendation method includes:

[0047] S1. Obtain current vehicle environment parameters and calculate a current vehicle environment comprehensive score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed;

[0048] S2. Query the current audio type that matches the current vehicle environment comprehensive score from the mapping relationship between the preset vehicle environment comprehensive score and the audio type.

[0049] S3. Obtain the face recognition information of the driver collected, and query the current user ID that matches the face recognition information from the preset user identity database.

[0050] S4. Use the latent factor algorithm to calculate a number of first score values for each user for each audio in the current audio type, and select a number of second score values of the current user ID for each audio in the current audio type. Arrange each audio in the current audio type corresponding to the second score values in descending order of the second score values to obtain an audio stream, so that the in-vehicle audio device automatically starts playing the audio according to the audio stream.

[0051] It can be understood that the audio is music or a program.

[0052] Specifically, in step S1, the calculation of the current vehicle environment comprehensive score according to the current vehicle environment parameters includes:

[0053] Obtain the score corresponding to the current vehicle environment parameters according to the current vehicle environment parameters.

[0054] Calculate the current vehicle environment comprehensive score according to the following formula:

[0055] Y = (0.5t + 0.3x + 0.2v)α

[0056] In the formula, t is the score corresponding to the current temperature inside the vehicle, x is the score corresponding to the current number of passengers inside the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient.

[0057] Preferably, α = 0.27. It can be understood that setting the preset influence coefficient to 0.27 can make the current vehicle environment comprehensive score between 0 and 1. The higher the current vehicle environment comprehensive score, the more active the vehicle interior and exterior environment.

[0058] Exemplarily, the vehicle interior temperature is divided into three categories: low, medium, and high, and the corresponding scores can be taken from 1 to 3. When the current vehicle interior temperature is less than 18°C, the corresponding score is 1. When the current vehicle interior temperature is greater than or equal to 18°C and less than 24°C, the corresponding score is 2. When the current vehicle interior temperature is greater than or equal to 24°C, the corresponding score is 3.

[0059] Exemplarily, the current number of passengers in the vehicle is regarded as the score corresponding to the current number of passengers in the vehicle, and the corresponding score can take values from 1 to 5. When the current number of passengers in the vehicle is one person, the corresponding score is 1; when the current number of passengers in the vehicle is two people, the corresponding score is 2; when the current number of passengers in the vehicle is three people, the corresponding score is 3; when the current number of passengers in the vehicle is four people, the corresponding score is 4; when the current number of passengers in the vehicle is five people, the corresponding score is 5.

[0060] Exemplarily, the vehicle speed is divided into three categories: low, medium, and high, and the corresponding score can take values from 1 to 3. When the current vehicle speed is less than or equal to 60 Km / h, the corresponding score is 1; when the current vehicle speed is greater than 60 Km / h and less than 110 Km / h, the corresponding score is 2; when the current vehicle speed is greater than or equal to 110 Km / h, the corresponding score is 3.

[0061] It can be understood that in the embodiments of the present invention, the weight of the vehicle interior temperature accounts for 50%, the weight of the number of passengers in the vehicle accounts for 30%, and the weight of the vehicle speed accounts for 20%. These proportions can be adjusted at any time in the cloud to meet the personalized needs of the user group.

[0062] Specifically, in step S2, the mapping relationship between the preset comprehensive vehicle environment score and the audio type includes:

[0063] When the comprehensive vehicle environment score is greater than 0.5 and less than or equal to 1, the audio type is the first audio type;

[0064] When the comprehensive vehicle environment score is greater than 0.2 and less than 0.5, the audio type is the second audio type;

[0065] When the comprehensive vehicle environment score is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

[0066] Exemplarily, when the current comprehensive vehicle environment score Y is in the range of (0.5 - 1], it indicates a high environmental activity level, and the current audio type is queried from the mapping relationship between the preset comprehensive vehicle environment score and the audio type as a quiet type of audio; when the current comprehensive vehicle environment score Y value is in the range of (0.2 - 0.5], it indicates a medium environmental activity level, and the current audio type is queried from the mapping relationship between the preset comprehensive vehicle environment score and the audio type as a gentle type of audio; when the current comprehensive vehicle environment score is in the range of (0.1 - 0.2], it indicates a low environmental activity level, and the current audio type is queried from the mapping relationship between the preset comprehensive vehicle environment score and the audio type as an agitated type of audio.

