Vehicle audio control system and method

Abstract

The present invention relates to the technical field of control systems, and in particular to a vehicle-mounted audio control system and method, comprising a vehicle machine device and a plurality of sound pickup devices, and also comprising a preference recording unit, which sets a voice connection instruction, identifies the voice connection instruction of each passenger, obtains music preference data of a target passenger and records it as target preference data; a position confirmation unit, which identifies the voice connection instruction and obtains audio data collected by each sound pickup device, determines a preset seat corresponding to the voice connection instruction and records it as a target seat, and binds the target seat to the target passenger; a feature binding unit, which obtains a pressure value on the target seat and records it as a target pressure of the target passenger, identifies the voiceprint feature of the target passenger and binds it to the target passenger, and forms an identification feature set in combination with the target pressure; and an passenger identification unit, which determines fixed passengers included in the passengers in the vehicle and recommends songs based on the preference data of the fixed passengers in the vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of control systems, and in particular to a vehicle-mounted audio control system and method. Background Art

[0002] The in-vehicle audio control system is an intelligent audio management system integrated into the vehicle, designed to enhance the audio experience of drivers and passengers. It can control audio playback through a variety of interactive methods (such as voice control, touch, steering wheel buttons), manage and optimize music, navigation, phone calls, etc., and perform operations such as play, pause, adjust volume, switch tracks, etc. through voice commands, improving driving safety and reducing driver distraction.

[0003] The in-vehicle audio playback equipment in the prior art can only control music playback based on the music preferences of the driver or the account bound to the vehicle computer. However, a vehicle is generally used as a means of transportation for more than one person. When there are multiple passengers in the vehicle, it is difficult for the in-vehicle audio control system to recommend music tracks based on the common preferences of multiple passengers, resulting in a low degree of personalization. Summary of the invention

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention provides a vehicle-mounted audio control system and method, which can effectively solve the problem in the prior art that the vehicle-mounted audio control system is difficult to recommend music to be played based on the common preferences of multiple passengers.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0006] The present invention provides a vehicle audio control system, which at least includes: a vehicle machine device and a plurality of sound pickup devices, and further includes:

[0007] a preference recording unit, which sets a voice connection instruction, identifies the voice connection instruction of each passenger and establishes a wireless connection with the passenger's handheld terminal, records the passenger as a target passenger, obtains the music preference data of the target passenger and records it as target preference data;

[0008] The position confirmation unit recognizes the voice connection command and obtains the audio data collected by each sound pickup device, obtains the preset seat inside the vehicle, and performs spectrum analysis on the audio data of multiple sound pickup devices, wherein:

[0009] Play audio at different preset seats, measure the amplitudes of different sound pickup devices and sort them, construct a feature sequence of the preset seats based on the sort of the sound pickup devices, record the sort of the amplitude of the sound pickup devices corresponding to the recognized voice connection instruction as the target sequence, compare the similarity between the target sequence and the feature sequence, determine the preset seat corresponding to the voice connection instruction as the target seat, and bind the target seat to the target occupant;

[0010] a feature binding unit, which obtains the pressure value on the target seat and records it as the target pressure of the target occupant, identifies the voiceprint feature of the target occupant and binds it to the target voiceprint feature, records the target occupant corresponding to the same target voiceprint feature as a fixed occupant, and combines the most recently recorded target pressure to form an identification feature set of the fixed occupant, wherein the identification feature set includes the target voiceprint feature and the target pressure;

[0011] The passenger identification unit performs pressure recognition, voiceprint recognition, and position recognition on passengers in preset seats in the car in turn, determines the fixed passengers included in the passengers in the car, and recommends songs based on the preference data of the current fixed passengers in the car.

[0012] Furthermore, the process of determining the target seat by the position confirmation unit is as follows:

[0013] Mark multiple pickup devices as , where i is the serial number of the sound pickup device, i=1,2,3,…,j, j is the total number of sound pickup devices, obtain p preset seats on the vehicle, play the preset test audio at a fixed height H on the preset seats, obtain the average amplitude recorded by different sound pickup devices during the test audio playback time, and record them as the sound distance reference value , where q is the number of the preset seat, q=1,2,3,…,p, the reference value of the sound distance Indicates the pickup device when playing the test audio at a fixed height H at the preset coordinate q The average value of the recorded amplitude is constructed from the sound distance reference value The p*j matrix A is composed of, and the matrix A is expressed as , obtain each row element in matrix A and sort them in order according to the sound distance reference value to reconstruct P 1*j feature matrices , q is the row order of the feature matrix in the original matrix A, extract the feature matrix The subscripts of the elements in form the characteristic sequence ;

