Vehicle voice data processing method, voice control system and vehicle
By applying a combination technology of neural network model and adaptive filters in the vehicle, the environmental noise and echo are eliminated, and the problem of difficulty in communicating in the vehicle's personnel in the noise scene is solved, achieving a clearer and more convenient conversation environment.
Patent Information
- Application Number
- CN202311066571.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-23
- Publication Date
- 2025-02-25
AI Technical Summary
In vehicles, especially in high-speed driving or in scenarios with high noise, it is difficult for people in the vehicle to communicate clearly, and the prior art is difficult to effectively eliminate environmental noise and echo, affecting the comfort and convenience of the conversation environment.
The neural network model is used to remove the ambient noise in the initial voice data, and the echo path in the vehicle is simulated through an adaptive filter, the basic multimedia voice data is processed to remove the echo characteristics, and finally the noise-removing data and the echo-removing data are eliminated to obtain clean vocal data, and the appropriate speaker is selected according to the voice playback scene for playback.
It effectively eliminates environmental noise and echo, significantly improves the conversation environment quality of the people in the car, and provides a more comfortable and convenient communication experience.
Smart Images
Figure CN119517021A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of data processing, and in particular to a method for processing vehicle voice data, a voice control system and a vehicle. Background Art
[0002] With the development of the automobile industry, people's feelings when riding in a car have attracted more and more attention.
[0003] In the related art, in a scene with relatively high noise, such as when a vehicle is driving at high speed, in order to make the passengers in the vehicle hear more clearly and feel better, the noise is usually removed after the sound is picked up by a microphone. Alternatively, only the microphone and the speaker are used as the sound receiving and playing devices, which cannot better improve the feeling of the passengers in the vehicle.
[0004] Therefore, it is important to provide a more comfortable and convenient conversation environment for people in the car. Summary of the invention
[0005] In an exemplary embodiment of the present application, a method for processing vehicle voice data, a voice control system and a vehicle are provided to provide a more comfortable and convenient conversation environment for occupants of the vehicle.
[0006] According to a first aspect of an exemplary embodiment, a method for processing vehicle voice data is provided, which is applied to a voice control system of a vehicle, wherein at least one microphone and at least one speaker are provided in the vehicle, and the method comprises:
[0007] Acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone;
[0008] Applying a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data;
[0009] Applying pre-set adaptive filters to simulate the echo path inside the vehicle;
[0010] Processing the basic multimedia voice data according to the parameters of the echo path to obtain second voice data;
[0011] Applying the second voice data to perform elimination processing on the first voice data to obtain target human voice data;
[0012] According to the current voice playback scenario, a target speaker is selected from at least one speaker, and the target speaker is used to play the target voice data.
[0013] According to a second aspect of an exemplary embodiment, there is provided a voice control system for a vehicle, wherein at least one microphone and at least one speaker are arranged in the vehicle, and the voice control system includes a processor and a data transmission unit;
[0014] The data transfer unit is configured to perform:
[0015] Acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone;
[0016] The processor is configured to execute:
[0017] Applying a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data;
[0018] Applying pre-set adaptive filters to simulate the echo path inside the vehicle;
[0019] Processing the basic multimedia voice data according to the parameters of the echo path to obtain second voice data;
[0020] Applying the second voice data to perform elimination processing on the first voice data to obtain target human voice data;
[0021] According to the current voice playback scenario, a target speaker is selected from at least one speaker, and the target speaker is used to play the target voice data.
[0022] According to a third aspect of an exemplary embodiment, a vehicle voice data processing device is provided, which is applied to a vehicle voice control system, wherein at least one microphone and at least one speaker are provided in the vehicle, and the device comprises:
[0023] A data acquisition unit, used to: acquire basic multimedia data in the vehicle, and initial voice data acquired by at least one microphone;
[0024] A data processing unit, configured to: apply a preset neural network model to remove environmental noise from the initial voice data to obtain first voice data;
[0025] The data processing unit is further used to: apply a pre-set adaptive filter to simulate an echo path in the vehicle;
[0026] The data processing unit is further used to: process the basic multimedia voice data according to the parameters of the echo path to obtain second voice data;
[0027] The data processing unit is further used to: apply the second voice data to eliminate the first voice data to obtain target human voice data;
[0028] The playing unit is used to: select a target speaker from at least one speaker according to a current voice playing scenario, and use the target speaker to play the target human voice data.