[0067] It can be understood that most of the media in the current vehicle are QQ Music, Kuwo Music, Kugou Music, and Himalaya. Among them, when the audio is played, the in-vehicle computer can clearly identify the original audio type of the audio, such as rock, folk, cross talk, storytelling, comedy, etc. After uploading the original audio type to the cloud through the in-vehicle computer, it is reclassified and combined. Rock, electronic music, jokes, etc. are defined as restless audio; folk, ancient style, storytelling, Sinology, etc. are defined as quiet audio; emotions, travel, cross talk, etc. are defined as gentle audio. Each audio type can be added, changed, moved, and deleted in the cloud database to ensure the flow and update of resources.

[0068] It can be understood that the current audio type is the audio type currently recommended to the user. In the embodiment of the present invention, combined with the vehicle interior temperature, the number of passengers in the vehicle, and the vehicle speed, a better audio experience in the current vehicle environment is brought to the user.

[0069] Specifically, in step S3, by collecting the face recognition information of the driver and obtaining the current user ID matching the face recognition information to identify the driver's identity and obtain the audio preference corresponding to the driver's identity, wherein the preset user identity database is the mapping relationship between the face recognition information and the user ID.

[0070] Specifically, in step S4, the use of the latent factor algorithm to calculate a number of first score values for each user for each audio in the current audio type includes:

[0071] Obtain the actual score value matrix R of each user for each audio in the current audio type;

[0072] Using the UV decomposition of the matrix, decompose the actual score value matrix R into two low-dimensional matrices P and Q, and use the gradient descent method to solve the objective function Obtain the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type; where r ui Represents the vector of the actual score value of user u for each audio i in the current audio type, q i Represents the vector of the user audio type preference of matrix P, p u Represents the vector of the elements of each audio in the current audio type of matrix Q, and T represents the transpose;

[0073] Multiply the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain a number of first score values for each user for each audio in the current audio type.

[0074] It can be understood that the user audio type preference matrix Q', that is, the user-latent factor matrix, represents the preference degrees of different users for different audio type elements (quiet, gentle, restless...); the element matrix P' of each audio in the current audio type, that is, the latent factor-audio / program matrix, represents the components of quiet, gentle, restless... contained in each audio; using these two matrices, the estimated rating score matrix ^R of different users for different audios can be obtained. If represented by a matrix, it is ^R = Q'P'. T , where T represents transpose; at this time, the estimated rating scores of different users for different audios are obtained through the following steps:

[0075] The R value can be obtained according to the actual operations of each user in the in-vehicle and out-of-vehicle environments, that is, scoring the operations of each audio played on the in-vehicle computer by each user in different in-vehicle and out-of-vehicle environments. For example, single-loop = 5, share = 4, favorite = 3, active play = 2, finish listening = 1, skip = -2, blacklist = -5, and the actual rating score matrix R that can be obtained during analysis;

[0076] Using the UV decomposition of the matrix, the R matrix is decomposed into two low-dimensional matrices P and Q, and the product of the estimated values P' and Q' of the obtained P and Q is used to estimate the actual rating score matrix, hoping that the estimated rating score matrix ^R and the actual rating score matrix R do not differ too much, that is, solving the objective function It can be understood that the low-dimensional matrix P is the actual user audio type preference matrix, and the low-dimensional matrix Q is the element matrix of each audio in the actual current audio type;

[0077] At this time, according to the actual R value, the estimated values P' and Q' of these two matrices P and Q can be obtained by using the gradient descent method. After multiplying P' and Q', the estimated rating score matrix, that is, several first rating scores of each user for each audio in the current audio type, is obtained. The single-audio is sorted from high to low according to the first rating score and pushed to the user.

[0078] In step S4, when several second rating scores are obtained, each audio in the current audio type is sorted from high to low according to its corresponding second rating score and pushed to the user.

[0079] It is worth noting that the in-vehicle audio recommendation method provided by the embodiments of the present invention can be executed in the cloud, which can reduce the computing power occupation of the in-vehicle computer end.

[0080] An in-vehicle audio recommendation method provided by an embodiment of the present invention obtains current vehicle environment parameters and calculates a current vehicle environment comprehensive score; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed; according to the current vehicle environment comprehensive score, obtains a current audio type that matches the current vehicle environment comprehensive score; obtains the face recognition information of the driver collected and the current user ID that matches the face recognition information; uses a latent factor algorithm to calculate a number of first score values of each user for each audio in the current audio type, and selects a number of second score values of the current user ID for each audio in the current audio type from them, and sorts each audio in the current audio type corresponding to the second score values in descending order of the second score values to obtain an audio stream, so that the in-vehicle device automatically starts playing audio according to the audio stream, realizing the active recommendation of in-vehicle audio. It can be seen that the embodiment of the present invention combines the in-vehicle temperature, the number of passengers in the vehicle, and the vehicle speed to bring a better audio experience to the user in the current vehicle environment, uses the latent factor algorithm to provide audio that conforms to the current user's preference to the user, and the embodiment of the present invention can actively recommend audio to the user, reduce the user's audio switching operation, reduce the boredom and anxiety of the user caused by listening to a single content, and thus improve driving safety.