[0014] The time interval during which the voice connection instruction lasts is recorded as the analysis interval, the audio data captured by each pickup device in the analysis interval is obtained and the environmental noise is removed, the spectrum corresponding to the voice connection instruction is retained, and the frequency spectrum of each pickup device is calculated. The average amplitude in the spectrum is recorded as the target amplitude value , set the target amplitude value Sort by size and remove the subscript to get the target sequence ;

[0015] Calculate the target sequence With each characteristic series The similarity between ;

[0016] Get similarity The subscript corresponding to the maximum value is recorded as the target seat by taking the preset seat corresponding to the subscript;

[0017] Furthermore, the similarity The calculation formula is:

[0018] Set the characteristic sequence for , target sequence for ,but

[0019] ,

[0020] in:

[0021] Represents the target sequence With characteristic series The similarity between

[0022] j is the total number of pickup devices;

[0023] q is the number of the preset seat;

[0024] g represents the element number in the sequence;

[0025] Furthermore, the process of constructing the identification feature set by the feature binding unit is as follows:

[0026] The pressure value on the target seat is obtained and recorded as the target pressure of the target occupant, the voice of the target occupant is collected in real time and voiceprint analysis is performed to obtain the voiceprint feature of the target occupant and record it as the target voiceprint feature;

[0027] When the target passenger is marked for the first time, the voice data during the driving process is recorded, the voice data of the person in the target seat is recorded as the target voice, and the voiceprint features of the target voice are extracted and recorded as the target voiceprint features;

[0028] The target crew members are marked in the order in which they are recorded. , m is the target crew member's serial number, based on the target crew member The fixed occupant set is constructed based on the target voiceprint features, and the fixed occupant set includes multiple target occupants with different voiceprint features, and the target occupants in the fixed occupant set are marked as fixed occupants. , n is the serial number of the fixed crew member;

[0029] When the voiceprint feature of the most recently marked target occupant is the same as the voiceprint feature of any fixed occupant, the target pressure corresponding to the target occupant is replaced with the target pressure of the fixed occupant, and the target voiceprint feature and target pressure of the fixed occupant are obtained and combined into an identification feature set. Each identification feature set corresponds to a fixed occupant, and each fixed occupant is bound to its target preference data.

[0030] Furthermore, the pressure recognition process of the occupant recognition unit is as follows:

[0031] The target pressure corresponding to the fixed occupant is recorded as , n is the serial number of the fixed occupant, and a minimum pressure threshold is preset. When the pressure value on the preset seat is detected to be greater than or equal to the preset minimum pressure threshold, the preset seat status is marked as in use;

[0032] The pressure value of the preset seat in use is obtained and recorded as the verification pressure value, and the verification pressure value is compared with the target pressure of each fixed occupant. Compare one by one and calculate the ratio. When the ratio between the check pressure value and any one or more target pressures is within a preset accurate range, the corresponding one or more target pressures are recorded as the riding pressure, and the fixed occupant corresponding to the riding pressure is recorded as the first riding occupant. One or more first riding occupants corresponding to different check pressure values ​​are bound to the corresponding preset seats.

[0033] The target preference data corresponding to all first-passenger passengers are obtained to form a comprehensive preference database.

[0034] Furthermore, the voiceprint recognition process of the occupant recognition unit is as follows:

[0035] Fixed crew The corresponding target voiceprint feature is recorded as , continuously monitor the voice of people in the car during driving and extract voiceprint features to form a voiceprint set in the car. The voiceprint set in the car includes multiple independent voiceprints, and compares each independent voiceprint with the target voiceprint features of the fixed occupants. Compare, if there is any When matching with an independent voiceprint, the independent voiceprint is recorded as a riding voiceprint, a fixed passenger corresponding to the riding voiceprint is recorded as a second riding passenger, and the intersection of the first riding passenger set and the second riding passenger set is obtained and recorded as a first target intersection;

[0036] The target preference data of all fixed passengers in the first target intersection are obtained to form a comprehensive preference database.