[0029] According to a fourth aspect of the exemplary embodiment, there is provided a vehicle comprising the voice control system of the second aspect.
[0030] According to a fifth aspect of an exemplary implementation, a computer storage medium is provided, in which computer program instructions are stored. When the instructions are executed on a computer, the computer executes the method for processing vehicle voice data as described in the first aspect.
[0031] In the embodiment of the present application, for the initial voice data collected by the microphone, a neural network model is first used to remove environmental noise; in order to remove multimedia noise, it is necessary to first obtain multimedia voice data that has been propagated and turned over. At this time, an adaptive filter is applied to obtain the parameters of the echo path, and the parameters are applied to perform echo processing on the basic multimedia data to obtain multimedia voice data with echo features added. Finally, the multimedia voice data with echo features added is eliminated from the data with environmental noise removed to obtain clean target voice data. In the actual application process, the corresponding speaker can also be selected to play clean target voice data according to whether the current voice playback scene is a communication scene or a privacy scene. Such a design can provide a more comfortable and convenient conversation environment for people in the car. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0033] Figure 1 A schematic diagram exemplarily shows the distribution of a microphone and a speaker provided in an embodiment of the present application;
[0034] Figure 2 A flowchart of a method for processing vehicle voice data provided by an embodiment of the present application is exemplarily shown;
[0035] Figure 3 A schematic diagram of generating a training sample provided by an embodiment of the present application is exemplarily shown;
[0036] Figure 4 A schematic diagram exemplarily shows a processing process of an adaptive algorithm provided in an embodiment of the present application;
[0037] Figure 5a A schematic diagram exemplarily shows a playback process in different scenarios provided by an embodiment of the present application;
[0038] Figure 5bA schematic diagram exemplarily shows an installation position of a smart terminal provided in an embodiment of the present application;
[0039] Figure 6 A schematic diagram of a display interface of a smart terminal provided in an embodiment of the present application is exemplarily shown;
[0040] Figure 7 A flowchart of a complete vehicle voice data processing method provided by an embodiment of the present application is exemplarily shown;
[0041] Figure 8 A schematic diagram exemplarily shows a process of processing vehicle voice data provided by an embodiment of the present application;
[0042] Fig. 9 The following is a schematic diagram showing the structure of a vehicle voice data processing device provided by an embodiment of the present application;
[0043] Fig.10 A structural schematic diagram of a vehicle voice control system provided in an embodiment of the present application is exemplarily shown. DETAILED DESCRIPTION
[0044] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.
[0045] Glossary:
[0046] (1) Acoustic echo: This refers to the sound broadcast by the loudspeaker being picked up by the microphone through multiple paths while being heard by the receiver. The result of multipath reflection produces echoes with different delays, including direct echoes and indirect echoes.
[0047] (2) Direct echo: refers to the sound broadcast by the speaker directly entering the microphone without any reflection. This type of echo has the shortest delay and is directly related to factors such as the voice energy of the far-end speaker, the distance and angle between the speaker and the microphone, the playback volume of the speaker, and the pickup sensitivity of the microphone.
[0048] (3) Indirect echo: refers to the collection of echoes generated by the sound broadcast by the speaker after one or more reflections from different paths (such as the house or any object in the house) and entering the microphone. Any change of any object in the house will change the channel of the echo. Therefore, the characteristics of this echo are multi-path and time-varying.
[0049] With the development of the automobile industry, people's feelings when riding in a car have attracted more and more attention.