[0081] See Figure 2 , Figure 2 is a structural block diagram of an in-vehicle audio recommendation device 10 provided by an embodiment of the present invention. The in-vehicle audio recommendation device 10 includes:

[0082] A current vehicle environment comprehensive score calculation module 11, configured to obtain current vehicle environment parameters and calculate a current vehicle environment comprehensive score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed;

[0083] A current audio type obtaining module 12, configured to query, according to the current vehicle environment comprehensive score, a current audio type that matches the current vehicle environment comprehensive score from a preset mapping relationship between the vehicle environment comprehensive score and the audio type;

[0084] A current user ID obtaining module 13, configured to obtain the face recognition information of the driver collected and query the current user ID that matches the face recognition information from a preset user identity database;

[0085] The audio recommendation module 14 is used to calculate a number of first scoring scores for each audio in the current audio type for each user by using the latent factor algorithm, and select a number of second scoring scores for each audio in the current audio type for the current user ID from them. Then, each audio in the current audio type corresponding to the second scoring scores is sorted in descending order of the second scoring scores to obtain an audio stream, so that the in-vehicle audio device automatically starts playing the audio according to the audio stream.

[0086] Preferably, calculating the current vehicle environment comprehensive score according to the current vehicle environment parameters includes:

[0087] Obtaining the score corresponding to the current vehicle environment parameter according to the current vehicle environment parameter;

[0088] Calculating the current vehicle environment comprehensive score according to the following formula:

[0089] Y=(0.5t + 0.3x + 0.2v)α

[0090] In the formula, t is the score corresponding to the current in-vehicle temperature, x is the score corresponding to the current number of passengers in the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient.

[0091] Preferably, the mapping relationship between the preset vehicle environment comprehensive score and the audio type includes:

[0092] When the vehicle environment comprehensive score is greater than 0.5 and less than or equal to 1, the audio type is the first audio type;

[0093] When the vehicle environment comprehensive score is greater than 0.2 and less than 0.5, the audio type is the second audio type;

[0094] When the vehicle environment comprehensive score is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

[0095] Preferably, calculating a number of first scoring scores for each audio in the current audio type for each user by using the latent factor algorithm includes:

[0096] Obtaining the actual scoring score matrix R of each audio in the current audio type for each user;

[0097] Using the UV decomposition of the matrix to decompose the actual scoring score matrix R into two low-dimensional matrices P and Q, and using the gradient descent method to solve the objective function Obtaining the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type; where, r uiA vector q representing the actual rating scores of user u for each audio i in the current audio type i A vector p representing the user's audio type preference of matrix P u A vector representing the elements in each audio of the current audio type of matrix Q, where T represents the transpose;

[0098] Multiply the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain a number of first rating scores for each user for each audio in the current audio type.

[0099] It should be noted that the working processes of the various modules in the in-vehicle audio recommendation device 10 described in the embodiments of the present invention can refer to the working process of the in-vehicle audio recommendation method described in the above embodiments, and will not be elaborated here.

[0100] An in-vehicle audio recommendation device 10 provided by an embodiment of the present invention obtains current vehicle environment parameters and calculates a current vehicle environment comprehensive score; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed; according to the current vehicle environment comprehensive score, obtains the current audio type matching the current vehicle environment comprehensive score; obtains the face recognition information of the driver collected and the current user ID matching the face recognition information; uses the latent factor algorithm to calculate a number of first rating scores for each user for each audio in the current audio type, and selects a number of second rating scores of the current user ID for each audio in the current audio type from them, sorts each audio in the current audio type corresponding to the second rating scores in descending order of the second rating scores to obtain an audio stream, so that the in-vehicle device automatically starts playing the audio according to the audio stream, realizing the active recommendation of in-vehicle audio. Thus, it can be seen that the embodiment of the present invention combines the in-vehicle temperature, the number of passengers in the vehicle, and the vehicle speed to bring a better audio experience to the user in the current vehicle environment, uses the latent factor algorithm to provide the audio that meets the current user's preference to the user, and the embodiment of the present invention can actively recommend audio to the user, reduce the user's audio switching operation, reduce the boredom and anxiety of the user caused by listening to a single content, and further improve driving safety.