[0037] Furthermore, the occupant identification unit position identification process is as follows:

[0038] When obtaining an independent voiceprint, determine the preset seat corresponding to the independent voiceprint and bind it to the independent voiceprint; the preset seat bound to the voiceprint is recorded as the first seat; each fixed occupant in the first target intersection corresponds to a first seat; the preset seat corresponding to the riding pressure is recorded as the second seat; each fixed occupant in the first target intersection corresponds to a second seat; when the first seat and the second seat of the same fixed occupant in the first target intersection are different, the fixed occupant is eliminated from the first target intersection, and based on the eliminated first target intersection, a common favorite playlist is regenerated and music playback is recommended.

[0039] The present invention also provides a vehicle audio control method, comprising the following steps:

[0040] Step 1: Identify the passenger's voice connection command and mark it as the target passenger, obtain the target passenger's target preference data, analyze the target passenger's voice connection command, determine the preset seat corresponding to the target passenger and record it as the target seat, extract the target passenger's voiceprint features during driving and bind it to the target passenger;

[0041] Step 2: Obtain the pressure value on the target seat as the target pressure of the target occupant, record the target occupant corresponding to the same voiceprint feature as a fixed occupant, combine the most recently obtained target pressure of the same fixed occupant with its voiceprint feature to form an identification feature set, each identification feature set corresponds to a fixed occupant, and each fixed occupant is bound to its target preference data;

[0042] Step 3: Obtain the pressure values, voiceprint features and seating positions of different passengers in the car, perform weight recognition, voiceprint recognition, and position recognition in sequence, and determine the fixed passengers included in the passengers in the car;

[0043] Step 4: The fixed passengers included in the passengers in the car are grouped into the first target intersection, the target preference data of all the fixed passengers in the first target intersection are obtained and formed into a comprehensive preference database, a common preference playlist is generated based on the comprehensive preference database analysis, and music is recommended for playback based on the common preference playlist.

[0044] Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects:

[0045] On the one hand, the present invention can obtain the music preference data of the target occupant and at the same time construct an identification feature set for each target occupant, so as to distinguish each target occupant and identify them to obtain multiple fixed occupants and their preference data. On the other hand, the present invention can identify the fixed occupants among the passengers in the vehicle based on the feature data in the identification feature set, and then automatically extract the preference data of all fixed occupants in the vehicle for comprehensive analysis without the need for the fixed occupants to perform any operations, and then make music playback recommendations. On the basis of confirming the identity of the occupant, a comprehensive preference database is generated in combination with their historical preference data, so as to dynamically generate a common preference playlist. This method not only improves the relevance of the recommendation, but also can balance the music preferences of different occupants and realize personalized recommendations. The entire process automatically generates a recommended playlist that meets the preferences of the current occupants in the vehicle through multiple data analyses, reducing user manual operations and realizing an efficient and seamless automatic music recommendation experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 It is the overall module block diagram of the present invention;

[0048] Figure 2 It is a schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] The present invention will be further described below in conjunction with the embodiments.

[0051] See also Figure 1 , a vehicle audio control system, comprising at least: a vehicle machine device with a wireless data transmission function and a plurality of sound pickup devices arranged inside the vehicle, the vehicle machine device can establish a wireless connection with a handheld terminal of an occupant (i.e., a passenger) and obtain user data on the handheld terminal, the sound pickup device is used to monitor and analyze the audio inside the vehicle, and also includes:

[0052] The preference recording unit sets a voice connection instruction, which consists of a paragraph of text, resets the Bluetooth connection state by identifying the voice connection instruction of the passenger, sends a Bluetooth connection request to the surrounding devices and sets a connection key, obtains the connection response and connection key (the connection key and the connection key are the same character string) of each surrounding mobile device, and when the connection key matches the connection key successfully, records the corresponding mobile device as the target connection device, establishes a Bluetooth connection between the target connection device and the vehicle machine device, records the passenger corresponding to the voice connection instruction as the target passenger, obtains the user music preference data stored on the target connection device and records it as the target preference data of the target passenger;

[0053] For example, in a specific embodiment, the voice connection command is set to "Connect my Bluetooth". When the user is recognized to say the voice connection command, the Bluetooth currently connected to the vehicle device is disconnected and re-matched. The user establishes a Bluetooth connection with the vehicle device through a handheld terminal. After that, the music preference data on the user's mobile terminal is obtained through Bluetooth data transmission.

[0054] It should be noted that the recognition of voice commands is a prior art and will not be described in detail here.