[0050] In the related art, in a scene with relatively high noise, such as when a vehicle is driving at high speed, in order to make the passengers in the vehicle hear more clearly and feel better, the noise is usually removed after the sound is picked up by a microphone. Alternatively, only the microphone and the speaker are used as the sound receiving and playing devices, which cannot better improve the feeling of the passengers in the vehicle.
[0051] Therefore, it is important to provide a more comfortable and convenient conversation environment for people in the car.
[0052] To this end, the embodiment of the present application provides a method for processing vehicle voice data, which is applied to a vehicle voice control system, wherein at least one microphone and at least one speaker are provided in the vehicle. For example, a microphone and a speaker are provided at each seat headrest, where the speaker refers to a separately configured speaker other than the built-in speaker of the vehicle. Figure 1 A schematic diagram showing the distribution of microphones and speakers, Figure 1 Taking the main driver's seat as an example, the other seats can be set up in the same way. In order to reduce interference from the neighboring seat, the microphone can be placed at a relatively far end, such as the headrest microphone position shown by the circular icon, and the rectangular icon corresponds to the position of the headrest speaker. In order to achieve a better effect, the headrest speaker can use an active directional speaker and correct the direction of the speaker as needed. This is just an example and does not constitute a specific limitation.
[0053] In the embodiment of the present application, the neural network model is first used to remove the environmental noise, and then the parameters of the echo path obtained by the adaptive filter are applied to the basic multimedia data for echo processing to obtain multimedia voice data with echo features added, and finally, the data with environmental noise removed is removed from the multimedia data with echo features to obtain clean human voice data. In the actual application process, the corresponding speaker can also be selected to play clean human voice data according to whether the current voice playback scene is a communication scene or a privacy scene. Such a design can provide a more comfortable and convenient conversation environment for people in the car.
[0054] After introducing the design ideas of the embodiments of the present application, the following briefly introduces the application scenarios to which the technical solutions of the embodiments of the present application can be applied. It should be noted that the application scenarios introduced below are only used to illustrate the embodiments of the present application and are not limited. In specific implementation, the technical solutions provided by the embodiments of the present application can be flexibly applied according to actual needs.
[0055] To further illustrate the technical solution provided by the embodiment of the present application, this is described in detail below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiment of the present application provides the method operation steps shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided in the embodiment of the present application.
[0056] Combine the following Figure 1 The application scenario shown is shown in Figure 2 A method for processing vehicle voice data is shown, which can be applied to a vehicle voice control system. At least one microphone and at least one speaker are provided in the vehicle. The method comprises at least the following steps:
[0057] S201: Acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone.
[0058] S202: Apply a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data.
[0059] S203: Apply a preset adaptive filter to simulate the echo path in the vehicle.
[0060] S204: Process the basic multimedia voice data according to the parameters of the echo path to obtain second voice data.
[0061] S205: Apply the second voice data to perform elimination processing on the first voice data to obtain target human voice data.
[0062] S206: Select a target speaker from at least one speaker according to the current voice playback scenario, and use the target speaker to play the target voice data.
[0063] In the embodiment of the present application, for the initial voice data collected by the microphone, a neural network model is first used to remove environmental noise; in order to remove multimedia noise, it is necessary to first obtain multimedia voice data that has been propagated and turned over. At this time, an adaptive filter is applied to obtain the parameters of the echo path, and the parameters are applied to perform echo processing on the basic multimedia data to obtain multimedia voice data with echo features added. Finally, the multimedia voice data with echo features added is eliminated from the data with environmental noise removed to obtain clean target voice data. In the actual application process, the corresponding speaker can also be selected to play clean target voice data according to whether the current voice playback scene is a communication scene or a privacy scene. Such a design can provide a more comfortable and convenient conversation environment for people in the car.
