WO2024120331A1 - 降噪方法、装置以及运载工具 - Google Patents

降噪方法、装置以及运载工具 Download PDF

Info

Publication number
WO2024120331A1
WO2024120331A1 PCT/CN2023/136081 CN2023136081W WO2024120331A1 WO 2024120331 A1 WO2024120331 A1 WO 2024120331A1 CN 2023136081 W CN2023136081 W CN 2023136081W WO 2024120331 A1 WO2024120331 A1 WO 2024120331A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
road condition
control
condition information
reference signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/136081
Other languages
English (en)
French (fr)
Inventor
王浩
邱小军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Publication of WO2024120331A1 publication Critical patent/WO2024120331A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
    • G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • G10K11/16—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C5/00—Registering or indicating the working of vehicles
    • G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C5/00—Registering or indicating the working of vehicles
    • G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
    • G07C5/0841—Registering performance data
    • G07C5/085—Registering performance data using electronic data carriers
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10K—SOUND-PRODUCING DEVICES; METHODS OR DEVICES FOR PROTECTING AGAINST, OR FOR DAMPING, NOISE OR OTHER ACOUSTIC WAVES IN GENERAL; ACOUSTICS NOT OTHERWISE PROVIDED FOR
    • G10K11/00—Methods or devices for transmitting, conducting or directing sound in general; Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • G10K11/16—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general
    • G10K11/175—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound
    • G10K11/178—Methods or devices for protecting against, or for damping, noise or other acoustic waves in general using interference effects; Masking sound by electro-acoustically regenerating the original acoustic waves in anti-phase
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04R—LOUDSPEAKERS, MICROPHONES, GRAMOPHONE PICK-UPS OR LIKE ACOUSTIC ELECTROMECHANICAL TRANSDUCERS; ELECTRIC HEARING AIDS; PUBLIC ADDRESS SYSTEMS
    • H04R3/00—Circuits for transducers