[0055] The position confirmation unit obtains the audio data collected by each sound pickup device while recognizing the voice connection instruction, performs spectrum analysis on the audio data of multiple sound pickup devices, and determines the occurrence position of the voice connection instruction. The analysis process is as follows:

[0056] Mark multiple pickup devices as , where i is the serial number of the sound pickup device, i=1,2,3,…,j, j is the total number of sound pickup devices, obtain p preset seats on the vehicle, play the preset test audio at a fixed height H (in a specific embodiment, H is 1.2 meters) on the preset seats (without the influence of other noises), obtain the average amplitudes recorded by different sound pickup devices during the test audio playback time and record them as the sound distance reference values , where q is the number of the preset seat, q=1,2,3,…,p, the reference value of the sound distance Indicates the pickup device when playing the test audio at a fixed height H at the preset coordinate q The average value of the recorded amplitude is constructed from the sound distance reference value The p*j matrix A is composed of, and the matrix A is expressed as , obtain each row element in matrix A and sort them in order according to the sound distance reference value to reconstruct P 1*j feature matrices , q is the row order of the feature matrix in the original matrix A (it also represents the number of the preset seats, and the two do not conflict), for example, are the elements of the first row in matrix A Rearrange the feature matrix in order of size and extract the feature matrix The subscripts of the elements in form the characteristic sequence , for example, when the feature matrix Expressed as When the characteristic sequence for , that is, each preset seat corresponds to a feature number series, through which the sound source position of the audio can be quickly determined, so as to determine the seat location of the passenger when the passenger speaks the voice connection command;

[0057] The time interval during which the voice connection instruction lasts is recorded as the analysis interval, the audio data captured by each pickup device in the analysis interval is obtained and the environmental noise is removed, the spectrum corresponding to the voice connection instruction is retained, and the frequency spectrum of each pickup device is calculated. The average amplitude in the spectrum is recorded as the target amplitude value , set the target amplitude value Sort by size and remove the subscript to get the target sequence ;

[0058] It should be noted that spectral denoising can be achieved by adopting the spectral human voice optimization method in the existing technology (such as spectral subtraction denoising, adjusting the spectral gain), extracting the noise spectrum in the non-speech part, and using the noise in the first few frames of the audio signal or the background silence. The noise spectrum is subtracted in the time-frequency domain, the noise component is filtered out, and only the main speech spectrum is retained to obtain a pure human voice spectrum. I will not go into details here.

[0059] Calculate the target sequence With each characteristic series The similarity between , similarity The calculation formula is:

[0060] Set the characteristic sequence for , target sequence for ,but

[0061] ,

[0062] in:

[0063] Represents the target sequence With characteristic series The similarity between

[0064] j is the total number of pickup devices;

[0065] q is the number of the preset seat;

[0066] g represents the element number in the sequence;

[0067] Get similarity The subscript corresponding to the maximum value is recorded as the target seat, with the preset seat corresponding to the subscript being the target seat.

[0068] By comparing and analyzing the amplitude performance of the voice connection instruction in different sound pickup devices, the amplitude sorting series (i.e., the target series) of each current sound pickup device is obtained, and the series is compared with the characteristic series obtained in the previous step. The element feature similarity between the series and each characteristic series can be calculated through the similarity algorithm. The greater the similarity, the more similar the two series are, so as to determine the characteristic series that is most similar to the current amplitude sorting series, and then determine the output position of the voice connection instruction. In general, the greater the similarity of the amplitude sorting series formed based on the amplitude performance of the sound in different sound pickup devices for sounds emitted at the same or similar positions, the greater the similarity. Therefore, the output position of the voice connection instruction can be inferred by the similarity with the characteristic series.

[0069] It should be noted that in order to ensure that the voice connection instruction can be recognized and the duration and time interval of the voice connection instruction can be recorded, the sound pickup device can be set to run continuously, so as to quickly determine the analysis interval after the voice connection instruction is recognized.