[0064] In relation to S201, the vehicle's voice control system may be a central control system or a subsystem in the central control system that is responsible for voice control. The central control system is used as an example for explanation below. In an embodiment of the present application, the basic multimedia data in the vehicle may be, for example, navigation data or music data. These data may be understood as data sources, that is, data that is not played. Data that is not played has not been propagated and does not contain other factors affected by the medium of the propagation path. The initial voice data collected by the microphone includes human voice data, multimedia data played by various speakers (built-in speakers in the vehicle and separately set speakers), and environmental noise.
[0065] Optionally, at least one microphone and at least one speaker are respectively arranged at a set position and a set height of a seat in the vehicle. In a specific example, the number of microphones and the number of speakers can be 4 respectively, and the setting positions can be referred to in Figure 1 , I will not go into details here.
[0066] In S202, a preset neural network model is applied to remove environmental noise in the initial voice data to obtain first voice data.
[0067] Among them, in order to improve the diversity of samples and thus improve the accuracy of the neural network, when making training samples, audio data at different speeds, different road conditions and different window opening degrees can be collected as training samples. Exemplarily, the audio here can include speech and noise, and the speech can include human voices and multimedia data played by the car's built-in speakers. When acquiring samples, speech and noise can be read in a loop, mixed into training samples, and corresponding features and tag data can be generated. In this example, the size of the data slice read can be 20ms, the step speed is 10ms, and there is 10ms of overlapping data.
[0068] Figure 3 A schematic diagram of generating a training sample provided in an embodiment of the present application includes at least the following steps:
[0069] S301: Initialize the object and parse the row parameters of the read command.
[0070] S302: Open an audio file for storing audio.
[0071] S303: Skip the first 150 frames of noise.
[0072] S304: Randomly generate scene parameters, which is executed once every 2821 rounds.
[0073] S305: Calculate features and mark data.
[0074] S306: Write the feature and tag data into the audio file.
[0075] S307: Arrange the processed audio files to obtain training samples, and repeat S304-S306.
[0076] In order to obtain a more effective model for the embodiment of the present application, speech with different signal-to-noise ratios can be used for training so that the obtained neural network model has better support. In the embodiment of the present application, the in-car noise and communication speech under different speeds, different road conditions and different window opening degrees are used to meet the conditions of different signal-to-noise ratios and obtain relatively high-quality training samples.
[0077] Among them, during the model training process, the following parameters can be used:
[0078] (1) The combination of speech gain and noise gain can represent different signal-to-noise ratios and different volume ranges. For example, the speech gain can be set to 20-40dBm, with a 10% probability of becoming silent. The speech gain can be set to 20-30dB, with a 10% probability of not adding noise.
[0079] (2) Two random second-order filters can simulate various special acoustic scenes, where the filter coefficient can be 0 to 0.75.
[0080] (3) Highest cutoff frequency, which can be used to simulate the frequency range of various acquisition environments. Lowpass: FREQ_SIZE*50rand(1) / 8, when FREQ_SIZE is 481, the value range is approximately 60 to 3000.
[0081] Regarding S203 and S204, as mentioned above, the environmental noise can be eliminated through the neural network model, but for a better user experience, the multimedia sound can also be removed.
[0082] The basic multimedia data directly obtained by the vehicle's voice control system has not been reflected or propagated, while the sound picked up by the microphone includes acoustic echoes, and the multi-path reflection results in echoes with different delays, including direct echoes and indirect echoes. Therefore, in order to obtain a purer human voice, the multimedia sound data to be removed must be multimedia voice data that includes acoustic echoes.
[0083] Among them, multimedia voice data including acoustic echo can be obtained through adaptive filtering algorithm. Optionally, the adaptive filtering algorithm includes adaptive echo cancellation function. The basic idea of adaptive echo cancellation here is to estimate the characteristic parameters of the echo path, generate a simulated echo path, obtain a simulated echo signal, and subtract the signal from the received signal to achieve echo cancellation. The key is to obtain the impulse response of the echo path. Since the echo path is usually unknown and time-varying, an adaptive filter is generally used to simulate the echo path. The notable feature of adaptive echo cancellation is real-time tracking and strong real-time performance.