Definitions

  • the embodiments of the present application relate to the field of noise reduction, and more specifically, to a noise reduction method, device and vehicle.
  • the more commonly used noise reduction scheme is to obtain the noise signal near the noise source as a reference signal, and determine the control parameters based on the reference signal to obtain the control signal. Since the control signal reaches the user's ear after being played through the speaker, the original noise at the user's ear is approximately the same in magnitude and approximately opposite in phase, so the noise heard by the user is reduced, achieving noise elimination.
  • the above method only obtains the control signal through the reference signal and plays the control signal through the speaker, and the noise reduction effect is poor, thereby affecting the user's driving experience.
  • the embodiments of the present application provide a noise reduction method, device and vehicle, which can improve the noise reduction effect and thus improve the user's driving experience.
  • a noise reduction method comprising: acquiring a reference signal, road condition information around a vehicle, and driving parameters of the vehicle, the reference signal being collected by a first sensor, and the first sensor being a sensor deployed near a noise source of the vehicle; and controlling a sound-emitting device of the vehicle to play a control signal according to the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, the control signal being used to reduce noise near the ears or head of a user, the user being located in a cabin of the vehicle.
  • the reference signal can be used to indicate the noise signal collected near the noise source.
  • the noise generated by the noise source may be the noise generated by the friction between the vehicle and the road surface and the air.
  • the noise generated by the noise source may be the noise generated by the friction between internal parts of the vehicle.
  • the first sensor may be an accelerometer or an acoustic sensor.
  • the driving parameters of the vehicle may include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • the road condition information around the vehicle may be used to indicate the type of road on which the vehicle is currently traveling.
  • the road type may include an asphalt road, a cement road, a gravel road, and the like.
  • the sound-generating device may include: a speaker, a sound, a horn or more.
  • the noise reduction effect when reducing noise, since the road condition information around the vehicle and the driving parameters of the vehicle are taken into consideration for noise reduction processing, the noise reduction effect can be improved, thereby improving the user's driving experience.
  • the method also includes: obtaining position information of the ear or head of the user in the cabin of the vehicle; and controlling the sound device of the vehicle to play the control signal based on the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, including: controlling the sound device of the vehicle to play the control signal based on the reference signal, the road condition information around the vehicle, the driving parameters of the vehicle, and the position information of the ear or head of the user.
  • image information of the user's head or ears can be acquired by a camera device deployed inside the vehicle, thereby determining the position information of the user's ears or head.
  • the position information of the user's ear or head may be determined by a sensor with a ranging function (eg, a laser sensor, etc.).
  • a sensor with a ranging function eg, a laser sensor, etc.
  • the position information of the user's ear or head may be measured position information or estimated position information, wherein the estimated position information This may include the relative or absolute position of the user's ears or head in the cockpit, etc.
  • the position information of the user's ears or head may be a photograph or measurement data.
  • the control signal played by the sound-emitting device can more specifically reduce the noise near the user's head or ears, which can further improve the user's driving experience.
  • the method also includes: obtaining an error signal, wherein the error signal is collected by a second sensor deployed near the position of the user's ear or head; and controlling the sound-emitting device of the vehicle to play the control signal according to the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, including: controlling the sound-emitting device of the vehicle to play the control signal according to the reference signal, the road condition information around the vehicle, the driving parameters of the vehicle, and the error signal.
  • the second sensor may be an error microphone for collecting noise signals near the ear or head of the user.
  • the error signal can be used to assist in controlling the sound-generating device of the vehicle to play the control signal, thereby achieving a better noise reduction effect for the control signal played by the sound-generating device.
  • controlling the sound-emitting device of the vehicle to play the control signal based on the reference signal, the road condition information around the vehicle and the driving parameters of the vehicle includes: determining control parameters based on the reference signal, the road condition information around the vehicle and the driving parameters of the vehicle, the control parameters including filter coefficients; and controlling the sound-emitting device of the vehicle to play the control signal based on the control parameters and the reference signal.
  • the error signal or the position information of the user's ear or head may be used as a reference.
  • controlling a sound-generating device of a vehicle to play a control signal according to the control parameter and the reference signal includes: determining the control signal according to the control parameter and the reference signal; and using the sound-generating device of the vehicle to play the control signal.
  • control signal played by the sound-generating device is determined by the control parameters, and in the currently commonly used noise reduction schemes, the control parameters are determined only by the reference signal, which often cannot obtain the optimal control parameters, thereby resulting in poor noise reduction effect.
  • the control parameters when determining the control parameters, the road condition information around the vehicle and the driving parameters of the vehicle are considered, so that the determined control parameters are more accurate and the noise reduction effect is better.
  • determining the control parameters based on the reference signal, the road condition information around the vehicle and the driving parameters of the vehicle includes: inputting the reference signal, the road condition information around the vehicle and the driving parameters of the vehicle into a first model to obtain the control parameters; wherein the first model is obtained based on training samples, and the training samples include: sample road condition information, sample reference signals, sample driving parameters and sample control parameters.
  • the first model can be a deep neural networks (DNN) model, which can also be composed of a conventional classification algorithm that can realize the classification function.
  • DNN deep neural networks
  • the process of training the first model may be: inputting sample reference signals, sample road condition information and sample driving parameters into the first model to obtain prediction control parameters; training the first model according to the prediction control parameters and the sample control parameters.
  • the prediction control parameters output by the first model can be continuously close to the sample control parameters, so that the first model can fit the best control parameters according to the input data in practical applications.
  • the sample control parameters can be the data labels corresponding to the prediction control parameters.
  • a reference signal, road condition information around the vehicle, and driving parameters of the vehicle can be input into a trained first model to obtain control parameters.
  • the control parameters can be obtained more efficiently and the determined control signal can be relatively more accurate.
  • obtaining the road condition information around the vehicle includes: obtaining the road condition information around the vehicle through data collected by a third sensor, and the third sensor is located on the vehicle.
  • the third sensor may include one or more sensors such as a digital video recorder (DVR), a gravity sensor, an external camera (e.g., a surround-view camera), etc.
  • DVR digital video recorder
  • a gravity sensor e.g., a gravity sensor
  • an external camera e.g., a surround-view camera
  • the third sensor can be used to collect image information from the outside of the vehicle.
  • the third sensor may include a camera device, which can be deployed on the outside of the vehicle, for example, at the front of the tire, at the vehicle logo, at the rearview mirror, at the reflector, etc.
  • the camera device may include a monocular camera, a binocular camera, a structured light camera, a panoramic camera, etc.
  • the data collected by the third sensor may be image information obtained by the camera device, and the road condition information on which the vehicle is traveling can be determined through the image information.
  • the image information may include static image information or video stream information.
  • the image information may be stored in the form of an image or video, or in the form of image or video parameters, for example, parameter information such as image brightness, grayscale, color distribution, contrast, and pixels.
  • the third sensor may be a DVR, and the road condition information on which the vehicle is traveling can be determined based on the image information on the outside of the vehicle collected by the DVR.
  • the third sensor can be one or more sensors.
  • some sensors can collect image information outside the vehicle, and other sensors can collect image information inside the vehicle. The image information outside the vehicle and the image information inside the vehicle are combined to determine the road condition information of the vehicle.
  • the third sensor when the third sensor is a plurality of sensors, some sensors can collect image information of the outside of the vehicle, and other sensors can collect component data inside the vehicle.
  • the other sensors are gravity sensors, which can collect data reflecting the bumpy conditions of the user's seat.
  • the road condition information of the vehicle can be determined by combining the image information of the outside of the vehicle and the data reflecting the bumpy conditions of the user's seat.
  • the vehicle can automatically obtain road condition information around the vehicle in specific driving scenarios, for example, when the vehicle is driving on a gravel road, a concave cement road, an asphalt road, and the like.
  • the road condition information around the vehicle can be obtained through the data collected by the third sensor, so that the vehicle can control the sound-emitting device to play the control signal according to the road condition information.
  • the current road classification (eg, expressway, national highway, etc.) may be acquired through navigation information to determine the road condition information around the vehicle.
  • navigation information can be used to obtain road condition information, so that the vehicle can control the sound-emitting device to play a control signal according to the road condition information.
  • a model training method which includes: obtaining secondary path information, a sample reference signal and a sample error signal; determining multiple groups of control parameters based on the secondary path information, the sample reference signal and the sample error signal; obtaining sample road condition information and sample driving parameters; inputting the sample reference signal, the sample road condition information and the sample driving parameters into a first model to obtain a predicted probability value corresponding to each group of control parameters in the multiple groups of control parameters; training the first model based on the predicted probability value and the target probability value, wherein the target probability value is a pre-set probability value.
  • the above target probability value may also be replaced by a score, that is, the first model may be trained by the predicted score and the target score of each set of control parameters.
  • the predicted probability value corresponding to each set of control parameters output by the first model can be continuously approached to the target probability value, so that the first model can obtain the optimal control parameters according to the input data in practical applications.
  • a model training method which includes: obtaining secondary path information, a sample reference signal and a sample error signal; determining sample control parameters based on the secondary path information, the sample reference signal and the sample error signal; obtaining sample road condition information around a vehicle and sample driving parameters of the vehicle; inputting the sample reference signal, the sample road condition information and the sample driving parameters into a first model to obtain predicted control parameters; and training the first model based on the predicted control parameters and the sample control parameters.
  • the sample control parameter may be a data label corresponding to the prediction control parameter.
  • the predicted control parameters output by the first model can be made to continuously approach the sample control parameters, so that the first model can fit the optimal control parameters according to the input data in practical applications.
  • the sample control parameter is a control parameter with a maximum probability value in a first control parameter set, and the first control parameter set includes one or more control parameters.
  • the sample control parameter may also be a control parameter with the highest score in the first control parameter set.
  • the sample control parameter may also be a control parameter with the highest score in the second control parameter set.
  • the sample driving parameters include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • a noise reduction device which includes: an acquisition unit and a processing unit; the acquisition unit is used to acquire a reference signal, road condition information around a vehicle, and driving parameters of the vehicle, the reference signal is collected by a first sensor, and the first sensor is a sensor deployed near the noise source of the vehicle; the processing unit is used to control the sound-emitting device of the vehicle to play a control signal according to the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, the control signal is used to reduce the noise near the ears or head of a user, and the user is located in the cabin of the vehicle.
  • the acquisition unit is further configured to acquire the information in the cabin of the vehicle.
  • the position information of the user's ear or head; the processing unit is specifically used to control the sound device of the vehicle to play the control signal according to the reference signal, the road condition information around the vehicle, the driving parameters of the vehicle and the position information of the user's ear or head.
  • the acquisition unit is also used to acquire an error signal, which is collected by a second sensor deployed near the user's ear or head; the processing unit is specifically used to control the sound device of the vehicle to play the control signal based on the reference signal, road condition information around the vehicle, driving parameters of the vehicle and the error signal.
  • the processing unit is specifically used to: determine control parameters based on the reference signal, road condition information around the vehicle and driving parameters of the vehicle, the control parameters including filter coefficients; and control the sound-emitting device of the vehicle to play the control signal based on the control parameters and the reference signal.
  • the processing unit is specifically used to input the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle into a first model to obtain the control parameters; wherein the first model is obtained based on training samples, and the training samples include: sample road condition information, sample reference signals, sample driving parameters, and sample control parameters.
  • the acquisition unit is further used to acquire navigation information sent by the server; the processing unit is further used to determine the road condition information around the vehicle based on the navigation information.
  • the driving parameters around the vehicle include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • a model training device which includes: an acquisition unit and a processing unit; the acquisition unit is used to acquire secondary path information, a sample reference signal and a sample error signal; the processing unit is used to determine multiple groups of control parameters based on the secondary path information, the sample reference signal and the sample error signal; the acquisition unit is also used to acquire sample road condition information and sample driving parameters; the processing unit is also used to input the sample reference signal, the sample road condition information and the sample driving parameters into a first model to obtain a predicted probability value corresponding to each group of control parameters in the multiple groups of control parameters; the first model is trained according to the predicted probability value and the target probability value, wherein the target probability value is a pre-set probability value.
  • a model training device which includes: an acquisition unit and a processing unit; the acquisition unit is used to acquire secondary path information, a sample reference signal and a sample error signal; the processing unit is used to: input the sample reference signal, the sample road condition information and the sample driving parameters into a first model to obtain predicted control parameters; and train the first model according to the predicted control parameters and the sample control parameters.
  • the sample control parameter is a control parameter with a maximum probability value in a first control parameter set, and the first control parameter set includes one or more control parameters.
  • the sample control parameter is a control parameter with a maximum probability value in a second control parameter set
  • the second control parameter set is obtained by classifying the first control parameter set
  • the sample driving parameters include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • a noise reduction device comprising: at least one processor and a memory, wherein the at least one processor is coupled to the memory and is used to read and execute instructions in the memory, so that the device implements any one of the implementation methods in the first aspect above.
  • a model training device which includes: at least one processor and a memory, wherein the at least one processor is coupled to the memory and is used to read and execute instructions in the memory, so that the device implements the method of any one of the implementation modes in the second aspect or the third aspect above.
  • a computer-readable medium stores a program code, and when the computer program code runs on a computer, the computer executes a method implemented in any one of the first to third aspects above.
  • a computer program product wherein the computer product includes a computer program, and when the computer program is executed, the computer executes a method implemented in any one of the first to third aspects above.
  • a chip comprising a circuit, the circuit being configured to execute any one of the first to third aspects above.
  • a method of implementing the method is provided, the chip comprising a circuit, the circuit being configured to execute any one of the first to third aspects above.
  • a noise reduction system comprising: a noise reduction device in any one of the implementations of the fourth aspect and one or more sound generating devices.
  • a noise reduction system comprising a computing platform and one or more sound-emitting devices, wherein the computing platform is used to: obtain road condition information around a vehicle, driving parameters of the vehicle, and a reference signal collected by a first sensor, wherein the first sensor is a sensor deployed near a noise source of the vehicle; and control the one or more sound-emitting devices to play a control signal based on the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, wherein the control signal is used to reduce noise near the ears or head of a user located in a cabin of the vehicle.
  • a server comprising a model training device in any one of the implementations of the fifth or sixth aspect.
  • a vehicle comprising: a noise reduction device in any one of the implementations of the fourth aspect, or a noise reduction system as described in the twelfth or thirteenth aspect.
  • the vehicle is a vehicle.
  • the technical solution provided by the embodiment of the present application can take into account the road conditions around the vehicle and the driving parameters of the vehicle when performing noise reduction processing, which can improve the noise reduction effect and thus improve the user's driving experience.
  • This solution can also take into account the position of the user's ears or head and the error signal when reducing noise, so that the control signal played by the sound-emitting device can more specifically reduce the noise near the user's head or ears, further improving the user's driving experience.
  • this solution can quickly determine the control parameters through the first model, so that the efficiency of obtaining the control parameters is higher and the determined control signal is relatively more accurate.
  • FIG1 is a functional schematic diagram of a vehicle provided in an embodiment of the present application.
  • FIG2 is a feedforward solution of a noise reduction method provided by an embodiment of the present application.
  • FIG3 is a feedback scheme of a noise reduction method provided in an embodiment of the present application.
  • FIG4 is a system architecture applicable to the noise reduction method provided in an embodiment of the present application.
  • FIG5 is a schematic diagram of a noise reduction method provided in an embodiment of the present application.
  • FIG6 is a flow chart of a model training method provided in an embodiment of the present application.
  • FIG7 is a schematic block diagram of another noise reduction method provided in an embodiment of the present application.
  • FIG8 is a flow diagram of an algorithm model training of a noise reduction method provided in an embodiment of the present application.
  • FIG9 is a schematic diagram of a process of using an algorithm model of a noise reduction method provided in an embodiment of the present application.
  • FIG10 is a flow diagram of another algorithm model training of the denoising method provided in an embodiment of the present application.
  • FIG11 is a schematic diagram of another algorithm model usage process of the noise reduction method provided in an embodiment of the present application.
  • FIG12 is a noise reduction or model training device provided in an embodiment of the present application.
  • FIG13 is another noise reduction or model training device provided in an embodiment of the present application.
  • Fig. 1 is a functional schematic diagram of a vehicle 100 provided in an embodiment of the present application. It should be understood that Fig. 1 and the related description are only examples and do not limit the vehicle in the embodiment of the present application.
  • the vehicle 100 may be configured in a fully or partially autonomous driving mode, or may be manually driven by a user.
  • the vehicle 100 may obtain environmental information about its surroundings through the perception system 120, and obtain an autonomous driving strategy based on the analysis of the surrounding environmental information to achieve fully autonomous driving, or present the analysis results to the user to achieve partially autonomous driving.
  • the vehicle 100 may include a variety of subsystems, such as a sensing system 120, a computing platform 130, and a sound generating device 140.
  • the vehicle 100 may include more or fewer subsystems, and each subsystem may include one or more components.
  • each subsystem and component of the vehicle 100 may be interconnected by wire or wirelessly.
  • the perception system 120 may include several sensors for sensing information about the environment around the vehicle 100.
  • the perception system 120 may include a positioning system, which may be a global positioning system (GPS), a Beidou system, or other positioning systems.
  • the perception system 120 may include one or more of an inertial measurement unit (IMU), a laser radar, a millimeter wave radar, an ultrasonic radar, and a camera.
  • IMU inertial measurement unit
  • the computing platform 130 may include processors 131 to 13n (n is a positive integer).
  • the processor is a circuit with signal processing capability.
  • the processor may be a circuit with instruction reading and execution capability, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP).
  • CPU central processing unit
  • GPU graphics processing unit
  • DSP digital signal processor
  • the processor may implement certain functions through the logical relationship of a hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable, such as a hardware circuit implemented by a processor such as an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA.
  • ASIC application-specific integrated circuit
  • PLD programmable logic device
  • the process of the processor loading a configuration document to implement the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units.
  • the processor may also be a hardware circuit designed for artificial intelligence, which may be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc.
  • the computing platform 130 may also include a memory, the memory is used to store instructions, and some or all of the processors 131 to 13n may call instructions in the memory to implement corresponding functions.
  • Computing platform 130 may control functions of vehicle 100 based on input received from various subsystems (eg, perception system 120). In some embodiments, computing platform 130 may be used to provide control over many aspects of vehicle 100 and its subsystems.
  • the sound-generating device 140 may include: a speaker, a sound, a horn, etc., which is used to play the control signal, thereby reducing the noise near the user's ears or head.
  • FIG. 1 should not be understood as a limitation on the embodiments of the present application.
  • a vehicle 100 traveling on a road can identify objects in its surroundings to determine adjustments to the current speed.
  • the objects can be other vehicles, traffic control devices, or other types of objects.
  • each identified object can be considered independently, and based on the respective characteristics of the object, such as the object's current speed, acceleration, spacing from the vehicle, etc., can be used to determine the speed to be adjusted for the vehicle 100.
  • the vehicle 100 or a sensing and computing device (e.g., computing platform 130) associated with the vehicle 100 can predict the behavior of the identified objects based on the characteristics of the identified objects and the state of the surrounding environment (e.g., traffic, rain, ice on the road, etc.).
  • each of the identified objects depends on the behavior of each other, so all of the identified objects can also be considered together to predict the behavior of a single identified object.
  • the vehicle 100 in the present application may include: road vehicles, water vehicles, air vehicles, industrial equipment, agricultural equipment, or entertainment equipment, etc.
  • the vehicle 100 may be a vehicle, which is a vehicle in a broad sense, and may be a vehicle (such as a commercial vehicle, a passenger car, a motorcycle, a flying car, a train, etc.), an industrial vehicle (such as a forklift, a trailer, a tractor, etc.), an engineering vehicle (such as an excavator, a bulldozer, a crane, etc.), agricultural equipment (such as a mower, a harvester, etc.), amusement equipment, a toy vehicle, etc.
  • the embodiment of the present application does not specifically limit the type of vehicle.
  • the vehicle 100 may be a vehicle such as an airplane or a ship.
  • Noise reduction can be performed using a feedforward solution or a feedback solution.
  • the difference between the feedforward solution and the feedback solution is that the feedforward solution needs to collect reference signals to determine control parameters, while the feedback solution does not need to collect reference signals to determine control parameters.
  • Figures 2 and 3 are model block diagrams of the feedforward solution and the feedback solution, respectively.
  • This control parameter can be the coefficient of a finite impulse response (FIR) filter.
  • FIR finite impulse response
  • control parameters do not require real-time calculation, one or more sets of control parameters need to be calculated in advance, and a suitable set of control parameters needs to be selected in advance for real-time filtering during operation. Since the control parameters calculated in advance are difficult to meet all scenarios, in some special scenarios, the optimal control parameters may not be selected. In this way, the control signal determined by the control parameters has a poor noise reduction effect after being played through the speaker, which in turn affects the user's driving experience.
  • the present invention provides a noise reduction method, device and vehicle, which can improve the noise reduction effect and thus improve the user's driving experience.
  • the system architecture to which the noise reduction method is applicable is first introduced.
  • FIG4 is a system architecture applicable to the noise reduction method provided in an embodiment of the present application.
  • the system architecture 400 includes: an accelerometer, an error microphone, a secondary source speaker, a preamplifier, an acceleration signal amplifier, a power amplifier, a vehicle computer/audio system/controller, a camera, and a vehicle computer.
  • the accelerometer is used to collect the vibration signal of the bottom of the vehicle as a reference signal, and the vibration signal can be a signal close to the noise source.
  • the accelerometer can be used to collect the signal at the spoiler, windshield, and rearview mirror as a reference signal, or the microphone can also replace the accelerometer to collect the reference signal.
  • the error microphone can be used to collect noise signals near the human ear.
  • the secondary source speaker can be a device that emits a reverse noise signal to offset or reduce the original noise.
  • the secondary source speaker can include one or more speakers, which can be arranged next to the user's ear, at the headrest, seat or door panel of the vehicle, etc.
  • the secondary source speaker can be one of the sound-generating devices.
  • the preamplifier and the acceleration signal amplifier can be devices that amplify the signal collected by the signal acquisition device such as the microphone and the accelerometer.
  • the power amplifier can be a device for amplifying the control signal so that the sound-generating device plays normally.
  • the vehicle machine/audio system/controller can include a computing unit for calculating the control signal.
  • the external camera can be placed in the following locations: the front of the tire, the car logo, the rearview mirror, the reflector, etc.
  • the external camera can obtain real-time road information and road condition information for the next few seconds (which can be determined in combination with the vehicle speed), or determine whether the vehicle is in a tunnel, rain, snow, and other special environments, and can assist in the selection of control parameters.
  • the navigation information can also be used to replace the function of the external camera, that is, the current road condition of the vehicle (highway, national highway, etc.) can be determined through the navigation information, and even the road condition of the current road can be determined through the navigation information (whether the road surface is a plate oil road, cement road, gravel road, etc.), so as to assist in the selection of control parameters.
  • the camera inside the car can be used to identify the user's head position and ear position to match the optimal control parameters and improve the user's driving experience.
  • the user's head position and ear position can also be obtained in other ways, for example, the user's head position and ear position can be determined by a laser sensor.
  • the vehicle control information may be obtained by the vehicle computer, and the vehicle control information may include at least one of the following information: tire pressure, vehicle speed, turning, acceleration, accelerator pedal opening, vehicle body posture, brake pedal opening, vehicle body tilt angle, etc.
  • the above information may affect the generation and propagation of road noise.
  • the vehicle body posture may cause the overall force of the vehicle body to be different, thereby affecting the size of the road noise.
  • the vehicle speed information in the vehicle control information combined with the use of the external camera, the road conditions on which the vehicle will travel in the future can be relatively accurately judged.
  • the noise reduction method since the information collected by the camera and/or the vehicle information is added as a reference when the control signal is output, the matching accuracy of the control parameters is improved, which is reflected in the user's driving experience and can improve the noise reduction effect under different vehicle conditions.
  • FIG5 is a flow chart of a noise reduction method provided by an embodiment of the present application.
  • the method 500 may be executed by the vehicle 100 in FIG1 , or may be executed by the computing platform 130 in the vehicle 100, or may be executed by a system-on-chip (SoC) in the computing platform 130, or may be executed by a processor in the computing platform 130.
  • SoC system-on-chip
  • the method 500 is introduced below with the vehicle as the execution subject, and the method 500 may include steps S501 to S502.
  • the reference signal may be collected by a first sensor, and the first sensor may be a sensor deployed near a noise source of the vehicle.
  • the reference signal may be used to indicate a noise signal collected near a noise source.
  • the noise generated by the noise source may be the noise generated by friction between the vehicle and the road surface or the air.
  • the noise generated by the noise source may also be the noise generated by friction between internal parts of the vehicle.
  • the first sensor may be an accelerometer or an acoustic sensor.
  • the first sensor When the first sensor is an accelerometer, it may be deployed at the bottom of the vehicle. Since the accelerometer can collect vibration signals at the bottom of the vehicle, the vibration signals collected by the accelerometer may be used as reference signals.
  • the first sensor When the first sensor is an acoustic sensor, it may be deployed at the spoiler, windshield or rearview mirror of the vehicle. Therefore, the signals collected by the acoustic sensor at the above positions may be used as reference signals.
  • the driving parameters of the vehicle may include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • the above driving parameters can be obtained by corresponding sensors on the vehicle, and the specific values of the above driving parameters can be displayed on the dashboard or display screen of the vehicle for the user to view.
  • the road condition information around the vehicle may be used to indicate the type of road on which the vehicle is currently traveling.
  • the current road type may include an asphalt road, a cement road, a gravel road, and the like.
  • step S501 the road condition information around the vehicle may be acquired through data collected by a third sensor, and the third sensor may be located on the vehicle.
  • the third sensor may include one or more sensors such as a DVR, a gravity sensor, an external camera (eg, a surround view camera), etc.
  • sensors such as a DVR, a gravity sensor, an external camera (eg, a surround view camera), etc.
  • the third sensor can be used to collect image information from the outside of the vehicle.
  • the third sensor may include a camera device, which can be deployed on the outside of the vehicle, for example, at the front of the tire, at the vehicle logo, at the rearview mirror, at the reflector, etc.
  • the camera device may include a monocular camera, a binocular camera, a structured light camera, a panoramic camera, etc.
  • the data collected by the third sensor may be image information obtained by the camera device, and the road condition information of the vehicle can be determined through the image information.
  • the image information may include static image information or video stream information.
  • the image information may be stored in the form of an image or video, or in the form of image or video parameters, for example, parameter information such as image brightness, grayscale, color distribution, contrast, pixels, etc.
  • the third sensor may be a DVR, and the road condition information of the vehicle may be determined based on the image information of the outside of the vehicle collected by the DVR.
  • the third sensor can be one or more sensors.