[0070] The feature binding unit identifies the voiceprint feature of the target occupant and binds it to the target occupant. The target occupant corresponding to the same voiceprint feature is recorded as a fixed occupant, and the identification feature set of the fixed occupant is formed in combination with the most recently recorded target pressure, where:

[0071] The pressure value on the target seat is obtained and recorded as the target pressure of the target occupant (the pressure value can be measured by a pressure sensor arranged under the seat), the voice of the target occupant is collected in real time and a voiceprint analysis is performed to obtain the voiceprint feature of the target occupant and record it as the target voiceprint feature;

[0072] The target voiceprint feature collection process is as follows:

[0073] When the target occupant is marked for the first time, the voice data of the driving process (a driving process refers to the process from the navigation starting point to the navigation end point) is recorded, the voice data of the target seat is recorded as the target voice, the spectrum graph corresponding to the target voice is drawn and recorded as the target spectrum graph, and the voiceprint features in the target spectrum graph identified based on voiceprint recognition technology are recorded as target voiceprint features.

[0074] It should be noted that common voiceprint feature parameters in the prior art include Mel-frequency cepstral coefficients (MFCC), log-Mel filter bank coefficients (Log-Mel Filter Bank), linear prediction coefficients (LPC), etc. Obtaining specific parameters and analyzing them can identify the voiceprint features of different people.

[0075] The target crew members are marked in the order in which they are recorded. , m is the target crew member's serial number, based on the target crew member The fixed occupant set is constructed based on the target voiceprint features, and the fixed occupant set includes multiple target occupants with different voiceprint features, and the target occupants in the fixed occupant set are marked as fixed occupants. , n is the serial number of the fixed occupant, and the maximum value of n is less than or equal to the maximum value of m. That is to say, when the voiceprint features of two target occupants are the same, they are recorded as the same fixed occupant, thereby avoiding the same occupant being repeatedly recorded in the fixed occupant set and reducing the amount of data storage;

[0076] When the voiceprint feature of the target occupant marked most recently is the same as the voiceprint feature of any fixed occupant, the target pressure corresponding to the target occupant replaces the target pressure of the fixed occupant. That is to say, no matter how many times the same fixed occupant is marked as a target occupant, the target pressure value at the time of the most recent marking is used as its target pressure. The target voiceprint feature and target pressure of the fixed occupant are obtained and combined into an identification feature set. Each identification feature set corresponds to a fixed occupant, so that there is a mutual mapping between the fixed occupant and the identification feature set. The fixed occupant can be confirmed by the features in the identification feature set, and each fixed occupant is bound to his or her target preference data.

[0077] The passenger identification unit performs pressure recognition, voiceprint recognition, and position recognition on passengers in preset seats in the car in turn, thereby determining the fixed passengers included in the passengers in the car, and then controls the music playback of the vehicle equipment based on the preference data of the current fixed passengers in the car, and recommends songs that meet the preferences of the fixed passengers in the car, including:

[0078] S1: Fixed the occupant The corresponding target pressure is recorded as ( is a specific value), a minimum pressure threshold is preset, and when it is detected that the pressure value on the preset seat is greater than or equal to the preset minimum pressure threshold, the preset seat status is marked as in use, the pressure value on the preset seat in use is obtained and recorded as the verification pressure value, and the verification pressure value is compared with the target pressure of each fixed occupant. Perform one-by-one comparison and calculate the ratio (i.e., the check pressure value divided by the target pressure value). When the ratio between the check pressure value and any one or more target pressures is within a preset accurate range, the corresponding one or more target pressures are recorded as the riding pressure, and the fixed occupant corresponding to the riding pressure is recorded as the first riding occupant. One or more first riding occupants corresponding to different check pressure values ​​are bound to the corresponding preset seats.

[0079] It should be noted that, for the convenience of understanding by those skilled in the art, the above steps can be summarized as: determining the pressure on different preset seats and comparing them with multiple target pressures; when similar pressures exist, identifying the person on the preset seat as a fixed occupant corresponding to the similar target pressures, that is, determining whether the person on the preset seat is a fixed occupant through pressure comparison, and identifying which specific fixed occupant it may be, thereby recommending music to be played based on the preference data of the fixed occupants, in line with the music preferences of all fixed occupants in the car.

[0080] Obtain the target preference data corresponding to all first-passenger passengers to form a comprehensive preference database, generate a common preference song list based on the comprehensive preference database analysis, and recommend music playback based on the common preference song list:

[0081] It should be noted that the prior art includes a variety of methods for generating a common preference playlist based on the music preference data of multiple users. For example, the clustering-based method mainly uses a clustering algorithm (such as K-means, hierarchical clustering) to divide each person's playlist into several groups (such as different styles, emotional classifications), and then takes the center of each group (or the most frequent songs) to form a common playlist. The most representative songs in each category can be selected to construct a playlist based on the similarity between songs and the clustering results. I will not go into details here.