[0084] That is, a preset adaptive filter can be used to simulate the echo path in the vehicle, and the basic multimedia voice data can be processed according to the parameters of the echo path to obtain the second voice data. The second voice data here is the multimedia voice data with the characteristics of the propagation medium added.
[0085] Figure 4 A schematic diagram of the processing process of an adaptive algorithm provided in an embodiment of the present application, wherein x(n) is basic multimedia data, w(n) is an adaptive filter, y(n) is second voice data, d(n) is first voice data, and e(n) is target voice data.
[0086] In S205, since the first voice data is human voice data and multimedia voice data that have passed through a transmission medium, and the second voice data is multimedia voice data that has passed through a transmission medium, the second voice data is used to eliminate the first voice data to obtain target human voice data. The target human voice data obtained here is pure target human voice data.
[0087] Regarding S206, in order to improve the user experience, a target speaker may be selected from at least one speaker according to the current voice playback scenario, and the target speaker may be used to play the target voice data.
[0088] Among them, the current voice playback scenarios are divided into communication scenarios and privacy scenarios. The communication scenario means that each user in the car can normally receive the voice played by the speaker; the privacy scenario means that user 1 can specify that user 2 can hear what he says, and other users cannot hear what user 2 says.
[0089] Figure 5a A schematic diagram of a playback process in different scenarios provided in an embodiment of the present application, refer to Figure 5a .
[0090] S500: Determine the current voice playback scenario.
[0091] S501: If the current voice playback scene is a communication scene, at least one speaker is determined as a target speaker, and target human voice data is played through the at least one speaker.
[0092] Among them, in the communication scenario, if the number of speakers is 4, all 4 speakers are used as target speakers, and the target human voice data is played through these 4 speakers.
[0093] S502: If the current voice playback scene is a privacy scene, a speaker with playback authority is determined as a target speaker among at least one speaker, and the target voice data is played through the target speaker.
[0094] In the privacy scenario, in response to the user's operation of selecting a privacy user, the speaker associated with the selected privacy user can be determined as a speaker with playback authority.
[0095] This process can be in response to the user's operation on the smart terminal or the button set in the car; the smart terminal includes the user's mobile smart terminal or the fixed smart terminal set in the car. The button in the car can be set in a position on the car body that is convenient for the user to operate, and the command of the button can be transmitted to the central control system of the car. The user's mobile smart terminal can be a smart phone used by the user, and the set application can be installed on the smart phone to transmit data with the central control system of the car. The fixed smart terminal can be a car-mounted tablet computer, which can be installed in a position on the headrest of the car seat for the user to view and operate.
[0096] Figure 5b A schematic diagram of the installation position of a smart terminal provided in an embodiment of the present application, wherein 51 is a vehicle-mounted tablet computer.
[0097] If user 1 has a need for private communication, he can select the object (or objects) with which he wants to have a private conversation by operating buttons, mobile phones or in-vehicle tablet computers, etc., and then generate a selection instruction, and send the selection instruction to the central control system of the car so that the central control system can recognize the selection instruction. For example, the selection instruction includes the identification of the target speaker. With this design, the central control system can determine that the target speaker is the speaker at the location of user 2 by analyzing the selection quality, and play the target human voice data through the target speaker.
[0098] Figure 6 A schematic diagram of a display interface of a smart terminal provided in an embodiment of the present application, wherein: Figure 6 (a) in the figure is the scene selection interface shown to the user. Figure 6 (b) in the figure is an interface for the user to select a privacy object in a privacy scenario. In this example, the selectable privacy objects are the driver A, the front passenger B, the left rear seat passenger C, and the right rear seat passenger D.
[0099] S503: If the current voice playback scene is a privacy scene, forward feedback active noise reduction processing is performed on the target voice data to obtain reverse voice data of the target voice data; and the reverse voice data is played through a speaker without playback permission.