[0101] An embodiment of the present invention further provides a vehicle, including: a vehicle body and the in-vehicle audio recommendation device 10 described in the above embodiment.

[0102] Specifically, the working process of the in-vehicle audio recommendation device 10 can refer to the working process of the in-vehicle audio recommendation device 10 described in the above embodiment, and will not be elaborated here.

[0103] A vehicle provided by an embodiment of the present invention obtains current vehicle environment parameters and calculates a comprehensive score of the current vehicle environment. The current vehicle environment parameters include the current temperature inside the vehicle, the current number of passengers inside the vehicle, and the current vehicle speed. According to the comprehensive score of the current vehicle environment, the current audio type matching the comprehensive score of the current vehicle environment is obtained. The face recognition information of the driver collected and the current user ID matching the face recognition information are obtained. The latent factor algorithm is used to calculate a number of first score values of each user for each audio in the current audio type, and a number of second score values of the current user ID for each audio in the current audio type are selected from them. According to the order from high to low of the second score values, each audio in the current audio type corresponding to the second score values is sorted to obtain an audio stream, so that the in-vehicle device automatically starts playing the audio according to the audio stream, realizing the active recommendation of in-vehicle audio. It can be seen that the embodiment of the present invention combines the temperature inside the vehicle, the number of passengers inside the vehicle, and the vehicle speed to bring a better audio experience to the user in the current vehicle environment, uses the latent factor algorithm to provide the audio that meets the current user's preferences to the user. The embodiment of the present invention can actively recommend audio to the user, reduce the user's audio switching operation, reduce the boredom and anxiety of the user caused by listening to a single content, and thus improve driving safety.

[0104] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A vehicle-mounted audio recommendation method, characterized in that, Including: Obtain the current vehicle environment parameters, and calculate the current comprehensive vehicle environment score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current in-vehicle temperature, the current number of passengers in the vehicle, and the current vehicle speed; According to the current comprehensive vehicle environment score, query the current audio type that matches the current comprehensive vehicle environment score from the preset mapping relationship between the comprehensive vehicle environment score and the audio type; Obtain the face recognition information of the driver collected, and query the current user ID that matches the face recognition information from the preset user identity database; Use the latent factor algorithm to calculate a number of first score values for each audio in the current audio type for each user, and select a number of second score values for each audio in the current audio type by the current user ID. Arrange each audio in the current audio type corresponding to the second score value in descending order of the second score value to obtain an audio stream, so that the in-vehicle audio device automatically starts playing the audio according to the audio stream; Wherein, the calculating the current comprehensive vehicle environment score according to the current vehicle environment parameters includes: According to the current vehicle environment parameters, obtain the score corresponding to the current vehicle environment parameters; Calculate the current comprehensive vehicle environment score according to the following formula: Y = (0.5t + 0.3x + 0.2v)α In the formula, t is the score corresponding to the current in-vehicle temperature, x is the score corresponding to the current number of passengers in the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient; When the current in-vehicle temperature is less than 18°C, the corresponding score is 1; when the current in-vehicle temperature is greater than or equal to 18°C and less than 24°C, the corresponding score is 2; when the current in-vehicle temperature is greater than or equal to 24°C, the corresponding score is 3; When the current number of passengers in the vehicle is one, the corresponding score is 1; when the current number of passengers in the vehicle is two, the corresponding score is 2; when the current number of passengers in the vehicle is three, the corresponding score is 3; when the current number of passengers in the vehicle is four, the corresponding score is 4; when the current number of passengers in the vehicle is five, the corresponding score is 5; When the current vehicle speed is less than or equal to 60 Km / h, the corresponding score is 1; when the current vehicle speed is greater than 60 Km / h and less than 110 Km / h, the corresponding score is 2; when the current vehicle speed is greater than or equal to 110 Km / h, the corresponding score is 3.