[0082] S2: Fixed the occupants The corresponding target voiceprint feature is recorded as , continuously monitor the voices in the car during driving and extract voiceprint features to form a voiceprint set in the car (the voiceprint set in the car contains the voiceprint features of all passengers who have spoken during driving). The voiceprint set in the car includes multiple independent voiceprints (each independent voiceprint corresponds to a passenger), and each independent voiceprint is compared with the target voiceprint features of the fixed passenger. Compare, if there is any When matching with an independent voiceprint, the independent voiceprint is recorded as a riding voiceprint, a fixed passenger corresponding to the riding voiceprint is recorded as a second riding passenger, and the intersection of the first riding passenger set and the second riding passenger set is obtained and recorded as a first target intersection;

[0083] By adding voiceprint verification, fixed passengers among the passengers can be identified more accurately, and then a playlist can be generated and played based on the preferences of the fixed passengers, so that the played music can be more in line with the preferences of all fixed passengers in the vehicle, thereby improving the experience.

[0084] Obtain the target preference data of all fixed passengers in the first target intersection to form a comprehensive preference database, generate a common preference song list based on the comprehensive preference database analysis, and recommend music playback based on the common preference song list:

[0085] S3: while obtaining the independent voiceprint, determine the preset seat corresponding to the independent voiceprint and bind it to the independent voiceprint (the process of determining the corresponding preset seat based on the independent voiceprint refers to the analysis step in the position confirmation unit), record the preset seat bound to the voiceprint as the first seat, and each fixed occupant in the first target intersection corresponds to a first seat. Obtain the preset seat corresponding to the seating pressure as the second seat, and each fixed occupant in the first target intersection corresponds to a second seat. When the first seat and the second seat of the same fixed occupant in the first target intersection are different, the fixed occupant is removed from the first target intersection, and based on the first target intersection after removal (the fixed occupant preference data in it), a new list of common preferences is generated and music is recommended for playback.

[0086] It should be noted that through further verification and confirmation, fixed passengers who are misjudged during the voiceprint recognition and pressure recognition process can be effectively eliminated, thereby ensuring that the final recommended playlist is generated based on all correct fixed passenger preference data, and try to avoid substituting preference data of non-current passengers in the car, thereby greatly improving the targeted nature of music recommendations and enhancing the listening experience of fixed passengers in the car.

[0087] See also Figure 2 , a vehicle audio control method, comprising the following steps:

[0088] Step 1: Identify the passenger's voice connection command and mark it as the target passenger, obtain the target passenger's target preference data, analyze the target passenger's voice connection command, determine the preset seat corresponding to the target passenger and record it as the target seat, extract the target passenger's voiceprint features during driving and bind it to the target passenger;

[0089] The target seat is determined by analyzing the voice connection instructions of the target occupant, and the characteristics of the target occupant can be further analyzed based on other data corresponding to the target seat, thereby assigning a set of identification features for individual identification to each unique target occupant.

[0090] Step 2: Obtain the pressure value on the target seat as the target pressure of the target occupant, record the target occupant corresponding to the same voiceprint feature as a fixed occupant, combine the most recently obtained target pressure of the same fixed occupant with its voiceprint feature to form an identification feature set, each identification feature set corresponds to a fixed occupant, and each fixed occupant is bound to its target preference data;

[0091] Step 3: Obtain the pressure values, voiceprint features and seating positions of different passengers in the car, perform weight recognition, voiceprint recognition, and position recognition in sequence, and determine the fixed passengers included in the passengers in the car;

[0092] Step 4: The fixed passengers included in the passengers in the car are grouped into the first target intersection, the target preference data of all the fixed passengers in the first target intersection are obtained and formed into a comprehensive preference database, a common preference playlist is generated based on the comprehensive preference database analysis, and music is recommended for playback based on the common preference playlist.

[0093] It should be noted that, on the one hand, the above method can obtain the music preference data of the target occupant and at the same time construct an identification feature set for each target occupant, so as to distinguish each target occupant and identify them to obtain multiple fixed occupants and their preference data. On the other hand, the fixed occupants among the passengers in the vehicle can be identified based on the feature data in the identification feature set, and then, without the fixed occupants having to perform any operations, the preference data of all fixed occupants in the vehicle are automatically extracted for comprehensive analysis, and then music playback recommendations are made. On the basis of confirming the identity of the occupant, a "comprehensive preference database" is generated in combination with their historical preference data, so as to dynamically generate a common preference playlist. This method not only improves the relevance of the recommendation, but also balances the music preferences of different occupants and realizes personalized recommendations. The entire process automatically generates a recommended playlist that meets the preferences of the current occupants in the vehicle through multiple data analyses, reduces user manual operations, and realizes an efficient and seamless automatic music recommendation experience.