[0100] In order to further reduce the possibility of other users in privacy scenarios hearing private conversations and to protect private conversations, forward feedback active noise reduction technology is used to process the target voice data, so that the reverse voice data of the target voice data can be obtained, and the reverse voice data is played on speakers without playback permission, and the corresponding non-privacy users cannot hear the target voice data.
[0101] In the embodiment of the present application, microphones and speakers are arranged reasonably at each seat headrest position. The microphone is used to obtain the voice of the driver or passenger (including multimedia voice and environmental noise) and send it to the central control system, which suppresses and eliminates the noise of the voice, and the speaker plays the processed voice. In scenes with high noise such as high-speed driving, or when some people in the car want to have a private conversation, better voice transmission services are provided, effectively improving the driving experience of people in the car when talking.
[0102] In order to make the technical solution of this application more perfect, a complete flowchart is used for explanation below. Figure 7 A flowchart of a complete method for processing vehicle voice data provided in an embodiment of the present application.
[0103] S701: Acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone.
[0104] The implementation method of this step can refer to step S201 and will not be repeated here.
[0105] S702: Apply a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data.
[0106] The implementation method of this step can refer to step S202, which will not be repeated here.
[0107] S703: Apply a pre-set adaptive filter to simulate the echo path in the vehicle.
[0108] The implementation method of this step can refer to step S203, which will not be repeated here.
[0109] S704: Process the basic multimedia voice data according to the parameters of the echo path to obtain second voice data.
[0110] The implementation method of this step can refer to step S204, which will not be repeated here.
[0111] S705: Apply the second voice data to perform elimination processing on the first voice data to obtain target human voice data.
[0112] The implementation method of this step can refer to step S207, which will not be repeated here.
[0113] S706: If the current voice playback scene is a communication scene, determine at least one speaker as a target speaker, and play the target human voice data through the at least one speaker.
[0114] The implementation method of this step can refer to step S401 and will not be repeated here.
[0115] S707: If the current voice playback scenario is a privacy scenario, then among at least one speaker, determine a speaker with playback permission as the target speaker, and play the target voice data through the target speaker; perform forward feedback active noise reduction processing on the target voice data to obtain reverse voice data of the target voice data; and play the reverse voice data through a speaker without playback permission.
[0116] The implementation method of this step can refer to steps S402 and S403, which will not be described in detail here.
[0117] In the embodiment of the present application, each seat headrest is reasonably arranged with a microphone and a speaker, which effectively achieves accurate noise elimination, improves the quality of conversation, and effectively improves the conversation experience of the passengers in the car. Each seat headrest can be equipped with multiple directional speakers, and through actual adjustment such as installation position and placement direction, better noise reduction and sound effects can be achieved.
[0118] In order to make the technical solution of the present application easier to understand, a complete flowchart is used below to illustrate the process of processing vehicle voice data. Figure 8 A schematic diagram of a vehicle voice data processing process provided by an embodiment of the present application, wherein training data 1 is 32 floating point data f32, training data 2 is in the format of hdf5, the first weight data is in the format of hdf5, and the second weight data is in the format applicable to a recurrent neural network (RNN). In this example, voice activity detection (VAD) is a technology used for voice processing, the purpose of which is to detect whether a voice signal exists.
[0119] like Fig. 9 As shown, based on the same inventive concept, an embodiment of the present application provides a vehicle voice data processing device, which is applied to a vehicle voice control system. At least one microphone and at least one speaker are arranged in the vehicle, and the device includes a data acquisition unit 91, a data processing unit 92 and a playback unit 93.
[0120] The data acquisition unit 91 is used to: acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone;
[0121] The data processing unit 92 is used to: apply a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data;
[0122] The data processing unit 92 is further used to: apply a preset adaptive filter to simulate an echo path in the vehicle;
[0123] The data processing unit 92 is further used to: process the basic multimedia voice data according to the parameters of the echo path to obtain second voice data;
[0124] The data processing unit 92 is further used to: apply the second voice data to eliminate the first voice data to obtain target human voice data;
[0125] The playing unit 93 is used to select a target speaker from at least one speaker according to a current voice playing scenario, and use the target speaker to play the target voice data.