2. The in-vehicle audio recommendation method according to claim 1, wherein The preset mapping relationship between the comprehensive vehicle environment score and the audio type includes: When the comprehensive vehicle environment score is greater than 0.5 and less than or equal to 1, the audio type is the first audio type; When the comprehensive vehicle environment score is greater than 0.2 and less than 0.5, the audio type is the second audio type; When the comprehensive vehicle environment score is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

3. The in-vehicle audio recommendation method according to claim 1, wherein The calculating, by using the latent factor algorithm, a number of first score values for each audio in the current audio type for each user includes: Obtain the actual rating score matrix R of each user for each audio in the current audio type; Using the UV decomposition of the matrix, decompose the actual score value matrix R into two low-dimensional matrices P and Q, and use the gradient descent method to solve the objective function Obtain the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type; where r ui represents the vector of the actual score values of user u for each audio i in the current audio type, q i represents the vector of the user audio type preferences of matrix P, p u represents the vector of the elements in each audio in the current audio type of matrix Q, and T represents the transpose; Multiply the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain a number of first rating scores of each user for each audio in the current audio type.

4. A vehicle-mounted audio recommendation device, characterized in that, Including: A current vehicle environment comprehensive score calculation module, configured to obtain current vehicle environment parameters and calculate a current vehicle environment comprehensive score according to the current vehicle environment parameters; wherein, the current vehicle environment parameters include: the current vehicle interior temperature, the current number of passengers in the vehicle, and the current vehicle speed; A current audio type acquisition module, configured to query, according to the current vehicle environment comprehensive score, a current audio type that matches the current vehicle environment comprehensive score from a preset mapping relationship between the vehicle environment comprehensive score and the audio type; A current user ID acquisition module, configured to obtain the face recognition information of the driver collected, and query a current user ID that matches the face recognition information from a preset user identity database; An audio recommendation module, configured to calculate a number of first rating scores of each user for each audio in the current audio type by using a latent factor algorithm, and select a number of second rating scores of the current user ID for each audio in the current audio type therefrom, and sort each audio in the current audio type corresponding to the second rating scores in descending order of the second rating scores to obtain an audio stream, so that the in-vehicle audio device automatically starts playing the audio according to the audio stream; Wherein, calculating the current vehicle environment comprehensive score according to the current vehicle environment parameters includes: Obtain the score corresponding to the current vehicle environment parameter according to the current vehicle environment parameter; Calculate the current vehicle environment comprehensive score according to the following formula: Y=(0.5t + 0.3x + 0.2v)α In the formula, t is the score corresponding to the current vehicle interior temperature, x is the score corresponding to the current number of passengers in the vehicle, v is the score corresponding to the current vehicle speed, and α is a preset influence coefficient; When the current vehicle interior temperature is less than 18°C, the corresponding score is 1, when the current vehicle interior temperature is greater than or equal to 18°C and less than 24°C, the corresponding score is 2, and when the current vehicle interior temperature is greater than or equal to 24°C, the corresponding score is 3; When the current number of passengers in the vehicle is one person, the corresponding score is 1, when the current number of passengers in the vehicle is two people, the corresponding score is 2, when the current number of passengers in the vehicle is three people, the corresponding score is 3, when the current number of passengers in the vehicle is four people, the corresponding score is 4, and when the current number of passengers in the vehicle is five people, the corresponding score is 5; When the current vehicle speed is less than or equal to 60 Km / h, the corresponding score is 1, when the current vehicle speed is greater than 60 Km / h and less than 110 Km / h, the corresponding score is 2, and when the current vehicle speed is greater than or equal to 110 Km / h, the corresponding score is 3.

5. The in-vehicle audio recommendation device according to claim 4, characterized in that, The preset mapping relationship between the vehicle environment comprehensive score and the audio type includes: When the comprehensive vehicle environment score is greater than 0.5 and less than or equal to 1, the audio type is the first audio type; When the comprehensive vehicle environment score is greater than 0.2 and less than 0.5, the audio type is the second audio type; When the comprehensive vehicle environment score is greater than 0.1 and less than or equal to 0.2, the audio type is the third audio type.

6. The in-vehicle audio recommendation device according to claim 4, wherein The calculation of a number of first scoring scores for each audio in the current audio type by each user using the latent factor algorithm includes: Obtain the actual scoring score matrix R of each audio in the current audio type for each user; Using the UV decomposition of the matrix, decompose the actual rating score matrix R into two low-dimensional matrices P and Q, and use the gradient descent method to solve the objective function Obtain the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type; where r ui represents the vector of the actual rating scores of user u for each audio i in the current audio type, q i represents the vector of the user audio type preferences of matrix P, p u represents the vector of the elements in each audio in the current audio type of matrix Q, and T represents the transpose; Multiply the user audio type preference matrix Q' and the element matrix P' of each audio in the current audio type to obtain a number of first scoring scores for each audio in the current audio type by each user.

7. A vehicle, characterized in that, Including: The vehicle body and the in-vehicle audio recommendation device according to claim 4 above.

Images