[0094] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle audio control system, comprising a vehicle device and a plurality of sound pickup devices, characterized in that: Also includes: a preference recording unit, which sets a voice connection instruction, identifies the voice connection instruction of each passenger and establishes a wireless connection with the passenger's handheld terminal, records the passenger as a target passenger, obtains the music preference data of the target passenger and records it as target preference data; The position confirmation unit obtains the audio data collected by each sound pickup device when recognizing the voice connection instruction, obtains the preset seat inside the vehicle, and performs spectrum analysis on the audio data of multiple sound pickup devices, wherein: Play audio at different preset seats, measure the amplitudes of different sound pickup devices and sort them, construct a feature sequence of the preset seats based on the sort of the sound pickup devices, record the sort of the amplitude of the sound pickup devices corresponding to the recognized voice connection instruction as the target sequence, compare the similarity between the target sequence and the feature sequence, determine the preset seat corresponding to the voice connection instruction as the target seat, and bind the target seat to the target occupant; The process of determining the target seat by the position confirmation unit is as follows: Mark multiple pickup devices as , where i is the serial number of the sound pickup device, i=1,2,3,…,j, j is the total number of sound pickup devices, obtain p preset seats on the vehicle, play the preset test audio at a fixed height H on the preset seats, obtain the average amplitude recorded by different sound pickup devices during the test audio playback time, and record them as the sound distance reference value , where q is the number of the preset seat, q=1,2,3,…,p, the reference value of the sound distance Indicates the pickup device when playing the test audio at a fixed height H at the preset coordinate q The average value of the recorded amplitude is constructed from the sound distance reference value The p*j matrix A is composed of, and the matrix A is expressed as , obtain each row element in matrix A and sort them in order according to the sound distance reference value to reconstruct P 1*j feature matrices , q is the row order of the feature matrix in the original matrix A, extract the feature matrix The subscripts of the elements in form the characteristic sequence ; The time interval during which the voice connection instruction lasts is recorded as the analysis interval, the audio data captured by each pickup device in the analysis interval is obtained and the environmental noise is removed, the spectrum corresponding to the voice connection instruction is retained, and the frequency spectrum of each pickup device is calculated. The average amplitude in the spectrum is recorded as the target amplitude value , set the target amplitude value Sort by size and remove the subscript to get the target sequence ; Calculate the target sequence With each characteristic series Similarity between ; Get similarity The subscript corresponding to the maximum value is recorded as the target seat by taking the preset seat corresponding to the subscript; Similarity The calculation formula is: Set the characteristic sequence for , target sequence for ,but ， in: Represents the target sequence With characteristic series The similarity between j is the total number of pickup devices; q is the number of the preset seat; g represents the element number in the sequence; a feature binding unit, which obtains the pressure value on the target seat and records it as the target pressure of the target occupant, identifies the voiceprint feature of the target occupant and binds it to the target voiceprint feature, records the target occupant corresponding to the same target voiceprint feature as a fixed occupant, and combines the most recently recorded target pressure to form an identification feature set of the fixed occupant, wherein the identification feature set includes the target voiceprint feature and the target pressure; The passenger identification unit performs pressure recognition, voiceprint recognition, and position recognition on passengers in preset seats in the car in turn, determines the fixed passengers included in the passengers in the car, and recommends songs based on the preference data of the current fixed passengers in the car.