[0126] In an optional implementation, the playback unit 93 is specifically configured to:
[0127] If the current voice playback scene is a communication scene, determining at least one speaker as a target speaker, and playing the target human voice data through the at least one speaker;
[0128] If the current voice playback scene is a privacy scene, a speaker with playback permission is determined as a target speaker among at least one speaker, and the target voice data is played through the target speaker.
[0129] In an optional implementation, the playback unit 93 is specifically configured to:
[0130] If the current voice playback scene is a privacy scene, forward feedback active noise reduction processing is performed on the target voice data to obtain reverse voice data of the target voice data;
[0131] Play reverse voice data through a speaker without playback permission.
[0132] In an optional implementation manner, the playing unit 93 is specifically configured to determine a speaker with playing authority in the following manner:
[0133] In response to the user's operation of selecting a privacy user, a speaker associated with the selected privacy user is determined to be a speaker with playback authority.
[0134] In an optional implementation, the playback unit 93 is specifically configured to:
[0135] Responding to the user's operation on a smart terminal or a button set in the car; wherein the smart terminal includes the user's mobile smart terminal or a fixed smart terminal set in the car.
[0136] In an optional embodiment, at least one microphone and at least one speaker are respectively arranged at a set position and a set height of a seat in the vehicle.
[0137] In an optional implementation, the initial voice data includes human voice data, multimedia data played by each speaker, and environmental noise.
[0138] In an optional implementation, the training samples of the preset neural network model during the training process meet the following conditions:
[0139] The training samples are samples at different speeds, different road conditions and different window opening degrees.
[0140] Since the device is the device in the method in the embodiment of the present application, and the principle of solving the problem by the device is similar to that of the method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.
[0141] like Fig.10 As shown, based on the same inventive concept, an embodiment of the present application provides a voice control system for a vehicle, wherein at least one microphone and at least one speaker are arranged in the vehicle, and the voice control system includes a processor 101 and a data transmission unit 102;
[0142] The data transmission unit 102 is configured to perform:
[0143] Acquire basic multimedia data in the vehicle and initial voice data collected by at least one microphone;
[0144] The processor 101 is configured to execute:
[0145] Applying a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data;
[0146] Applying pre-set adaptive filters to simulate the echo path inside the vehicle;
[0147] Processing the basic multimedia voice data according to the parameters of the echo path to obtain second voice data;
[0148] Applying the second voice data to perform elimination processing on the first voice data to obtain target human voice data;
[0149] According to the current voice playback scenario, a target speaker is selected from at least one speaker, and the target speaker is used to play the target voice data.
[0150] In an optional implementation, the processor 101 is specifically configured to execute:
[0151] If the current voice playback scene is a communication scene, determining at least one speaker as a target speaker, and playing the target human voice data through the at least one speaker;
[0152] If the current voice playback scene is a privacy scene, a speaker with playback permission is determined as a target speaker among at least one speaker, and the target voice data is played through the target speaker.
[0153] In an optional implementation, the processor 101 is specifically configured to execute:
[0154] If the current voice playback scene is a privacy scene, forward feedback active noise reduction processing is performed on the target voice data to obtain reverse voice data of the target voice data;
[0155] Play reverse voice data through a speaker without playback permission.
[0156] In an optional implementation, the processor 101 is specifically configured to perform the following method to determine the speaker with the playback permission:
[0157] In response to the user's operation of selecting a privacy user, a speaker associated with the selected privacy user is determined to be a speaker with playback authority.
[0158] In an optional implementation, the processor 101 is specifically configured to execute:
[0159] Responding to the user's operation on a smart terminal or a button set in the car; wherein the smart terminal includes the user's mobile smart terminal or a fixed smart terminal set in the car.