  • some sensors can collect image information outside the vehicle, and other sensors can collect image information inside the vehicle. The image information outside the vehicle and the image information inside the vehicle are combined to determine the road condition information of the vehicle.
  • the third sensor when the third sensor is a plurality of sensors, some sensors can collect image information of the outside of the vehicle, and other sensors can collect component data inside the vehicle.
  • the other sensors are gravity sensors, which can collect data reflecting the bumpy conditions of the user's seat.
  • the road condition information of the vehicle can be determined by combining the image information of the outside of the vehicle and the data reflecting the bumpy conditions of the user's seat.
  • the vehicle can automatically obtain road condition information around the vehicle in specific driving scenarios. For example, when the vehicle is driving on a gravel road, a concave cement road, an asphalt road, etc., the third sensor is automatically triggered to obtain the road condition information of the vehicle.
  • step S501 navigation information sent by a server may be obtained, and road condition information around the vehicle may be determined based on the navigation information.
  • classification information of the road eg, expressway, national highway, etc.
  • classification information of the road may be acquired through navigation information, thereby determining road condition information around the vehicle based on the classification information.
  • the noise reduction processing is performed by taking into account information such as the road conditions around the vehicle and the driving parameters of the vehicle.
  • the control signal played by the sound-emitting device can be approximately equal in magnitude and opposite in phase to the noise near the user's head or ears, thereby achieving a good noise reduction effect and improving the user's driving experience.
  • method 500 before step S502, also includes: obtaining position information of the ear or head of a user in the cabin of the vehicle, and step S502 includes: controlling the sound-emitting device of the vehicle to play a control signal based on a reference signal, road condition information around the vehicle, driving parameters of the vehicle and the position information of the ear or head of the user.
  • image information of the user's head or ears can be obtained by a camera device deployed inside the vehicle, so that the position of the user's ears or head can be determined based on the image information.
  • the position of the user's ear or head may be determined by a sensor with a distance measurement function (eg, a laser sensor, etc.).
  • a sensor with a distance measurement function eg, a laser sensor, etc.
  • the position information of the user's ear or head may be a photo or measurement data.
  • the position information of the user's ear or head may be measured position information or estimated position information, wherein the estimated position information may include the relative position or absolute position of the user's ear or head in the vehicle cabin, etc.
  • the control signal played by the sound-emitting device can more specifically reduce the noise near the user's head or ears, which can further improve the user's driving experience.
  • method 500 before step S502, also includes: obtaining an error signal, which is collected by a second sensor deployed near the user's ear or head; step S502 includes: controlling the sound-emitting device of the vehicle to play a control signal based on a reference signal, road condition information around the vehicle, driving parameters of the vehicle and the error signal.
  • the second sensor may be an error microphone, which may be disposed near the ear of the user or the head to collect noise signals near the ear or the head of the user.
  • the error microphone may be disposed near the headrest of the user's seat.
  • the error signal can be used to assist in controlling the sound-generating device of the vehicle to play the control signal, thereby achieving a better noise reduction effect of the control signal played by the sound-generating device.
  • step S502 includes: according to the reference signal, the road condition information around the vehicle and the driving parameters of the vehicle, Determine a control parameter, the control parameter including a coefficient of a filter; and control a sound generating device of a vehicle to play a control signal according to the control parameter and a reference signal.
  • a first control parameter can be obtained according to step S502.
  • a second control parameter can be obtained according to step S502.
  • the value of the second control parameter can be greater than the value of the first control parameter, so that the vehicle can achieve a good noise reduction effect when traveling on different road surfaces.
  • the third control parameter when the vehicle travels on the oil road at the second speed, the third control parameter can be obtained according to step S502.
  • the fourth control parameter can be obtained according to step S502.
  • the third speed is greater than the second speed, the value of the fourth control parameter can be greater than the value of the third control parameter, so that the vehicle can achieve a good noise reduction effect when traveling at different speeds.
  • the error signal or the position information of the user's ear or head may be used as a reference.
  • the sound-emitting device of the vehicle is controlled to play the control signal, including: determining the control signal according to the control parameters and the reference signal; using the sound-emitting device of the vehicle to play the control signal. Since the control signal played by the sound-emitting device is determined by the control parameters, and in the currently commonly used noise reduction schemes, the control parameters are determined only by the reference signal, and the optimal control parameters are often not obtained, which leads to poor noise reduction effect.
  • the control parameters when determining the control parameters, the road condition information around the vehicle and the driving parameters of the vehicle are taken into consideration, so that the determined control parameters are more accurate and the noise reduction effect is better.
  • step S502 includes: inputting a reference signal, road condition information around the vehicle, and driving parameters of the vehicle into a first model to obtain control parameters; wherein the first model is obtained based on a training sample, and the training sample includes: sample road condition information, sample reference signal, sample driving parameters, and sample control parameters.
  • the first model may be a deep neural network DNN model, which may also be composed of a conventional classification algorithm that can implement a classification function.
  • the process of training the first model may be: inputting sample reference signals, sample road condition information and sample driving parameters into the first model to obtain prediction control parameters; training the first model according to the prediction control parameters and the sample control parameters.
  • the prediction control parameters output by the first model can be continuously close to the sample control parameters, so that the first model can fit the best control parameters according to the input data in practical applications.
  • the sample control parameters can be the data labels corresponding to the prediction control parameters.
  • a reference signal, road condition information around the vehicle, and driving parameters of the vehicle can be input into a trained first model to obtain control parameters.
  • the control parameters can be obtained more efficiently and the determined control signal can be relatively more accurate.
  • FIG. 6 is a flow chart of a model training method provided in an embodiment of the present application.
  • Method 600 is a method 500.
  • the training process of the first model used in the method 600 is introduced.
  • the method 600 may include steps S601 to S605.
  • S601 obtaining secondary path information, a sample reference signal, and a sample error signal.
  • the secondary path information may be path information between the secondary source speaker and the error microphone
  • the sample reference signal may be used to indicate a noise signal near the noise source, which may be collected by a sensor deployed near the noise source
  • the sample error signal may be used to indicate a noise signal near the user's ear or head, which may be collected by an error sensor deployed near the user's ear or head.
  • the sensor deployed near the noise source may be the first sensor in method 500
  • the error sensor may be the second sensor in method 500 .
  • the control parameter may include a filter coefficient.
  • the secondary path information, the sample reference signal and the sample error signal can be input into a neural network model to obtain multiple groups of control parameters.
  • the neural network model can be a deep neural network DNN model, or a conventional classification algorithm.
  • the sample road condition information may be used to indicate the type of road on which the vehicle is currently traveling, for example, the road type may include an asphalt road, a cement road, a gravel road, etc.
  • the sample road condition information may be acquired by a camera device deployed outside the vehicle, for example, the camera device may be deployed at the front of the tire, the vehicle logo, the rearview mirror, the reflector, etc. of the vehicle.
  • the sample driving parameter includes at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • the above driving parameters can be obtained by corresponding sensors on the vehicle, and the specific values of the above driving parameters can be displayed on the dashboard or display screen of the vehicle for the user to view.
  • the target probability value can be a pre-set probability value, and the target probability value can be a data label corresponding to the model training.
  • the above target probability value may also be replaced by a score, that is, the first model may be trained by the predicted score and the target score of each set of control parameters.
  • the predicted probability value corresponding to each set of control parameters output by the first model can be continuously approached to the target probability value, so that the first model can obtain the optimal control parameters according to the input data in practical applications.
  • step S604 can be replaced by: inputting the sample reference signal, the sample road condition information and the sample driving parameter into the first model to obtain the predicted control parameter.
  • step S605 can be replaced by: training the first model according to the predicted control parameter and the sample control parameter.
  • the sample control parameter can be a data label corresponding to the prediction control parameter, and the sample control parameter can be obtained based on the sample reference signal, sample road condition information and sample driving parameter.
  • the predicted control parameters output by the first model can be made to continuously approach the sample control parameters, so that the first model can fit the optimal control parameters according to the input data in practical applications.
  • the sample control parameter may be a control parameter with a maximum probability value or a highest score in a first control parameter set, where the first control parameter set includes one or more control parameters.
  • the sample control parameter may be a control parameter with the largest probability value or the highest score in a second control parameter set, where the second control parameter set is obtained by classifying the first control parameter set.
  • the efficiency of model training can be improved during model training.
  • Figure 7 is a schematic block diagram of another noise reduction method provided in an embodiment of the present application.
  • the model training phase in method 700 is a detailed introduction to the model training process of method 600, and the model use phase in method 700 is a detailed introduction to the process of using the first model to determine the control parameters in method 500.
  • method 700 can be divided into a model training phase and a model use phase.
  • the vehicle can collect reference signals, error signals, camera signals, vehicle control signals, etc. through sensors, and use the error signals, reference signals and secondary path information to calculate control parameters, and all control parameters can constitute a control parameter set.
  • the control parameter set can be classified using different methods, and each type of control parameter set only saves one set of control parameters.
  • the vehicle can pre-train the algorithm model based on the reference signal, error signal (optional), camera signal, and vehicle control signal.
  • the algorithm model can be a classification algorithm or a preferential algorithm, which is not limited to a neural network algorithm or a non-neural network algorithm.
  • the algorithm model can select the optimal parameters from the control parameter set.
  • the above algorithm model can be the first model in method 500
  • the vehicle control signal can correspond to the driving parameters of the vehicle in method 500
  • the camera signal can be used to obtain the road condition information around the vehicle in method 500.
  • the vehicle can collect reference signals, error signals (optional), camera signals, vehicle control signals, and navigation information through sensors, and select the optimal control parameters based on the trained algorithm model. Then the control signal is obtained based on the reference signal and the control parameter, and the control signal is output through the sound device.
  • the step of optimizing the control parameters can be performed every once in a while, for example, every 100 milliseconds or 50 milliseconds.
  • obtaining the control signal using the reference signal and outputting it through the sound device requires calculation at each sampling point, and it needs to be performed continuously and in real time.
  • FIG8 is a flow chart of an algorithm model training of a denoising method provided in an embodiment of the present application.
  • Method 800 is a detailed introduction to the model training phase in method 700.
  • Method 800 may include steps S801 to S805.
  • the vehicle can control the secondary source speaker to play white noise and use the error microphone to receive the white noise.
  • the vehicle can obtain the secondary path information from the secondary source speaker to the error microphone by using the least mean square algorithm (LMS) or the Wiener filter algorithm.
  • LMS least mean square algorithm
  • corresponding secondary path information needs to be obtained between each secondary source speaker and the corresponding error microphone.
  • the secondary path is different when the position of the user's ear (head) is different, the secondary path can be measured multiple times and the position of the user's ear (head) at this time can be recorded in combination with the camera in the vehicle.
  • S802 Determine a control parameter and add the control parameter to a control parameter set.
  • the vehicle can collect reference signals, error signals, camera signals, vehicle control signals and other information, and can use LMS or Wiener filtering algorithm to combine the obtained reference signals, error signals and corresponding secondary path information to determine the control
  • the control parameter is added to the control parameter set A.
  • the signal collected by the camera deployed outside the vehicle can determine the road condition information around the vehicle, and through this road condition information, the vehicle can determine the current and future road conditions in a few seconds.
  • the vehicle can determine the position of the user's ear (head), thereby accurately determining the area that needs noise reduction.
  • the position of the user's ear (head) can also be obtained by other sensors, such as laser sensors.
  • the vehicle can classify the control parameter set A, replace the same or similar set of control parameters with a new parameter, and ensure that the noise reduction performance of the control parameter decreases less after the new parameter is replaced.
  • the control parameters replaced by the same parameter can be classified into the same category, and all new control parameters can form a set B.
  • the classification method may be: performing Fourier transform on the control parameters, classifying all sets with similar frequency domain amplitudes and phase amplitudes into the same category, and using the mean value Fourier transform result of the frequency domain as a substitute value for the new parameter.
  • the model may be a deep neural network DNN model, or may be composed of a conventional classification algorithm, which can implement a classification function.
  • the algorithm model may be the first model in method 600.
  • the model After the model design is completed, the collected reference signal, error signal, camera signal, vehicle control signal and other information are input into the model, and the model can output the probability values of different categories of control parameters in set B.
  • the information input to the model can be one or more of the following:
  • the driver’s ear (head) position information which can be obtained by (visual) algorithms from sensors such as cameras deployed in the vehicle.
  • Vehicle speed information Vehicle speed information, acceleration and deceleration pedal information, vehicle body tilt angle information, tire pressure and other information.
  • the above-mentioned reference signal can be a sample reference signal in method 600
  • the vehicle control signal can be a sample driving parameter in method 600
  • the camera signal can include the sample road condition information in method 600.
  • the target probability value of the corresponding optimal category of control parameters is 1, and the target probability values of other control parameters are 0.
  • the training process allows the output results of the neural network model to iterate and approach the above target probability values.
  • the optimal category of control parameters can be the sample control parameters in method 600, and the output results of the neural network model can be the predicted probability values in method 600.
  • the collected reference signal, error signal, camera signal, vehicle control signal and other information can be input into the algorithm model to obtain the predicted control parameters; then the algorithm model is trained based on the predicted control parameters and the corresponding sample control parameters (the sample control parameters belong to set B).
  • the training process is to allow the predicted control parameters output by the neural network model to iterate and approximate the corresponding sample control parameters.
  • the model can be trained using reference signals, error signals, camera signals, and vehicle control signals, so that the model can output optimal control parameters when actually used.
  • FIG9 is a flow chart of the use of the algorithm model of the denoising method provided in the embodiment of the present application.
  • the method 900 may be executed by the vehicle 100 in FIG1 , or by the computing platform 130 in the vehicle 100, or by the system on chip SoC in the computing platform 130, or by the processor in the computing platform 130.
  • the method 900 is a detailed introduction to the model use stage in the method 700.
  • the method 900 is introduced below with the vehicle as the execution subject.
  • the method 900 may include steps S901 to S905.
  • S901 collecting signals during the driving of the vehicle.
  • the signal may include: a reference signal, an error signal, a camera signal and a vehicle control signal.
  • the collected reference signal, error signal, camera signal, vehicle control signal, etc. are input into a trained algorithm model (for example, an algorithm model trained by method 800) to obtain control parameters.
  • a trained algorithm model for example, an algorithm model trained by method 800
  • S903 Determine a control signal according to the control parameter and the reference signal.
  • control parameter determined in step S902 is selected and combined with the collected reference signal to calculate the control signal.
  • the sound emitting device may be the secondary source speaker in method 800 .
  • control parameters are updated.
  • the reference signal collected in method 900 and step S904 requires uninterrupted calculation at each sampling point.
  • the other process can be calculated once at regular intervals to complete the update of control parameters.
  • the vehicle can use a trained algorithm model to obtain optimal control parameters.
  • the selected control parameters can be made more accurate, and the noise reduction effect can be improved under different vehicle conditions and when the vehicle conditions change.
  • Figure 10 is a flow chart of another algorithm model training of the denoising method provided in an embodiment of the present application.
  • Method 1000 is a detailed introduction to the model training stage in method 700.
  • Method 1000 may include steps S1001 to S1004.
  • the vehicle can control the secondary source speaker to play white noise and use the error microphone to receive the white noise.
  • the vehicle can obtain the secondary path information from the secondary source speaker to the error microphone by using the least mean square algorithm (LMS) or the Wiener filter algorithm.
  • LMS least mean square algorithm
  • corresponding secondary path information needs to be obtained between each secondary source speaker and the corresponding error microphone.
  • the secondary path is different when the position of the user's ear (head) is different, the secondary path can be measured multiple times and the position of the user's ear (head) at this time can be recorded in combination with the in-vehicle camera.
  • the vehicle can collect reference signals, error signals, camera signals, vehicle control signals and other information, and use LMS or Wiener filtering algorithm to combine the acquired reference signals, error signals and corresponding secondary path information to determine the control parameters, and add the control parameters to the control parameter set A.
  • the vehicle can determine the current and next few seconds of road conditions.
  • the vehicle can determine the position of the user's ear (head), thereby accurately determining the area that needs noise reduction.
  • the position of the user's ear (head) can also be obtained through other sensors, such as laser sensors.
  • the model may be a DNN model, or may be composed of a conventional classification algorithm, which can implement a classification function.
  • the algorithm model may be the first model in method 600 .
  • the model After the model is designed, the collected reference signal, error signal, camera signal, vehicle control signal and other information are input, and the model can output the probability values of different categories in set A.
  • the information input to the model can be one or more of the following:
  • the position information of the driver’s ear (head) which can be obtained by (visual) algorithms from sensors such as cameras deployed in the vehicle.
  • Vehicle speed information Vehicle speed information, acceleration and deceleration pedal information, vehicle body tilt angle information, tire pressure and other information.
  • the above-mentioned reference signal can be a sample reference signal in method 600
  • the vehicle control signal can be a sample driving parameter in method 600
  • the camera signal can include: the sample road condition information in method 600.
  • the target probability value of the corresponding optimal category of control parameters is 1, and the target probability values of other control parameters are 0.
  • the training process allows the output results of the neural network model to iterate and approach the above target probability values.
  • the optimal category of control parameters can be the sample control parameters in method 600, and the output results of the neural network model can be the predicted probability values in method 600.
  • the collected reference signal, error signal, camera signal, vehicle control signal and other information can be input into the algorithm model to obtain the predicted control parameters; then the algorithm model is trained based on the predicted control parameters and the corresponding sample control parameters (the sample control parameters belong to set A).
  • the training process is to allow the predicted control parameters output by the neural network model to iterate and approximate the corresponding sample control parameters.
  • the model can be trained using reference signals, error signals, camera signals, and vehicle control signals, so that the model can output optimal control parameters when actually used.
  • FIG11 is a flow chart of another algorithm model used in the noise reduction method provided in an embodiment of the present application.
  • the method 1100 may be executed by the vehicle 100 in FIG1 , or by the computing platform 130 in the vehicle 100, or by the system on chip SoC in the computing platform 130, or by the processor in the computing platform 130.
  • the method 1100 is a detailed introduction to the model use stage in the method 700.
  • the method 1100 is introduced below with the vehicle as the execution subject.
  • the method 1100 may include steps S1101 to S1105.
  • the vehicle collects signals while driving.
  • the signal may include information such as a reference signal, a camera signal, and a vehicle control signal.
  • the collected information such as reference signals, camera signals and vehicle control signals are input into a trained algorithm (for example, an algorithm trained by method 1000) to obtain control parameters.
  • a trained algorithm for example, an algorithm trained by method 1000
  • control parameter determined in step S1102 is selected and combined with the collected control signal to calculate the control signal.
  • the sound emitting device may be the secondary source speaker in method 1000 .
  • the reference signal collected in method 1100 and step S1104 requires uninterrupted calculation at each sampling point, and other processes of method 1100 can be calculated once at regular intervals to complete the update of the control parameters.
  • the trained algorithm model can be used to obtain the optimal control parameters.
  • the selected control parameters can be made more accurate, and the noise reduction effect can be improved under different vehicle conditions and when the vehicle conditions change.
  • method 1100 does not need to collect and use error signals during the application process, and can obtain the optimal control parameters more simply and efficiently.
  • the embodiment of the present application also provides a device for implementing any of the above methods, and the device includes a unit for implementing each step performed by the vehicle 100 in any of the above methods.
  • FIG12 is a schematic diagram of a denoising or model training device 1200 provided in an embodiment of the present application, and the device 1200 may include an acquisition unit 1210, a storage unit 1220, and a processing unit 1230.
  • the acquisition unit 1210 is used to acquire instructions and/or data, and the acquisition unit 1210 may also be referred to as a communication interface or a communication unit.
  • the storage unit 1220 is used to implement a corresponding storage function and store corresponding instructions and/or data.
  • the processing unit 1230 is used to perform data processing.
  • the processing unit 1230 can read the instructions and/or data in the storage unit so that the device 1200 implements the aforementioned denoising method or model training method.
  • the device 1200 includes: an acquisition unit 1210 and a processing unit 1230; the acquisition unit 1210 is used to acquire a reference signal, road condition information around a vehicle, and driving parameters of the vehicle, the reference signal is collected by a first sensor, and the first sensor is a sensor deployed near a noise source of the vehicle; the processing unit 1230 is used to control a sound-emitting device of the vehicle to play a control signal based on the reference signal, road condition information around the vehicle, and driving parameters of the vehicle, the control signal being used to reduce noise near the ears or head of a user, and the user is located in the cabin of the vehicle.
  • the acquisition unit 1210 is also used to obtain the position information of the ear or head of the user in the cabin of the vehicle; the processing unit 1230 is specifically used to control the sound device of the vehicle to play the control signal based on the reference signal, the road condition information around the vehicle, the driving parameters of the vehicle and the position information of the ear or head of the user.
  • the acquisition unit 1210 is also used to acquire an error signal, which is collected by a second sensor deployed near the user's ear or head; the processing unit 1230 is specifically used to control the sound-emitting device of the vehicle to play a control signal based on the reference signal, road condition information around the vehicle, driving parameters of the vehicle and the error signal.
  • the processing unit 1230 is specifically used to: determine control parameters based on a reference signal, road condition information around the vehicle and driving parameters of the vehicle, the control parameters including filter coefficients; and control a sound-emitting device of the vehicle to play a control signal based on the control parameters and the reference signal.
  • the processing unit 1230 is specifically used to input the reference signal, road condition information and driving parameters into the first model to obtain control parameters; wherein the first model is obtained based on training samples, and the training samples include: sample road condition information, sample reference signal, sample driving parameters and sample control parameters.
  • the processing unit 1230 is specifically used to obtain road condition information around the vehicle through data collected by a third sensor, where the third sensor is located on the vehicle and is used to collect image information outside the vehicle.
  • the acquisition unit 1210 is further used to acquire navigation information sent by the server; the processing unit 1230 is further used to determine the road condition information around the vehicle based on the navigation information.
  • the driving parameters around the vehicle include at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • the device 1200 includes: an acquisition unit 1210 and a processing unit 1230; the acquisition unit 1210 is used to acquire secondary path information, sample reference signals and sample error signals; the processing unit 1230 is used to determine multiple groups of control parameters according to the secondary path information, sample reference signals and sample error signals; the acquisition unit 1210 is also used to acquire sample road condition information and sample driving parameters; the processing unit 1230, is also used to: input the sample reference signal, the sample road condition information and the sample driving parameter into the first model to obtain the predicted probability value corresponding to each group of control parameters in the multiple groups of control parameters; train the first model according to the predicted probability value and the target probability value, wherein the target probability value is a pre-set probability value.
  • the device 1200 includes: an acquisition unit 1210 and a processing unit 1230; the acquisition unit 1210 is used to acquire secondary path information, a sample reference signal and a sample error signal; the processing unit 1230 is used to: input the sample reference signal, the sample road condition information and the sample driving parameters into the first model to obtain the predicted control parameters; and train the first model according to the predicted control parameters and the sample control parameters.
  • the sample control parameter is a control parameter with a maximum probability value in a first control parameter set, and the first control parameter set includes one or more control parameters.
  • the sample control parameter is a control parameter with a maximum probability value in a second control parameter set
  • the second control parameter set is obtained by classifying the first control parameter set
  • the sample driving parameter includes at least one of the following: tire pressure, vehicle speed, acceleration, accelerator pedal opening, brake pedal opening, vehicle body posture and vehicle body tilt angle.
  • the processing unit 1230 may be the processor 131 shown in FIG. 1 .
  • FIG13 is a schematic diagram of a noise reduction or model training device 1300 provided in an embodiment of the present application.
  • the device 1300 can be applied to the vehicle 100 of FIG1 .
  • the device 1300 includes: a memory 1310, a processor 1320, and a communication interface 1330.
  • the memory 1310, the processor 1320, and the communication interface 1330 are connected through an internal connection path, the memory 1310 is used to store instructions, and the processor 1320 is used to execute the instructions stored in the memory 1310 to control the communication interface 1330 to obtain information, or to enable the denoising device to perform the denoising method or model training method in the above-mentioned embodiments.
  • the memory 1310 can be coupled to the processor 1320 through an interface, or can be integrated with the processor 1320.
  • the processor 1320 stores one or more computer programs, and the one or more computer programs include instructions.
  • the noise reduction device 1300 executes the noise reduction method in each of the above embodiments.
  • each step of the above method can be completed by the hardware integrated logic circuit in the processor 1320 or the instruction in the form of software.
  • the method disclosed in conjunction with the embodiment of the present application can be directly embodied as a hardware processor for execution, or it can be executed by a combination of hardware and software modules in the processor.
  • the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc.
  • the storage medium is located in the memory 1310, and the processor 1320 reads the information in the memory 1310 and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it is not described in detail here.
  • the communication interface 1330 in FIG. 13 may implement the acquisition unit 1210 in FIG. 12
  • the memory 1310 in FIG. 13 may implement the storage unit 1230 in FIG. 12
  • the processor 1320 in FIG. 13 may implement the processing unit 1230 in FIG. 12 .
  • the device 1200 or the device 1300 may be a computing platform, and the computing platform may be a vehicle-mounted computing platform or a cloud computing platform.
  • the device 1200 or the device 1300 may be located in the vehicle 100 in FIG. 1 .
  • the device 1200 or the device 1300 may be the computing platform 130 in the vehicle of FIG. 1 .
  • An embodiment of the present application further provides a computer-readable medium, wherein the computer-readable medium stores a program code.
  • the computer program code When the computer program code is executed on a computer, the computer executes any one of the methods in FIG. 5 to FIG. 11 .
  • An embodiment of the present application also provides a chip, including: a circuit, wherein the circuit is used to execute any one of the methods in FIG. 5 to FIG. 11 above.
  • An embodiment of the present application also provides a vehicle, comprising any one of the devices shown in Figure 12 or Figure 13.
  • An embodiment of the present application further provides a noise reduction system, comprising any one of the noise reduction devices shown in FIG. 12 or FIG. 13 and one or more sound generating devices.
  • An embodiment of the present application also provides a noise reduction system, including a computing platform and one or more sound-emitting devices, wherein the computing platform is used to: obtain road condition information around a vehicle, driving parameters of the vehicle, and a reference signal collected by a first sensor, wherein the first sensor is a sensor deployed near a noise source of the vehicle; and control the one or more sound-emitting devices to play a control signal based on the reference signal, the road condition information around the vehicle, and the driving parameters of the vehicle, wherein the control signal is used to reduce noise near the ears or head of a user, wherein the user is located in the cabin of the vehicle.
  • a noise reduction system including a computing platform and one or more sound-emitting devices, wherein the computing platform is used to: obtain road condition information around a vehicle, driving parameters of the vehicle, and a reference signal collected by a first sensor, wherein the first sensor is a sensor deployed near a noise source of the vehicle; and control the one or more sound-emitting devices to play a control
  • the computing platform may be a computing platform 130 in a vehicle.
  • An embodiment of the present application further provides a computer program product, which includes a computer program.
  • the computer program When the computer program is executed, the computer executes any one of the methods in Figures 5 to 11 above.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art.
  • the computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • General Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Fittings On The Vehicle Exterior For Carrying Loads, And Devices For Holding Or Mounting Articles (AREA)