2. The vehicle audio control system according to claim 1, characterized in that: The process of constructing the identification feature set by the feature binding unit is as follows: The pressure value on the target seat is obtained and recorded as the target pressure of the target occupant, the voice of the target occupant is collected in real time and voiceprint analysis is performed to obtain the voiceprint feature of the target occupant and record it as the target voiceprint feature; When the target passenger is marked for the first time, the voice data during the driving process is recorded, the voice data of the person in the target seat is recorded as the target voice, and the voiceprint features of the target voice are extracted and recorded as the target voiceprint features; The target crew members are marked in the order in which they are recorded. , m is the target crew member's serial number, based on the target crew member The fixed occupant set is constructed based on the target voiceprint features, and the fixed occupant set includes multiple target occupants with different voiceprint features, and the target occupants in the fixed occupant set are marked as fixed occupants. , n is the serial number of the fixed crew member; When the voiceprint feature of the most recently marked target occupant is the same as the voiceprint feature of any fixed occupant, the target pressure corresponding to the target occupant is replaced with the target pressure of the fixed occupant, and the target voiceprint feature and target pressure of the fixed occupant are obtained and combined into an identification feature set. Each identification feature set corresponds to a fixed occupant, and each fixed occupant is bound to its target preference data.

3. The vehicle audio control system according to claim 1, characterized in that: The pressure recognition process of the occupant recognition unit is as follows: The target pressure corresponding to the fixed occupant is recorded as , n is the serial number of the fixed occupant, and a minimum pressure threshold is preset. When the pressure value on the preset seat is detected to be greater than or equal to the preset minimum pressure threshold, the preset seat status is marked as in use; The pressure value of the preset seat in use is obtained and recorded as the verification pressure value, and the verification pressure value is compared with the target pressure of each fixed occupant. Compare one by one and calculate the ratio. When the ratio between the check pressure value and any one or more target pressures is within a preset accurate range, the corresponding one or more target pressures are recorded as the riding pressure, and the fixed occupant corresponding to the riding pressure is recorded as the first riding occupant. One or more first riding occupants corresponding to different check pressure values ​​are bound to the corresponding preset seats. The target preference data corresponding to all first-passenger passengers are obtained to form a comprehensive preference database.

4. The vehicle audio control system according to claim 3, characterized in that: The occupant recognition unit voiceprint recognition process is as follows: Fixed crew The corresponding target voiceprint feature is recorded as , continuously monitor the voice of people in the car during driving and extract voiceprint features to form a voiceprint set in the car. The voiceprint set in the car includes multiple independent voiceprints, and compares each independent voiceprint with the target voiceprint features of the fixed occupants. Compare, if there is any When matching with an independent voiceprint, the independent voiceprint is recorded as a riding voiceprint, a fixed passenger corresponding to the riding voiceprint is recorded as a second riding passenger, and the intersection of the first riding passenger set and the second riding passenger set is obtained and recorded as a first target intersection; The target preference data of all fixed passengers in the first target intersection are obtained to form a comprehensive preference database.

5. The vehicle audio control system according to claim 4, characterized in that: The occupant identification unit position identification process is as follows: When obtaining an independent voiceprint, determine the preset seat corresponding to the independent voiceprint and bind it to the independent voiceprint; the preset seat bound to the voiceprint is recorded as the first seat; each fixed occupant in the first target intersection corresponds to a first seat; the preset seat corresponding to the riding pressure is recorded as the second seat; each fixed occupant in the first target intersection corresponds to a second seat; when the first seat and the second seat of the same fixed occupant in the first target intersection are different, the fixed occupant is eliminated from the first target intersection, and based on the eliminated first target intersection, a common favorite playlist is regenerated and music playback is recommended.

6. A vehicle audio control method, applied to a vehicle audio control system according to any one of claims 1 to 5, characterized in that: The following steps are involved: Step 1: Identify the passenger's voice connection command and mark it as the target passenger, obtain the target passenger's target preference data, analyze the target passenger's voice connection command, determine the preset seat corresponding to the target passenger and record it as the target seat, extract the target passenger's voiceprint features during driving and bind it to the target passenger; Step 2: Obtain the pressure value on the target seat as the target pressure of the target occupant, record the target occupant corresponding to the same voiceprint feature as a fixed occupant, combine the most recently obtained target pressure of the same fixed occupant with its voiceprint feature to form an identification feature set, each identification feature set corresponds to a fixed occupant, and each fixed occupant is bound to its target preference data; Step 3: Obtain the pressure values, voiceprint features and seating positions of different passengers in the car, perform weight recognition, voiceprint recognition, and position recognition in sequence, and determine the fixed passengers included in the passengers in the car; Step 4: The fixed passengers included in the passengers in the car are grouped into the first target intersection, the target preference data of all the fixed passengers in the first target intersection are obtained and formed into a comprehensive preference database, a common preference playlist is generated based on the comprehensive preference database analysis, and music is recommended for playback based on the common preference playlist.

Images