[0160] In an optional embodiment, at least one microphone and at least one speaker are respectively arranged at a set position and a set height of a seat in the vehicle.
[0161] In an optional implementation, the initial voice data includes human voice data, multimedia data played by each speaker, and environmental noise.
[0162] In an optional implementation, the training samples of the preset neural network model during the training process meet the following conditions:
[0163] The training samples are samples at different speeds, different road conditions and different window opening degrees.
[0164] Based on the same inventive concept, an embodiment of the present application provides a vehicle, comprising the voice control system in the aforementioned embodiment.
[0165] The embodiment of the present application also provides a computer storage medium, in which computer program instructions are stored. When the instructions are executed on a computer, the computer executes the steps of the above-mentioned vehicle voice data processing method.
[0166] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0167] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0168] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0169] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0170] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.
Claims
1. A method for processing vehicle voice data, characterized in that: A voice control system applied to the vehicle, wherein at least one microphone and at least one speaker are arranged in the vehicle, and the method comprises: Acquiring basic multimedia data in the vehicle and initial voice data collected by the at least one microphone; Applying a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data; applying a predetermined adaptive filter to simulate an echo path within the vehicle; Processing the basic multimedia voice data according to the parameters of the echo path to obtain second voice data; Applying the second voice data to perform elimination processing on the first voice data to obtain target human voice data; According to the current voice playing scenario, a target speaker is selected from the at least one speaker, and the target speaker is used to play the target human voice data.
2. The method according to claim 1, characterized in that The step of selecting a target speaker from the at least one speaker according to the current voice playback scenario, and using the target speaker to play the target human voice data includes: If the current voice playback scene is a communication scene, determining the at least one speaker as a target speaker, and playing the target human voice data through the at least one speaker; If the current voice playback scene is a privacy scene, a speaker with playback authority is determined as a target speaker among the at least one speaker, and the target human voice data is played through the target speaker.
3. The method according to claim 2, characterized in that The method further comprises: If the current voice playback scene is a privacy scene, performing forward feedback active noise reduction processing on the target human voice data to obtain reverse voice data of the target human voice data; The reverse voice data is played through a speaker without playback permission.
4. The method according to claim 2, characterized in that: The method further includes determining a speaker with playback permission by: In response to the user's operation of selecting a privacy user, a speaker associated with the selected privacy user is determined to be a speaker with playback authority.
5. The method according to claim 4, characterized in that The operation in response to the user selecting a privacy user includes: Responding to the user's operation on a smart terminal or a button set in the car; wherein the smart terminal includes the user's mobile smart terminal or a fixed smart terminal set in the car.
6. The method according to claim 1, characterized in that The at least one microphone and the at least one speaker are respectively disposed at a set position and a set height of a seat in the vehicle.
7. The method according to claim 1, characterized in that The initial voice data includes human voice data, multimedia data played by various speakers in the vehicle, and environmental noise.
8. The method according to any one of claims 1 to 7, characterized in that: The training samples of the preset neural network model during the training process meet the following conditions: The training samples are samples at different speeds, different road conditions and different window opening degrees.
9. A voice control system for a vehicle, characterized in that: At least one microphone and at least one speaker are arranged in the vehicle, and the voice control system includes a processor and a data transmission unit; The data transmission unit is configured to perform: Acquiring basic multimedia data in the vehicle and initial voice data collected by the at least one microphone; The processor is configured to perform: Applying a preset neural network model to remove environmental noise in the initial voice data to obtain first voice data; applying a predetermined adaptive filter to simulate an echo path within the vehicle; Processing the basic multimedia voice data according to the parameters of the echo path to obtain second voice data; Applying the second voice data to perform elimination processing on the first voice data to obtain target human voice data; According to the current voice playing scenario, a target speaker is selected from the at least one speaker, and the target speaker is used to play the target human voice data.
10. A vehicle, characterized in that: Includes the voice control system as claimed in claim 9.