Abstract

一种降噪方法、装置以及运载工具，该方法包括：获取参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，参考信号由部署在噪声源附近处的第一传感器采集（S501）；根据参考信号、运载工具周围的路况信息和运载工具的行驶参数，控制运载工具的发声装置播放控制信号，用于降低用户耳朵或头部所在位置附近的噪声（S502）。该方法在降噪时结合运载工具周围的路况信息和运载工具的行驶参数等信息进行降噪处理。

Description

降噪方法、装置以及运载工具
本申请要求在2022年12月6日提交中国国家知识产权局、申请号为202211558434.4、申请名称为“降噪方法、装置以及运载工具”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及降噪领域，并且更具体地，涉及一种降噪方法、装置以及运载工具。
背景技术
车辆在高速行驶的过程中存在较强的噪声(包括：路噪、风噪和环境噪声等)，十分影响驾驶人员的舒适度。目前，比较常用的降噪方案是：获取噪声源附近的噪声信号作为参考信号，并根据该参考信号确定控制参数，进而得到控制信号。由于该控制信号通过扬声器播放后到达用户人耳时与用户人耳处原始噪声大小近似相同、相位近似相反，使得用户听到的噪声降低，实现了噪声消除。
然而，上述方法仅仅通过参考信号得到控制信号并经过扬声器播放该控制信号，降噪的效果较差，进而影响了用户的驾驶体验。
发明内容
本申请实施例提供一种降噪方法、装置以及运载工具，能够提高降噪效果，从而改善用户的驾驶体验。
第一方面，提供了一种降噪方法，该方法包括：获取参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，所述参考信号是第一传感器采集的，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
示例性地，参考信号可以用于指示噪声源附近处所采集到的噪声信号，例如，噪声源产生的噪声可以是运载工具与路面、空气摩擦产生的噪声，再例如，噪声源产生的噪声可以是运载工具内部零件之间摩擦产生的噪声。
可选地，第一传感器可以是加速度计或声传感器。
可选地，运载工具的行驶参数可以包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
可选地，运载工具周围的路况信息可以用于指示运载工具当前行驶的道路类型，例如，道路类型可以包括板油路、水泥路、石子路等。
可选地，发声装置可以包括：扬声器、音响、喇叭一种或多种。
本申请实施例中，在降噪时，由于考虑到了运载工具周围的路况信息和运载工具的行驶参数等信息进行降噪处理，能够提高降噪效果，从而改善用户的驾驶体验。
结合第一方面，在第一方面的某些实现方式中，该方法还包括：获取所述运载工具的座舱内所述用户的耳朵或头部的位置信息；所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述用户的耳朵或头部的位置信息，控制所述运载工具的发声装置播放所述控制信号。
可选地，可以通过部署在运载工具内部的摄像装置获取用户的头部或耳朵处的图像信息，从而确定用户耳朵或头部的位置信息。
可选地，可以通过具有测距功能的传感器(例如，激光传感器等)来确定用户的耳朵或头部的位置信息。
可选地，用户的耳朵或头部的位置信息可以是实测位置信息或估算位置信息，其中，估算位置信息 可以包括用户的耳朵或头部在座舱内的相对位置或绝对位置等等。
可选地，用户的耳朵或头部的位置信息可以是照片或者测量数据。
本申请实施例中，在降噪时，由于考虑到了运载工具座舱内用户头部或耳朵的位置信息，能够使得发声装置播放的控制信号更有针对性地降低用户头部或人耳附近的噪声，能够进一步的提高用户的驾驶体验。
结合第一方面，在第一方面的某些实现方式中，该方法还包括：获取误差信号，所述误差信号是部署在所述用户耳朵或头部所在位置附近的第二传感器采集的；所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述误差信号，控制所述运载工具的发声装置播放所述控制信号。
可选地，第二传感器可以是误差传声器，用于采集用户耳朵或头部附近的噪声信号。
本申请实施例中，在降噪时，可以通过误差信号辅助控制运载工具的发声装置播放控制信号，从而使得发声装置播放的控制信号的降噪效果更好。
结合第一方面，在第一方面的某些实现方式中，所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：根据所述参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，所述控制参数包括滤波器的系数；根据所述控制参数和所述参考信号，控制所述运载工具的发声装置播放所述控制信号。
可选地，在确定控制参数时，还可以以误差信号、用户耳朵或头部的位置信息作为参考。
可选地，根据控制参数和参考信号，控制运载工具的发声装置播放控制信号，包括：根据控制参数和参考信号，确定控制信号；使用运载工具的发声装置播放该控制信号。
由于发声装置播放的控制信号是由控制参数确定的，而目前常用的降噪方案中，仅仅通过参考信号确定控制参数，往往不能得到最佳的控制参数，进而导致降噪的效果差。本申请实施例中，在确定控制参数时，考虑了运载工具周围的路况信息和运载工具的行驶参数，使得确定的控制参数更加准确，降噪的效果也更好。
结合第一方面，在第一方面的某些实现方式中，所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，包括：将所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数输入到第一模型中，得到所述控制参数；其中，所述第一模型是基于训练样本得到的，所述训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
可选地，第一模型可以是深度神经网络(deep neural networks，DNN)模型，该模型也可以是由常规分类算法构成，该分类算法能够实现分类功能。
示例性地，第一模型训练的过程可以是：将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到预测控制参数；根据预测控制参数和样本控制参数训练第一模型。这样，能够使得第一模型输出的预测控制参数不断接近样本控制参数，从而使得第一模型在实际应用中能够根据输入的数据拟合出最佳的控制参数。其中，该样本控制参数可以是预测控制参数对应的数据标签。
本申请实施例中，可以将参考信号、运载工具周围路况信息和运载工具的行驶参数输入到训练好的第一模型中，从而得到控制参数，通过这样的方式，得到控制参数的效率更高，确定的控制信号也相对更准确。
结合第一方面，在第一方面的某些实现方式中，所述获取运载工具周围的路况信息，包括：通过第三传感器采集的数据获取所述运载工具周围的路况信息，所述第三传感器位于所述运载工具上。
可选地，该第三传感器可以包括行车记录仪(digital video recorder，DVR)、重力传感器、车外摄像头(例如，环视摄像头)等传感器中的一种或多种。
可选地，该第三传感器可以用于采集运载工具外部的图像信息。例如，第三传感器可以包括摄像装置，该摄像装置可以部署在运载工具的外部，例如，轮胎前部、车标处、后视镜处、反光镜处等等。该摄像装置可以包括单目相机、双目相机、结构光相机以及全景相机等。第三传感器采集的数据可以是摄像装置获取的图像信息，通过该图像信息能够确定运载工具行驶的路况信息。其中，图像信息可以包括静态图像信息，也可以包括视频流信息。图像信息可以以图像或视频的形式存储，也可以以图像或视频的参数的形式存储，例如，图像的亮度、灰度、色彩分布、对比度、像素等参数信息。又例如，第三传感器可以是DVR，可以根据DVR采集的运载工具外部的图像信息确定运载工具行驶的路况信息。
可选地，该第三传感器可以是一个或多个传感器，在第三传感器是多个传感器时，一部分传感器可以采集运载工具外部的图像信息，另一部分传感器可以采集运载工具内部的图像信息，结合运载工具外部的图像信息和运载工具内部的图像信息可以确定运载工具行驶的路况信息。
可选地，在第三传感器是多个传感器时，一部分传感器可以采集运载工具外部的图像信息，另一部分传感器可以采集运载工具内部的部件数据，例如，该另一部分传感器是重力传感器，其可以采集反映用户座椅上颠簸情况的数据，结合运载工具外部的图像信息和反映用户座椅颠簸情况的数据可以确定运载工具行驶的路况信息。
可选地，运载工具可以在特定的驾驶场景下，自动获取运载工具周围的路况信息，例如，车辆行驶在石子路、具有凹陷的水泥路，板油路等等特定的场景下。
本申请实施例中，可以通过第三传感器采集的数据获取运载工具周围的路况信息，从而使得运载工具能够根据该路况信息，控制发声装置播放控制信号。
结合第一方面，在第一方面的某些实现方式中，所述获取运载工具周围的路况信息，包括：获取服务器发送的导航信息；根据所述导航信息确定所述运载工具周围的路况信息。
可选地，可以通过导航信息获取当前行驶的路分类(例如，高速路、国道等)，从而确定运载工具周围的路况信息。
本申请实施例中，可以使用导航信息获取路况信息，从而使得运载工具能够根据该路况信息，控制发声装置播放控制信号。
第二方面，提供了一种模型训练方法，该方法包括：获取次级路径信息、样本参考信号和样本误差信号；根据所述次级路径信息、所述样本参考信号和所述样本误差信号，确定多组控制参数；获取样本路况信息和样本行驶参数；将所述样本参考信号、所述样本路况信息和所述样本行驶参数输入到第一模型中，得到多组控制参数中每一组控制参数对应的预测概率值；根据所述预测概率值和目标概率值训练第一模型，其中，所述目标概率值是预先设置的概率值。
可选地，上述目标概率值也可以替换为得分，即可以通过每一组控制参数的预测得分和目标得分，训练第一模型。
本申请实施例中，能够使得第一模型输出的每一组控制参数对应的预测概率值不断接近目标概率值，从而使得第一模型在实际应用中能够根据输入的数据得到最佳的控制参数。
第三方面，提供了一种模型训练方法，该方法包括：获取次级路径信息、样本参考信号和样本误差信号；根据所述次级路径信息、所述样本参考信号和所述样本误差信号，确定样本控制参数；获取运载工具周围的样本路况信息和所述运载工具的样本行驶参数；将所述样本参考信号、所述样本路况信息和所述样本行驶参数输入到第一模型中，得到预测控制参数；根据所述预测控制参数和样本控制参数，训练所述第一模型。
其中，该样本控制参数可以是预测控制参数对应的数据标签。
本申请实施例中，能够使得第一模型输出的预测控制参数不断接近样本控制参数，从而使得第一模型在实际应用中能够根据输入的数据拟合出最佳的控制参数。
结合第三方面，在第三方面的某些实现方式中，所述样本控制参数是第一控制参数集中概率值最大的控制参数，所述第一控制参数集中包括一个或多个控制参数。
可选地，该样本控制参数也可以是第一控制参数集中得分最高的控制参数。
结合第三方面，在第三方面的某些实现方式中，所述样本控制参数是第二控制参数集中概率值最大的控制参数，所述第二控制参数集是对所述第一控制参数集分类得到的。
可选地，所述样本控制参数也可以是第二控制参数集中得分最高的控制参数。
结合第三方面，在第三方面的某些实现方式中，所述样本行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
第四方面，提供了一种降噪装置，该装置包括：获取单元和处理单元；所述获取单元，用于获取参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，所述参考信号是第一传感器采集的，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；所述处理单元，用于根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
结合第四方面，在第四方面的某些实现方式中，所述获取单元，还用于获取所述运载工具的座舱内 所述用户的耳朵或头部的位置信息；所述处理单元，具体用于根据所述参考信号、运载工具周围的路况信息、所述运载工具的行驶参数和所述用户的耳朵或头部的位置信息，控制所述运载工具的发声装置播放所述控制信号。
结合第四方面，在第四方面的某些实现方式中，所述获取单元，还用于获取误差信号，所述误差信号是部署在所述用户耳朵或头部所在位置附近的第二传感器采集的；所述处理单元，具体用于根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述误差信号，控制所述运载工具的发声装置播放所述控制信号。
结合第四方面，在第四方面的某些实现方式中，所述处理单元，具体用于：根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，所述控制参数包括滤波器的系数；根据所述控制参数和所述参考信号，控制所述运载工具的发声装置播放所述控制信号。
结合第四方面，在第四方面的某些实现方式中，所述处理单元，具体用于将所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数输入到第一模型中，得到所述控制参数；其中，所述第一模型是基于训练样本得到的，所述训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
结合第四方面，在第四方面的某些实现方式中，所述处理单元，具体用于通过第三传感器采集的数据获取所述运载工具周围的路况信息，所述第三传感器位于所述运载工具上，所述第三传感器用于采集所述运载工具外部的图像信息。
结合第四方面，在第四方面的某些实现方式中，所述获取单元，还用于获取服务器发送的导航信息；所述处理单元，还用于根据所述导航信息，确定所述运载工具周围的路况信息。
结合第四方面，在第四方面的某些实现方式中，所述运载工具周围的行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
第五方面，提供一种模型训练装置，该装置包括：获取单元和处理单元；所述获取单元用于获取次级路径信息、样本参考信号和样本误差信号；所述处理单元，用于根据所述次级路径信息、所述样本参考信号和所述样本误差信号，确定多组控制参数；所述获取单元，还用于获取样本路况信息和样本行驶参数；所述处理单元，还用于将所述样本参考信号、所述样本路况信息和所述样本行驶参数输入到第一模型中，得到所述多组控制参数中每一组控制参数对应的预测概率值；根据所述预测概率值和所述目标概率值，训练所述第一模型，其中，所述目标概率值是预先设置的概率值。
第六方面，提供一种模型训练装置，该装置包括：获取单元和处理单元；所述获取单元用于获取次级路径信息、样本参考信号和样本误差信号；所述处理单元，用于：将所述样本参考信号、所述样本路况信息和所述样本行驶参数输入到第一模型中，得到预测控制参数；根据所述预测控制参数和样本控制参数，训练所述第一模型。
结合第六方面，在第六方面的某些实现方式中，所述样本控制参数是第一控制参数集中概率值最大的控制参数，所述第一控制参数集中包括一个或多个控制参数。
结合第六方面，在第六方面的某些实现方式中，所述样本控制参数是第二控制参数集中概率值最大的控制参数，所述第二控制参数集是对所述第一控制参数集分类得到的。
结合第六方面，在第六方面的某些实现方式中，所述样本行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
第七方面，提供一种降噪装置，该装置包括：至少一个处理器和存储器，所述至少一个处理器与所述存储器耦合，用于读取并执行所述存储器中的指令，使得该装置实现上述第一方面中任意一种实现方式的方法。
第八方面，提供一种模型训练装置，该装置包括：至少一个处理器和存储器，所述至少一个处理器与所述存储器耦合，用于读取并执行所述存储器中的指令，使得该装置实现上述第二方面或第三方面中任意一种实现方式的方法。
第九方面，提供一种计算机可读介质，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面至第三方面中任意一种实现方式的方法。
第十方面，供一种计算机程序产品，所述计算机产品包括计算机程序，当所述计算机程序被运行时，使得计算机执行上述第一方面至第三方面中任意一种实现方式的方法。
第十一方面，提供一种芯片，该芯片包括电路，该电路用于执行上述第一方面至第三方面中任意一 种实现方式的方法。
第十二方面，提供了一种降噪系统，包括：上述第四方面中任意一种实现方式中的降噪装置以及一个或多个发声装置。
第十三方面，提供了一种降噪系统，包括计算平台以及一个或多个发声装置，所述计算平台用于：获取运载工具周围的路况信息、所述运载工具的行驶参数和第一传感器采集的参考信号，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述一个和多个发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
第十四方面，提供了一种服务器，所述服务器包括上述第五方面或第六方面任意一种实现方式中的模型训练装置。
第十五方面，提供了一种运载工具，包括：上述第四方面中任意一种实现方式中的降噪装置，或者，包括上述十二方面或十三方面所述的降噪系统。
结合第十五方面，在第十五方面的某些实现方式中，该运载工具为车辆。
本申请实施例提供的技术方案能够在降噪时，考虑到运载工具周围的路况信息和运载工具的行驶参数等信息进行降噪处理，能够提高降噪效果，从而改善用户的驾驶体验。本方案还可以在降噪时，考虑到用户耳朵或头部所在位置以及误差信号，使得发声装置播放的控制信号更有针对性地降低用户头部或人耳附近的噪声，进一步的提高用户的驾驶体验。此外，在控制参数的选择上，本方案可以通过第一模型快速确定控制参数，这样得到控制参数的效率更高，确定的控制信号也相对更准确。
附图说明
图1是本申请实施例提供的一种运载工具的功能性示意图；
图2是本申请实施例提供一种降噪方法的前馈方案；
图3是本申请实施例提供一种降噪方法的反馈方案；
图4是本申请实施例提供的降噪方法所适用的系统架构；
图5是本申请实施例提供的一种降噪方法流程性示意图；
图6是本申请实施例提供的一种模型训练方法流程性示意图；
图7是本申请实施例提供的另一种降噪方法的示意性框图；
图8是本申请实施例提供的降噪方法的一种算法模型训练的流程性示意图；
图9是本申请实施例提供的降噪方法的一种算法模型使用流程性示意图；
图10是本申请实施例提供的降噪方法的另一种算法模型训练的流程性示意图；
图11是本申请实施例提供的降噪方法的另一种算法模型使用流程性示意图；
图12是本申请实施例提供的一种降噪或模型训练装置；
图13是本申请实施例提供的另一种降噪或模型训练装置。
具体实施方式
下面将结合附图，对本申请实施例中的技术方案进行描述。
图1是本申请实施例提供的运载工具100的一个功能性示意图。应理解，图1及相关描述仅为一种举例，并不对本申请实施例中的运载工具进行限定。
在实施过程中，运载工具100可以被配置为完全或部分自动驾驶模式，也可以由用户进行人工驾驶。例如：运载工具100可以通过感知系统120获取其周围的环境信息，并基于对周边环境信息的分析得到自动驾驶策略以实现完全自动驾驶，或者将分析结果呈现给用户以实现部分自动驾驶。
运载工具100可包括多种子系统，例如感知系统120、计算平台130和发声装置140。可选地，该运载工具100可包括更多或更少的子系统，并且每个子系统都可包括一个或多个部件。另外，该运载工具100的每个子系统和部件可以通过有线或者无线的方式实现互连。
感知系统120可包括用于感测关于运载工具100周边的环境的信息的若干种传感器。例如，感知系统120可以包括定位系统，该定位系统可以是全球定位系统(global positioning system，GPS)，也可以是北斗系统或者其他定位系统。感知系统120可以包括惯性测量单元(inertial measurement unit，IMU)等、激光雷达、毫米波雷达、超声波雷达以及摄像装置中的一种或者多种。
运载工具100的部分或所有功能可以由计算平台130控制。计算平台130可包括处理器131至13n(n为正整数)，处理器是一种具有信号的处理能力的电路，在一种实现中，处理器可以是具有指令读取与运行能力的电路，例如中央处理单元(central processing unit，CPU)、微处理器、图形处理器(graphics processing unit，GPU)(可以理解为一种微处理器)、或数字信号处理器(digital signal processor，DSP)等；在另一种实现中，处理器可以通过硬件电路的逻辑关系实现一定功能，该硬件电路的逻辑关系是固定的或可以重构的，例如处理器为专用集成电路(application-specific integrated circuit，ASIC)或可编程逻辑器件(programmable logic device，PLD)实现的硬件电路，例如FPGA。在可重构的硬件电路中，处理器加载配置文档，实现硬件电路配置的过程，可以理解为处理器加载指令，以实现以上部分或全部单元的功能的过程。此外，处理器还可以是针对人工智能设计的硬件电路，其可以理解为一种ASIC，例如神经网络处理单元(neural network processing unit，NPU)、张量处理单元(tensor processing unit，TPU)、深度学习处理单元(deep learning processing unit，DPU)等。此外，计算平台130还可以包括存储器，存储器用于存储指令，处理器131至13n中的部分或全部处理器可以调用存储器中的指令，以实现相应的功能。
计算平台130可基于从各种子系统(例如，感知系统120)接收的输入来控制运载工具100的功能。在一些实施例中，计算平台130可用于对运载工具100及其子系统的许多方面提供控制。
发声装置140可以包括：扬声器、音响、喇叭等等，用于播放控制信号，从而减少用户耳朵或头部附近的噪声。
可选地，上述组件只是一个示例，实际应用中，上述各个模块中的组件有可能根据实际需要增添或者删除，图1不应理解为对本申请实施例的限制。
在道路行进的运载工具100，可以识别其周围环境内的物体以确定对当前速度的调整。所述物体可以是其它车辆、交通控制设备、或者其它类型的物体。在一些示例中，可以独立地考虑每个识别的物体，并且基于物体的各自的特性，诸如物体的当前速度、加速度、与运载工具的间距等，可以用来确定运载工具100所要调整的速度。
可选地，运载工具100或者与运载工具100相关联的感知和计算设备(例如，计算平台130)可以基于所识别的物体的特性和周围环境的状态(例如，交通、雨、道路上的冰、等等)来预测所述识别的物体的行为。可选地，每一个所识别的物体都依赖于彼此的行为，因此还可以将所识别的所有物体全部一起考虑来预测单个识别的物体的行为。
本申请中的运载工具100可以包括：路上交通工具、水上交通工具、空中交通工具、工业设备、农业设备、或娱乐设备等。例如，运载工具100可以为车辆，该车辆为广义概念上的车辆，可以是交通工具(如商用车、乘用车、摩托车、飞行车、火车等)，工业车辆(如：叉车、挂车、牵引车等)，工程车辆(如挖掘机、推土车、吊车等)，农用设备(如割草机、收割机等)，游乐设备，玩具车辆等，本申请实施例对车辆的类型不作具体限定。再如，运载工具100可以为飞机、或轮船等交通工具。
以下以运载工具为车辆为例，说明本申请需要解决的技术问题以及所采用的技术方案。
车辆在高速行驶的过程中存在较强的噪声(包括：路噪、风噪和环境噪声等)，十分影响驾驶人员的舒适度。其中，路噪由车胎与地面、悬架与车身的碰撞产生，并通过车架传递至座舱。在车架部分位置设置加速度计，能够采集车辆的震动情况，加速度计产生的信号可以作为车辆主动噪声控制(active noise control，ANC)系统的参考信号。由于车辆的结构复杂，噪声传播的路径往往难以估计，并且常常受到工况的影响而变化，导致ANC系统的降噪变得困难。
可以采用前馈方案或反馈方案进行降噪，前馈方案与反馈方案的区别在于：前馈方案需要采集参考信号来确定控制参数，而反馈方案不需要采集参考信号确定控制参数。图2和图3分别是前馈方案和反馈方案的模型框图，图2和图3所示的前馈方案和反馈方案在实施的过程中，均需要对框图中的控制参数进行调整，这个控制参数可以是有限长滤波响应(finite impulse response，FIR)滤波器的系数，该FIR滤波器在ANC系统控制噪声的过程中，每个采样点都需要实时进行计算，而对控制参数的调整、更新虽然不需要实时计算，但是需要提前计算好一组或多组控制参数，在运行过程中选取合适的一组控制参数来实时滤波。由于提前计算的控制参数很难满足所有的场景，在一些特殊的场景下，可能无法选择最优的控制参数，这样，通过控制参数确定的控制信号在通过扬声器播放后，降噪效果不佳，进而影响了用户的驾驶体验。
本申请实施例提供了一种降噪方法、装置以及运载工具，能够提高降噪效果，从而改善用户的驾驶 体验。在介绍本申请实施例提供的降噪方法之前，首先介绍降噪方法所适用的系统架构。
图4是本申请实施例提供的降噪方法所适用的系统架构。
如图4所示，该系统架构400包括：加速度计、误差传声器、次级源扬声器、前置放大器、加速度信号放大器、功率放大器、车机/音频系统/控制器、摄像头和车机。
其中，对于路噪而言，加速度计用于采集车底的震动信号作为参考信号，该震动信号可以是靠近噪声源处的信号，对于风噪而言，可以用加速度计采集扰流板、挡风玻璃、后视镜处的信号作为参考信号，或者传声器也可以代替加速度计采集参考信号。误差传声器可以用于采集人耳附近的噪声信号。次级源扬声器可以是发出反向噪声信号来抵消或降低原始噪声的设备，次级源扬声器可以包括一个或多个扬声器，其可以布置在用户耳朵旁、车辆的头枕处、座椅处或门板处等位置，该次级源扬声器可以是发声装置中的一种。前置放大器和加速度信号放大器可以是给传声器、加速度计等信号采集设备采集的信号放大的设备。功率放大器可以是用于放大控制信号，从而使得发声装置正常播放的设备。车机/音频系统/控制器可以包括计算单元，用于计算控制信号。
在摄像头的布置上，车外摄像头可以布置在以下几个位置：轮胎前部、车标处、后视镜处、反光镜处等等。车外摄像头能够获取实时路面信息和未来数秒的路况信息(可以结合车速确定)，或者确定车辆是否处于隧道、下雨、积雪等特殊的环境下，能够辅助进行控制参数的选择。如果不设置车外摄像头，也可以使用导航信息代替实现车外摄像头的功能，即通过导航信息确定车辆当前行驶的路况(高速路、国道等)，甚至可以通过导航信息判断当前道路的路况(路面是否为板油路、水泥路、石子路等)，从而辅助进行控制参数的选择。车内摄像头可以用于识别用户头部位置以及耳朵位置，以匹配最优控制参数，提高用户的驾驶体验。此外，用户的头部位置以及耳朵位置的获取也可以采取其他方式，例如，可以通过激光传感器确定用户的头部位置以及耳朵位置。
车控信息可以是车机获取的，车控信息可以包括以下信息至少一种：车胎压、车速、转弯、加速度、加速踏板开度、车身姿态、刹车踏板开度和车身倾斜角等等。以上信息可以影响路噪的产生与传播，例如，车身姿态能够使得车身的整体受力不同，从而影响路噪的大小。通过获取上述车控信息，有利于车辆选择最优的控制参数，此外，利用车控信息中的车速信息结合车外摄像头的使用，能够相对准确的判断出车辆未来行驶的路面情况。
本申请实施例提供的降噪方法的实际应用中，由于控制信号输出时增设了摄像头采集的信息和/或车机信息作为参考，改善了控制参数的匹配精准度，反映到用户的驾驶体验上是可以在不同的车况下改善了降噪的效果。
图5是本申请实施例提供的一种降噪方法流程性示意图，方法500可以由图1中的运载工具100执行，或者也可以由运载工具100中的计算平台130执行，或者还可以由计算平台130中的片上系统(system-on-chip，SoC)执行，或者，还可以由计算平台130中的处理器执行。下面以运载工具作为执行主体介绍方法500，方法500可以包括步骤S501至S502。
S501，获取参考信号，运载工具周围的路况信息和运载工具的行驶参数。
其中，参考信号可以是第一传感器采集的，该第一传感器可以是部署在所述运载工具的噪声源附近处的传感器。参考信号可以用于指示噪声源附近所采集的噪声信号，例如，噪声源产生的噪声可以是运载工具与路面、空气摩擦产生的噪声，再例如，噪声源产生的噪声也可以是运载工具内部零件之间摩擦产生的噪声。
可选地，第一传感器可以是加速度计或声传感器。其中，当第一传感器是加速度计时，其可以部署在运载工具的底部，由于加速度计可以采集运载工具底部的震动信号，因此，可以将加速度计采集的震动信号作为参考信号。当第一传感器是声传感器时，其可以部署在运载工具的扰流板、挡风玻璃或者后视镜位置，因此，可以将声传感器在上述位置处采集的信号作为参考信号。
可选地，运载工具的行驶参数可以包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。上述行驶参数可以通过运载工具上相应的传感器获取，上述行驶参数具体的数值可以显示在运载工具的仪表盘或者显示屏上供用户查看。
可选地，运载工具周围的路况信息可以用于指示运载工具当前行驶的道路类型，例如，当前道路类型可以包括板油路、水泥路、石子路等。
一个实施例中，在步骤S501中，可以通过第三传感器采集的数据获取运载工具周围的路况信息，该第三传感器可以位于运载工具上。
可选地，该第三传感器可以包括DVR、重力传感器、车外摄像头(例如，环视摄像头)等传感器中的一种或多种。
可选地，该第三传感器可以用于采集所述运载工具外部的图像信息。例如，第三传感器可以包括摄像装置，该摄像装置可以部署在运载工具的外部，例如，轮胎前部、车标处、后视镜处、反光镜处等等。该摄像装置可以包括单目相机、双目相机、结构光相机以及全景相机等。第三传感器采集的数据可以是摄像装置获取的图像信息，通过该图像信息能够确定运载工具行驶的路况信息。其中，图像信息可以包括静态图像信息，也可以包括视频流信息。图像信息可以以图像或视频的形式存储，也可以以图像或视频的参数的形式存储，例如，图像的亮度、灰度、色彩分布、对比度、像素等参数信息。又例如，第三传感器可以是DVR，可以根据DVR采集的运载工具外部的图像信息确定运载工具行驶的路况信息。
可选地，该第三传感器可以是一个或多个传感器，在第三传感器是多个传感器时，一部分传感器可以采集运载工具外部的图像信息，另一部分传感器可以采集运载工具内部的图像信息，结合运载工具外部的图像信息和运载工具内部的图像信息可以确定运载工具行驶的路况信息。
可选地，在第三传感器是多个传感器时，一部分传感器可以采集运载工具外部的图像信息，另一部分传感器可以采集运载工具内部的部件数据，例如，该另一部分传感器是重力传感器，其可以采集反映用户座椅上颠簸情况的数据，结合运载工具外部的图像信息和反映用户座椅颠簸情况的数据可以确定运载工具行驶的路况信息。
可选地，运载工具可以在特定的驾驶场景下，自动获取运载工具周围的路况信息，例如，运载工具行驶在石子路、具有凹陷的水泥路，板油路等等特定的场景下自动触发第三传感器获取运载工具作为的路况信息。
一个实施例中，在步骤S501中，可以通过获取服务器发送的导航信息，并根据导航信息确定运载工具周围的路况信息。
可选地，可以通过导航信息获取运载工具当前行驶的道路(例如，高速路、国道等)的分类信息，从而基于该分类信息确定运载工具周围的路况信息。
S502，根据参考信号，运载工具周围的路况信息和运载工具的行驶参数，控制运载工具的发声装置播放控制信号。
本申请实施例中，在降噪时，由于考虑到了运载工具周围的路况信息和运载工具的行驶参数等信息进行降噪处理，这样能够使得发声装置播放的控制信号与用户头部或者耳朵附近处的噪声大小近似相等、相位相反，从而能够达到良好降噪的效果，改善用户的驾驶体验。
一个实施例中，在步骤S502之前，方法500还包括：获取运载工具的座舱内用户的耳朵或头部的位置信息，步骤S502包括：根据参考信号、运载工具周围的路况信息、运载工具的行驶参数和用户耳朵或头部的位置信息，控制运载工具的发声装置播放控制信号。
可选地，可以通过部署在运载工具内部的摄像装置获取用户的头部或耳朵处的图像信息，从而可以基于该图像信息确定用户耳朵或头部所在位置。
可选地，可以通过具有测距功能的传感器(例如，激光传感器等)来确定用户耳朵或头部所在位置。
示例性地，上述用户的耳朵或头部的位置信息可以是照片或者测量数据。
示例性地，用户的耳朵或头部的位置信息可以是实测位置信息或估算位置信息，其中，估算位置信息可以包括用户的耳朵或头部在运载工具座舱内的相对位置或绝对位置等等。
本申请实施例中，在降噪时，由于考虑到了运载工具座舱内用户头部或耳朵的位置信息，能够使得发声装置播放的控制信号更有针对性地降低用户头部或人耳附近的噪声，能够进一步的提高用户的驾驶体验。
一个实施例中，在步骤S502之前，方法500还包括：获取误差信号，该误差信号是部署在用户耳朵或头部所在位置附近的第二传感器采集的；步骤S502包括：根据参考信号、运载工具周围的路况信息、运载工具的行驶参数和误差信号，控制运载工具的发声装置播放控制信号。
可选地，第二传感器可以是误差传声器，其可以部署在用户或头部耳朵附近，用于采集用户耳朵或头部附近的噪声信号。例如，误差传声器可以部署在用户的座椅头枕附近。
本申请实施例中，在降噪时，可以通过误差信号辅助控制运载工具的发声装置播放控制信号，从而使得发声装置播放的控制信号降噪效果更好。
一个实施例中，步骤S502包括：根据参考信号、运载工具周围的路况信息和运载工具的行驶参数， 确定控制参数，该控制参数包括滤波器的系数；根据控制参数和参考信号，控制运载工具的发声装置播放控制信号。
例如，运载工具以第一速度行驶在板油路上时，根据步骤S502可以得到第一控制参数，运载工具以第一速度行驶在石子路上时，根据步骤S502可以得到第二控制参数，此时，第二控制参数的数值可以大于第一控制参数的数值，从而使得运载工具行驶在不同的路面上都能达到良好的降噪效果。
又例如，运载工具以第二速度行驶在板油路上时，根据步骤S502可以得到第三控制参数，运载工具以第三速度行驶在板油路上时，根据步骤S502可以得到第四控制参数，当第三速度大于第二速度时，第四控制参数的数值可以大于第三控制参数的数值，从而使得运载工具行驶在不同的行驶速度时都能达到良好的降噪效果。
可选地，在确定控制参数时，还可以以误差信号、用户耳朵或头部的位置信息作为参考。
可选地，根据控制参数和参考信号，控制运载工具的发声装置播放控制信号，包括：根据控制参数和参考信号，确定控制信号；使用运载工具的发声装置播放该控制信号。由于发声装置播放的控制信号是由控制参数确定的，而目前常用的降噪方案中，仅仅通过参考信号确定控制参数，往往不能得到最佳的控制参数，进而导致降噪的效果差。本申请实施例中，在确定控制参数时，考虑了运载工具周围的路况信息和运载工具的行驶参数，使得确定的控制参数更加准确，降噪的效果也更好。
一个实施例中，步骤S502包括：将参考信号、运载工具周围的路况信息和运载工具的行驶参数输入到第一模型中，得到控制参数；其中，该第一模型是基于训练样本得到的，该训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
可选地，第一模型可以是深度神经网络DNN模型，该模型也可以是由常规分类算法构成，该分类算法能够实现分类功能。
示例性地，第一模型训练的过程可以是：将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到预测控制参数；根据预测控制参数和样本控制参数训练第一模型。这样，能够使得第一模型输出的预测控制参数不断接近样本控制参数，从而使得第一模型在实际应用中能够根据输入的数据拟合出最佳的控制参数。其中，该样本控制参数可以是预测控制参数对应的数据标签。
本申请实施例中，可以将参考信号、运载工具周围的路况信息和运载工具的行驶参数输入到训练好的第一模型中，从而得到控制参数，通过这样的方式，得到控制参数的效率更高，确定的控制信号也相对更准确。
图6是本申请实施例提供的一种模型训练方法流程性示意图，方法600是对方法500
中使用的第一模型的训练过程进行介绍。方法600可以包括步骤S601至S605。
S601，获取次级路径信息、样本参考信号和样本误差信号。
其中，次级路径信息可以是次级源扬声器与误差麦克风之间的路径信息，样本参考信号可以用于指示噪声源附近处的噪声信号，其可以是部署在噪声源附近的传感器采集的。样本误差信号可以用于指示用户耳朵或头部附近的噪声信号，其可以是部署在用户耳朵或头部附近的误差传感器采集的。
上述部署在噪声源附近处的传感器可以是方法500中的第一传感器，误差传感器可以是方法500中的第二传感器。
S602，根据次级路径信息、样本参考信号和样本误差信号，确定多组控制参数。
其中，该控制参数可以包括滤波器的系数。
示例性地，可以将次级路径信息、样本参考信号和样本误差信号输入到神经网络模型中，得到多组控制参数。上述神经网络模型可以是深度神经网络DNN模型，该神经网络模型也可以是常规的分类算法。
S603，获取运载工具周围的样本路况信息和运载工具的样本行驶参数。
可选地，该样本路况信息可以用于指示运载工具当前行驶的道路类型，例如，道路类型可以包括板油路、水泥路、石子路等。该样本路况信息可以通过部署在运载工具外部的摄像装置获取，例如，摄像装置可以部署在运载工具的轮胎前部、车标处、后视镜处、反光镜处等等。
可选地，该样本行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。上述行驶参数可以通过运载工具上相应的传感器获取，上述行驶参数具体的数值可以显示在运载工具的仪表盘或者显示屏上供用户查看。
S604，将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到多组控制参数中每一组控制参数对应的预测概率值。
S605，根据预测概率值和目标概率值，训练第一模型。
其中，该目标概率值可以是预先设置的概率值，该目标概率值可以是模型训练对应的数据标签。
可选地，上述目标概率值也可以替换为得分，即可以通过每一组控制参数的预测得分和目标得分，训练第一模型。
本申请实施例中，能够使得第一模型输出的每一组控制参数对应的预测概率值不断接近目标概率值，从而使得第一模型在实际应用中能够根据输入的数据得到最佳的控制参数。
一个实施例中，步骤S604可以替换为：将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到预测控制参数。步骤S605可以替换为：根据预测控制参数和样本控制参数，训练第一模型。
其中，该样本控制参数可以是预测控制参数对应的数据标签，样本控制参数可以是基于样本参考信号、样本路况信息和样本行驶参数得到的。
本申请实施例中，能够使得第一模型输出的预测控制参数不断接近样本控制参数，从而使得第一模型在实际应用中能够根据输入的数据拟合出最佳的控制参数。
可选地，上述样本控制参数可以是第一控制参数集中概率值最大或得分最高的控制参数，该第一控制参数集中包括一个或多个控制参数。
可选地，上述样本控制参数可以是第二控制参数集中概率值最大或得分最高的控制参数，该第二控制参数集是对所述第一控制参数集分类得到的。这样，能够在模型训练时，提高模型训练的效率。
图7是本申请实施例提供的另一种降噪方法的示意性框图，方法700中的模型训练阶段是对方法600的模型训练过程的详细介绍，方法700中的模型使用阶段是对方法500中使用第一模型确定控制参数的过程的详细介绍。
如图7所示，方法700可以分为模型训练阶段和模型使用阶段。
(1)模型训练阶段：
运载工具可以通过传感器采集参考信号、误差信号、摄像头信号、车控信号等，并利用误差信号、参考信号和次级路径信息计算出控制参数，所有的控制参数可以构成控制参数集。可选地，控制参数集可以使用不同的方法进行分类，每一类控制参数集仅保存一组控制参数。将所有的控制参数均保存后，运载工具可以基于参考信号、误差信号(可选)、摄像头信号、车控信号对算法模型进行预先训练。其中，该算法模型可以是分类算法或择优算法，其不仅仅限于神经网络算法或非神经网络算法，该算法模型能够从控制参数集中选择最优的参数。
应理解，上述算法模型可以是方法500中的第一模型，车控信号可以对应方法500中的运载工具的行驶参数，摄像头信号可以用于得到方法500中的运载工具周围的路况信息。
(2)模型使用阶段：
运载工具可以通过传感器采集参考信号、误差信号(可选)、摄像头信号、车控信号、导航信息，并基于训练得到的算法模型选择最优的控制参数。然后根据参考信号和控制参数得到控制信号，并将控制信号通过发声装置输出。其中，控制参数择优这一步骤可以每隔一段时间进行一次，例如，每隔100毫秒、50毫秒进行一次。并且，利用参考信号得到控制信号并经由发声装置输出需要每个采样点都进行计算，而且需要连续、实时进行。
图8是本申请实施例提供的降噪方法的一种算法模型训练的流程性示意图，方法800是对方法700中模型训练阶段的详细介绍，方法800可以包括步骤S801至S805。
S801，获取次级路径信息。
具体地，运载工具可以控制次级源扬声器播放白噪声，并使用误差麦克风接收该白噪声，在这个过程中，利用最小均方算法(least mean square，LMS)或者维纳滤波算法，运载工具能够获取次级源扬声器到误差麦克风的次级路径信息。
可选地，如果运载工具上设置有多个次级源扬声器或多个误差麦克风，则每个次级源扬声器和对应的误差麦克风之间均需要获取相应的次级路径信息。考虑到用户的耳朵(头部)的位置不同时，次级路径不同，因此可以通过多次测量次级路径，并结合运载工具内的摄像头记录此时用户的耳朵(头部)位置。
S802，确定控制参数，并将控制参数添加至控制参数集。
在运载工具行驶的过程中，运载工具能够采集参考信号、误差信号、摄像头信号、车控信号等信息，并且可以利用LMS或者维纳滤波算法结合获取的参考信号、误差信号以及对应的次级路径信息，确定控 制参数，并将该控制参数添加至控制参数集合A中。
其中，通过部署在运载工具外的摄像头采集的信号，能够确定运载工具周围的路况信息，通过该路况信息，运载工具可以确定当前以及未来数秒内的路况。通过部署在运载工具内的摄像头采集的信号，运载工具可以判出用户的耳朵(头部)的位置，从而精确地确定需要降噪的区域。此外，用户的耳朵(头部)的位置也可以通过其他传感器获取，例如，激光传感器。
S803，对控制参数集进行分类。
具体地，运载工具可以对控制参数集A进行分类，将相同或近似的一组控制参数用一个新的参数代替，且保证有新的参数代替后，控制参数的降噪性能下降较少。由同一个参数代替的控制参数可以分为同一类，所有新的控制参数可以构成集合B。
可选地，上述分类的方法可以是：对控制参数进行傅里叶变换，将所有频域幅度和相位幅度接近的集合分为同一类，使用频域的均值傅里叶变换结果作为新的参数的代替值。
S804，设计算法模型。
可选地，该模型可以是深度神经网络DNN模型，也可以是由常规分类算法构成，该分类算法能够实现分类功能。该算法模型可以是方法600中的第一模型。
在模型设计完成后，将采集的参考信号、误差信号、摄像头信号、车控信号等信息输入到模型中，该模型能够输出集合B中不同类别的控制参数的概率值。以该模型为神经网络模型为例，输入模型的信息可以是以下内容中的一种或多种：
(1)最近5秒内部署在运载工具外的摄像头拍摄的多组照片，拍摄照片的时间间隔可以是200毫秒。
(2)驾驶人员的耳朵(头部)位置信息，该位置信息可以由(视觉)算法从部署在运载工具内摄像头等传感器中获取。
(3)车速信息、加减速踏板信息、车身倾斜角度信息、车胎压等信息。
(4)参考信号、误差信号。
应理解，上述参考信号可以是方法600中的样本参考信号，车控信号可以是方法600中的样本行驶参数，摄像头信号可以包括方法600中的样本路况信息。
S805，对算法模型进行训练。
在模型训练时，上述信息输入模型后，对应的最优类别的控制参数的目标概率值为1，其他的控制参数的目标概率值为0，训练过程让神经网络模型的输出结果迭代并逼近上述目标概率值。其中，最优类别的控制参数可以是方法600中的样本控制参数，神经网络模型输出结果可以是方法600中的预测概率值。
在另一种实现方式中，可以将采集的参考信号、误差信号、摄像头信号、车控信号等信息输入到算法模型中，得到预测控制参数；再根据预测控制参数和对应的样本控制参数(该样本控制参数属于集合B)训练算法模型，训练的过程是让神经网络模型输出的预测控制参数迭代并逼近对应的样本控制参数。
本申请实施例中，可以通过参考信号、误差信号、摄像头信号、车控信号训练模型，从而使得该模型在实际使用时，能够输出最优的控制参数。
图9是本申请实施例提供的降噪方法算法模型使用的流程性示意图，方法900可以由图1中的运载工具100执行，或者也可以由运载工具100中的计算平台130执行，或者还可以由计算平台130中的片上系统SoC执行，或者，还可以由计算平台130中的处理器执行。方法900是对方法700中模型使用阶段的详细介绍，下面以运载工具作为执行主体介绍方法900，方法900可以包括步骤S901至S905。
S901，在运载工具行驶的过程中，采集信号。
示例性地，该信号可以包括：参考信号、误差信号、摄像头信号和车控信号。
S902，将信号输入到算法模型中，获得控制参数。
具体地，将采集到的参考信号、误差信号、摄像头信号和车控信号等输入到已训练的算法模型中(例如，通过方法800训练的算法模型)，获得控制参数。
S903，根据控制参数和参考信号，确定控制信号。
具体地，选择步骤S902中确定的控制参数，结合采集到的参考信号，计算控制信号。
S904，使用控制源(发声装置)播放该控制信号。
具体地，该发声装置可以是方法800中的次级源扬声器。
S905，控制参数更新。
具体地，在方法900中采集的参考信号和步骤S904，需要每个采样点不间断的计算，方法900的其 他过程可以每隔一段时间计算一次，从而完成控制参数的更新。
本申请实施例中，运载工具能够使用经过训练的算法模型，得到最优的控制参数，在这个过程中，由于考虑到了摄像头信号和车控信号，能够使得选择的控制参数更加准确，可以在不同的车况下以及车况变化时改善降噪的效果。
图10是本申请实施例提供的降噪方法的另一种算法模型训练的流程性示意图，方法1000是对方法700中模型训练阶段的详细介绍，方法1000可以包括步骤S1001至S1004。
S1001，获取次级路径信息。
具体地，运载工具能够控制次级源扬声器播放白噪声，并使用误差麦克风接收该白噪声，在这个过程中，利用最小均方算法(least mean square，LMS)或者维纳滤波算法，运载工具能够获取次级源扬声器到误差麦克风的次级路径信息。
可选地，如果运载工具上设置有多个次级源扬声器或多个误差麦克风，则每个次级源扬声器和对应的误差麦克风之间均需要获取相应的次级路径信息。考虑到用户的耳朵(头部)的位置不同时，次级路径不同，因此可以通过多次测量次级路径，并结合车内摄像头记录此时用户的耳朵(头部)位置。
S1002，确定控制参数，并将控制参数添加至控制参数集。
在运载工具行驶的过程中，运载工具能够采集参考信号、误差信号、摄像头信号、车控信号等信息，并利用LMS或者维纳滤波算法结合获取的参考信号、误差信号以及对应的次级路径信息，确定控制参数，并将该控制参数添加至控制参数集合A中。
通过该路况信息，运载工具可以确定当前以及未来数秒内的路况。通过部署在运载工具内的摄像头采集的信号，运载工具可以判出用户的耳朵(头部)的位置，从而精确地确定需要降噪的区域。此外，用户的耳朵(头部)的位置也可以通过其他传感器获取，例如，激光传感器。
S1003，设计算法模型。
可选地，该模型可以是DNN模型，也可以是由常规分类算法构成，该分类算法能够实现分类功能。该算法模型可以是方法600中的第一模型。
在模型设计好以后，输入采集的参考信号、误差信号、摄像头信号、车控信号等信息，该模型能够输出集合A中不同类别的概率值。以该模型为神经网络模型为例，输入模型的信息可以是以下内容中的一种或多种：
(1)最近5秒内部署在运载工具外的摄像头拍摄的多组照片，拍摄照片的时间间隔可以是200毫秒。
(2)驾驶人员的人耳(人头)位置信息，该位置信息可以由(视觉)算法从部署在运载工具内摄像头等传感器中获取。
(3)车速信息、加减速踏板信息、车身倾斜角度信息、车胎压等信息。
(4)参考信号。
应理解，上述参考信号可以是方法600中的样本参考信号，车控信号可以是方法600中的样本行驶参数，摄像头信号可以包括：方法600中的样本路况信息。
S1004，对算法模型进行训练。
在模型训练时，上述信息输入模型后，对应的最优类别的控制参数目标概率值为1，其他控制参数的目标概率值为0，训练过程让神经网络模型的输出结果迭代并逼近上述目标概率值。其中，最优类别的控制参数可以是方法600中的样本控制参数，神经网络模型输出结果可以是方法600中的预测概率值。
在另一种实现方式中，可以将采集的参考信号、误差信号、摄像头信号、车控信号等信息输入到算法模型中，得到预测控制参数；再根据预测控制参数和对应的样本控制参数(该样本控制参数属于集合A)训练算法模型，训练的过程是让神经网络模型输出的预测控制参数迭代并逼近对应的样本控制参数。
本申请实施例中，可以通过参考信号、误差信号、摄像头信号、车控信号训练模型，从而使得该模型在实际使用时，能够输出最优的控制参数。
图11是本申请实施例提供的降噪方法的另一种算法模型使用的流程性示意图，方法1100可以由图1中的运载工具100执行，或者也可以由运载工具100中的计算平台130执行，或者还可以由计算平台130中的片上系统SoC执行，或者，还可以由计算平台130中的处理器执行。方法1100是对方法700中模型使用阶段的详细介绍，下面以运载工具作为执行主体介绍方法1100，方法1100可以包括步骤S1101至S1105。
S1101，运载工具在行驶的过程中，采集信号。
示例性地，该信号可以包括：参考信号、摄像头信号和车控信号等信息。
S1102，将信号输入到算法中，获得控制参数。
具体地，将采集到的参考信号、摄像头信号和车控信号等信息输入到已训练的算法中(例如，通过方法1000训练的算法)，获得控制参数。
S1103，根据控制参数和参考信号，确定控制信号。
具体地，选择步骤S1102确定的控制参数，结合采集到的控制信号，计算控制信号。
S1104，使用控制源(发声装置)播放该控制信号。
具体地，该发声装置可以是方法1000中的次级源扬声器。
S1105，控制参数更新。
具体地，在方法1100中采集的参考信号和步骤S1104，需要每个采样点不间断的计算，方法1100的其他过程可以每隔一段时间计算一次，从而完成控制参数的更新。
本申请实施例中，能够使用经过训练的算法模型，得到最优的控制参数，在这个过程中，由于考虑到了摄像头信号和车控信号，能够使得选择的控制参数更加准确，可以在不同的车况下以及车况变化时改善降噪的效果。此外，相对于方法900，方法1100在应用的过程中无需采集和使用误差信号，能够更简单、高效地得到最优的控制参数。
本申请实施例还提供用于实现以上任一种方法的装置，该装置包括用于实现以上任一种方法中运载工具100所执行的各步骤的单元。
图12是本申请实施例提供的降噪或模型训练装置1200的示意图，该装置1200可以包括获取单元1210、存储单元1220和处理单元1230。获取单元1210用于获取指令和/或数据，获取单元1210还可以称为通信接口或通信单元。存储单元1220，用于实现相应的存储功能，存储相应的指令和/或数据。处理单元1230用于进行数据处理。处理单元1230可以读取存储单元中的指令和/或数据，以使得装置1200实现前述降噪方法或模型训练方法。
作为一种设计，该装置1200包括：获取单元1210和处理单元1230；获取单元1210，用于获取参考信号、运载工具周围的路况信息和运载工具的行驶参数，该参考信号是第一传感器采集的，该第一传感器是部署在运载工具的噪声源附近处的传感器；处理单元1230，用于根据参考信号、运载工具周围的路况信息和运载工具的行驶参数，控制运载工具的发声装置播放控制信号，该控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
一种可能的实现方式中，获取单元1210，还用于获取运载工具的座舱内用户的耳朵或头部的位置信息；处理单元1230，具体用于根据参考信号、运载工具周围的路况信息、运载工具的行驶参数和用户的耳朵或头部的位置信息，控制运载工具的发声装置播放控制信号。
一种可能的实现方式中，获取单元1210，还用于获取误差信号，该误差信号是部署在用户耳朵或头部所在位置附近的第二传感器采集的；处理单元1230，具体用于根据参考信号、运载工具周围的路况信息、运载工具的行驶参数和误差信号，控制运载工具的发声装置播放控制信号。
一种可能的实现方式中，处理单元1230，具体用于：根据参考信号、运载工具周围的路况信息和运载工具的行驶参数，确定控制参数，该控制参数包括滤波器的系数；根据控制参数和参考信号，控制运载工具的发声装置播放控制信号。
一种可能的实现方式中，处理单元1230，具体用于将参考信号、路况信息和行驶参数输入到第一模型中，得到控制参数；其中，该第一模型是基于训练样本得到的，训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
一种可能的实现方式中，处理单元1230，具体用于通过第三传感器采集的数据获取运载工具周围的路况信息，该第三传感器位于运载工具上，该第三传感器用于采集运载工具外部的图像信息。
一种可能的实现方式中，获取单元1210，还用于获取服务器发送的导航信息；处理单元1230，还用于根据导航信息，确定运载工具周围的路况信息。
一种可能的实现方式中，该运载工具周围的行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
作为另一种设计，该装置1200包括：获取单元1210和处理单元1230；获取单元1210用于获取次级路径信息、样本参考信号和样本误差信号；处理单元1230，用于根据次级路径信息、样本参考信号和样本误差信号，确定多组控制参数；获取单元1210，还用于获取样本路况信息和样本行驶参数；处理单元 1230，还用于：将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到多组控制参数中每一组控制参数对应的预测概率值；根据预测概率值和目标概率值，训练第一模型，其中，该目标概率值是预先设置的概率值。
作为另一种设计，该装置1200包括：获取单元1210和处理单元1230；获取单元1210用于获取次级路径信息、样本参考信号和样本误差信号；处理单元1230，用于：将样本参考信号、样本路况信息和样本行驶参数输入到第一模型中，得到预测控制参数；根据预测控制参数和样本控制参数，训练第一模型。
一种可能的实现方式中，该样本控制参数是第一控制参数集中概率值最大的控制参数，该第一控制参数集中包括一个或多个控制参数。
一种可能的实现方式中，该样本控制参数是第二控制参数集中概率值最大的控制参数，该第二控制参数集是对第一控制参数集分类得到的。
一种可能的实现方式中，该样本行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
可选地，若该装置1200位于运载工具100中，上述处理单元1230可以是图1所示的处理器131。
图13是本申请实施例提供的降噪或模型训练装置1300示意图，该装置1300可应用于图1的运载工具100中。
该装置1300包括：存储器1310、处理器1320、以及通信接口1330。其中，存储器1310、处理器1320，通信接口1330通过内部连接通路相连，该存储器1310用于存储指令，该处理器1320用于执行该存储器1310存储的指令，以控制通信接口1330获取信息，或者使所述降噪装置执行上述各实施例中的降噪方法或模型训练方法。可选地，存储器1310既可以和处理器1320通过接口耦合，也可以和处理器1320集成在一起。
需要说明的是，上述通信接口1330使用例如但不限于收发器一类的收发装置。上述通信接口1330还可以包括输入/输出接口(input/output interface)。
处理器1320存储有一个或多个计算机程序，该一个或多个计算机程序包括指令。当该指令被所述处理器1320运行时，使得该降噪装置1300执行上述各实施例中降噪方法。
在实现过程中，上述方法的各步骤可以通过处理器1320中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的方法可以直接体现为硬件处理器执行完成，或者用处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器1310，处理器1320读取存储器1310中的信息，结合其硬件完成上述方法的步骤。为避免重复，这里不再详细描述。
可选地，图13中的通信接口1330可以实现图12中的获取单元1210，图13中的存储器1310可以实现图12中的存储单元1230，图13中的处理器1320可以实现图12中的处理单元1230。
可选地，该装置1200或装置1300可以是计算平台，该计算平台可以是车载计算平台或云端计算平台。
可选地，该装置1200或装置1300可以位于图1中的运载工具100中。
可选地，该装置1200或装置1300可以为图1运载工具中的计算平台130。
本申请实施例还提供一种计算机可读介质，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得所述计算机执行上述图5至图11中的任一种方法。
本申请实施例还提供一种芯片，包括：电路，该电路用于执行上述图5至图11中的任一种方法。
本申请实施例还提供一种运载工具，包括图12或图13任一种装置。
本申请实施例还提供了一种降噪系统，包括图12或图13任一种降噪装置以及一个或多个发声装置。
本申请实施例还提供了一种降噪系统，包括计算平台以及一个或多个发声装置，所述计算平台用于：获取运载工具周围的路况信息、所述运载工具的行驶参数和第一传感器采集的参考信号，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述一个和多个发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
可选地，上述计算平台可以是运载工具中的计算平台130。
本申请实施例还提供一种计算机程序产品，该计算机产品包括计算机程序，当所述计算机程序被运行时，使得计算机执行上述图5至图11中的任一种方法。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (22)

  1. 一种降噪方法，其特征在于，所述方法包括：
    获取参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，所述参考信号是第一传感器采集的，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；
    根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
  2. 如权利要求1所述的方法，其特征在于，所述方法还包括：
    获取所述用户的耳朵或头部的位置信息；
    所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：
    根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述用户的耳朵或头部的位置信息，控制所述运载工具的发声装置播放所述控制信号。
  3. 如权利要求1或2所述的方法，其特征在于，所述方法还包括：
    获取误差信号，所述误差信号是部署在所述用户耳朵或头部所在位置附近的第二传感器采集的；
    所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：
    根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述误差信号，控制所述运载工具的发声装置播放所述控制信号。
  4. 如权利要求1至3任一项所述的方法，其特征在于，所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，包括：
    根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，所述控制参数包括滤波器的系数；
    根据所述控制参数和所述参考信号，控制所述运载工具的发声装置播放所述控制信号。
  5. 如权利要求4所述的方法，其特征在于，所述根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，包括：
    将所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数输入到第一模型中，得到所述控制参数；
    其中，所述第一模型是基于训练样本得到的，所述训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
  6. 如权利要求1至5任一项所述的方法，其特征在于，所述获取运载工具周围的路况信息，包括：
    通过第三传感器采集的数据获取所述运载工具周围的路况信息，所述第三传感器位于所述运载工具上，所述第三传感器用于采集所述运载工具外部的图像信息。
  7. 如权利要求1至5任一项所述的方法，其特征在于，所述获取运载工具周围的路况信息，包括：
    获取服务器发送的导航信息；
    根据所述导航信息确定所述运载工具周围的路况信息。
  8. 如权利要求1至7任一种所述的方法，其特征在于，所述运载工具的行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
  9. 一种降噪装置，其特征在于，所述装置包括：获取单元和处理单元；
    所述获取单元，用于获取参考信号、运载工具周围的路况信息和所述运载工具的行驶参数，所述参考信号是第一传感器采集的，所述第一传感器是部署在所述运载工具的噪声源附近处的传感器；
    所述处理单元，用于根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，控制所述运载工具的发声装置播放控制信号，所述控制信号用于降低用户耳朵或头部所在位置附近的噪声，所述用户位于所述运载工具的座舱内。
  10. 如权利要求9所述的装置，其特征在于，
    所述获取单元，还用于获取所述用户的耳朵或头部的位置信息；
    所述处理单元，具体用于根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述用户的耳朵或头部的位置信息，控制所述运载工具的发声装置播放所述控制信号。
  11. 如权利要求9或10所述的装置，其特征在于，
    所述获取单元，还用于获取误差信号，所述误差信号是部署在所述用户耳朵或头部所在位置附近的第二传感器采集的；
    所述处理单元，具体用于根据所述参考信号、所述运载工具周围的路况信息、所述运载工具的行驶参数和所述误差信号，控制所述运载工具的发声装置播放所述控制信号。
  12. 如权利要求9至11任一项所述的装置，其特征在于，
    所述处理单元，具体用于：
    根据所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数，确定控制参数，所述控制参数包括滤波器的系数；
    根据所述控制参数和所述参考信号，控制所述运载工具的发声装置播放所述控制信号。
  13. 如权利要求9至12任一项所述的装置，其特征在于，
    所述处理单元，具体用于将所述参考信号、所述运载工具周围的路况信息和所述运载工具的行驶参数输入到第一模型中，得到所述控制参数；
    其中，所述第一模型是基于训练样本得到的，所述训练样本包括：样本路况信息、样本参考信号、样本行驶参数和样本控制参数。
  14. 如权利要求9至13任一项所述的装置，其特征在于，
    所述处理单元，具体用于通过第三传感器采集的数据获取所述运载工具周围的路况信息，所述第三传感器位于所述运载工具上，所述第三传感器用于采集所述运载工具外部的图像信息。
  15. 如权利要求9至13任一项所述的装置，其特征在于，
    所述获取单元，还用于获取服务器发送的导航信息；
    所述处理单元，还用于根据所述导航信息，确定所述运载工具周围的路况信息。
  16. 如权利要求9至15任一项所述的装置，其特征在于，所述运载工具的行驶参数包括以下内容至少一种：车胎压、车速、加速度、加速踏板开度、刹车踏板开度、车身姿态和车身倾斜角。
  17. 一种降噪装置，其特征在于，包括：处理器和存储器，所述处理器与所述存储器耦合，用于读取并执行所述存储器中的指令，以执行如权利要求1至8中任一项所述的方法。
  18. 一种计算机可读介质，其特征在于，所述计算机可读介质存储有程序代码，当所述计算机程序代码在计算机上运行时，使得所述计算机执行如权利要求1至8中任一项所述的方法。
  19. 一种芯片，其特征在于，包括：电路，所述电路用于执行如权利要求1至8中任一项所述的方法。
  20. 一种降噪系统，其特征在于，包括：如权利要求9至17中任意一项所述的降噪装置以及一个或多个发声装置。
  21. 一种运载工具，其特征在于，包括：如权利要求9至17中任意一项所述的降噪装置，或者，如权利要求20所述的降噪系统。
  22. 如权利要求21所述的运载工具，其特征在于，所述运载工具为车辆。
PCT/CN2023/136081 2022-12-06 2023-12-04 降噪方法、装置以及运载工具 Ceased WO2024120331A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211558434.4 2022-12-06
CN202211558434.4A CN118155594A (zh) 2022-12-06 2022-12-06 降噪方法、装置以及运载工具

Publications (1)

Publication Number Publication Date
WO2024120331A1 true WO2024120331A1 (zh) 2024-06-13

Family

ID=91287622

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/136081 Ceased WO2024120331A1 (zh) 2022-12-06 2023-12-04 降噪方法、装置以及运载工具

Country Status (2)

Country Link
CN (1) CN118155594A (zh)
WO (1) WO2024120331A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN121011164A (zh) * 2024-08-28 2025-11-25 深圳引望智能技术有限公司 一种噪声控制系统、方法、装置及车辆
CN119049444A (zh) * 2024-09-25 2024-11-29 东风汽车集团股份有限公司 车内降噪方法、装置、设备、存储介质及产品

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10347236B1 (en) * 2018-02-28 2019-07-09 Harman International Industries, Incorporated Method and apparatus for continuously optimized road noise cancellation
CN110992924A (zh) * 2019-12-26 2020-04-10 无锡吉兴汽车声学部件科技有限公司 一种无反馈机制的汽车主动降噪方法
CN111261134A (zh) * 2018-11-30 2020-06-09 上汽通用汽车有限公司 车辆道路噪声的降噪方法及其系统
CN111833841A (zh) * 2020-06-12 2020-10-27 清华大学苏州汽车研究院（相城） 一种用于汽车道路噪声的主动控制系统、方法及车辆系统
CN111933105A (zh) * 2020-09-17 2020-11-13 南昌工程学院 一种汽车噪声控制装置及其控制方法
CN112550191A (zh) * 2020-12-10 2021-03-26 奇瑞汽车股份有限公司 用于汽车主动降噪系统的降噪参考信号生成方法
CN113228161A (zh) * 2019-01-04 2021-08-06 哈曼国际工业有限公司 高频宽带空气传播噪声主动噪声消除
CN113291248A (zh) * 2021-05-25 2021-08-24 浙江大学 一种多通道解耦分列的汽车车厢噪声主动控制方法和系统
US20210323562A1 (en) * 2020-04-21 2021-10-21 Hyundai Motor Company Noise control apparatus, vehicle having the same and method for controlling the vehicle
CN114255726A (zh) * 2021-04-13 2022-03-29 北京安声科技有限公司 主动降噪方法、车载主动降噪系统以及汽车

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10347236B1 (en) * 2018-02-28 2019-07-09 Harman International Industries, Incorporated Method and apparatus for continuously optimized road noise cancellation
CN111261134A (zh) * 2018-11-30 2020-06-09 上汽通用汽车有限公司 车辆道路噪声的降噪方法及其系统
CN113228161A (zh) * 2019-01-04 2021-08-06 哈曼国际工业有限公司 高频宽带空气传播噪声主动噪声消除
CN110992924A (zh) * 2019-12-26 2020-04-10 无锡吉兴汽车声学部件科技有限公司 一种无反馈机制的汽车主动降噪方法
US20210323562A1 (en) * 2020-04-21 2021-10-21 Hyundai Motor Company Noise control apparatus, vehicle having the same and method for controlling the vehicle
CN111833841A (zh) * 2020-06-12 2020-10-27 清华大学苏州汽车研究院（相城） 一种用于汽车道路噪声的主动控制系统、方法及车辆系统
CN111933105A (zh) * 2020-09-17 2020-11-13 南昌工程学院 一种汽车噪声控制装置及其控制方法
CN112550191A (zh) * 2020-12-10 2021-03-26 奇瑞汽车股份有限公司 用于汽车主动降噪系统的降噪参考信号生成方法
CN114255726A (zh) * 2021-04-13 2022-03-29 北京安声科技有限公司 主动降噪方法、车载主动降噪系统以及汽车
CN113291248A (zh) * 2021-05-25 2021-08-24 浙江大学 一种多通道解耦分列的汽车车厢噪声主动控制方法和系统

Also Published As

Publication number Publication date
CN118155594A (zh) 2024-06-07

Similar Documents

Publication Publication Date Title
US12399086B2 (en) Simulating realistic test data from transformed real-world sensor data for autonomous machine
US12321829B2 (en) System and method for contextualized vehicle operation determination
US11704781B2 (en) Enhanced high-dynamic-range imaging and tone mapping
US20250191466A1 (en) Object identification using audible cues for autonomous and semi-autonomous systems and applications
US20250322608A1 (en) Environmental reconstruction for path planning in robotics systems and applications
US20220283638A1 (en) Gaze determination using glare as input
US20210286923A1 (en) Sensor simulation and learning sensor models with generative machine learning methods
US20250148810A1 (en) Systems and methods for performing operations in a vehicle using gaze detection
US20220141450A1 (en) Applications for detection capabilities of cameras
WO2024120331A1 (zh) 降噪方法、装置以及运载工具
CN115344117A (zh) 适配的眼睛追踪机器学习模型引擎
CN110248861A (zh) 在车辆操纵过程中使用机器学习模型来引导车辆
WO2022061277A1 (en) Measuring the effects of augmentation artifacts on a machine learning network
JP2018107759A (ja) 画像処理装置、画像処理方法、及び画像処理システム
CN114435351A (zh) 用于基于神经网络的自动驾驶的系统和方法
JP2024526591A (ja) 音発生装置制御方法、音発生システム及び車両
WO2022061289A1 (en) Simulating viewpoint transformations for sensor independent scene understanding in autonomous systems
KR20240017952A (ko) 비히클을 위한 트랜스패런트 오디오 모드
WO2024203100A1 (ja) 騒音抑制装置および騒音抑制方法、並びにプログラム
US10382862B2 (en) Noise testing in an autonomous vehicle
US12363269B2 (en) Deferred color correction in image processing pipelines for autonomous systems and applications
WO2024203098A1 (ja) 騒音抑制装置および騒音抑制方法、並びにプログラム
US20260025475A1 (en) Image signal processing for rgb-ir sensor arrays
US12507003B2 (en) Voice pre-processing pipeline for exterior communications on autonomous vehicle
US20250356529A1 (en) Shading calibration for radial sensor lenses

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23899909

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 23899909

Country of ref document: EP

Kind code of ref document: A1