WO2020161914A1

WO2020161914A1 - Apparatus, method and program for noise reduction, and computer-readable recording medium recording noise reduction program

Info

Publication number: WO2020161914A1
Application number: PCT/JP2019/004727
Authority: WO
Inventors: Hiroo YAMAZAKI
Original assignee: Volvo Truck Corporation
Priority date: 2019-02-08
Filing date: 2019-02-08
Publication date: 2020-08-13

Abstract

A noise reduction apparatus has a microphone for detecting noise, digital audio device including a digital signal processor, and at least one speaker connected to the digital audio device. The digital signal processor applies the Kalman filter to the noise detected by the microphone to estimate noise at the head of a driver, superimposes a signal for cancelling the estimated noise on an audio signal of the digital audio device, and drives the at least one speaker with the audio signal.

Description

APPARATUS, METHOD AND PROGRAM FOR NOISE REDUCTION, AND COMPUTER-READABLE RECORDING MEDIUM RECORDING NOISE REDUCTION PROGRAM

The present invention relates to a noise reduction apparatus for reducing noise the driver feels, to a noise reduction method, to a noise reduction program, and to a computer-readable recording medium recording the noise reduction program.

In order to reduce engine noise of an automobile and the like, a technique for adaptively controlling a speaker placed in a vehicle interior based on noise detected by a microphone has been proposed, as disclosed in JP 2004-51036 A.

JP 2004-51036 A

However, it has been difficult in conventionally proposed techniques to retrofit a noise reduction apparatus to an existing automobile, because in addition to placing a microphone in the vehicle interior, a control unit for processing the output signal of the microphone to drive a speaker is necessary.

Therefore, an object of the present invention is to provide a noise reduction apparatus and a noise reduction method which can be retrofitted in an existing automobile by using functions of digital audio device. Another object of the present invention is to provide a noise reduction program for realizing the noise reduction apparatus and the noise reduction method, and a computer-readable recording medium recording the noise reduction program.

According to an embodiment of the present invention, the noise reduction apparatus comprises a microphone for detecting noise, digital audio device including a digital signal processor, and at least one speaker connected to the digital audio device. The digital signal processor applies the Kalman filter to the noise detected by the microphone to estimate noise at the head of a driver, superimposes a signal for cancelling the estimated noise on an audio signal of the digital audio device, and drives the at least one speaker with the audio signal.

According to another embodiment of the present invention, a digital signal processor, capable of reading an output signal of a microphone for detecting noise and capable of outputting a driving signal to at least one speaker connected to digital audio device, applies the Kalman filter to the noise detected by the microphone to estimate noise at the head of a driver. Then, the digital signal processor superimposes a signal for cancelling the estimated noise at the head of a driver on an audio signal of the digital audio device, and drives the at least one speaker with the audio signal.

According to still another embodiment of the present invention, a noise reduction program causes a computer to execute the steps of applying the Kalman filter to an output signal of a microphone for detecting noise to estimate noise at the head of a driver, superimposing a signal for cancelling the estimated noise on an audio signal of the digital audio device, and driving at least one speaker connected to the digital audio device with the audio signal. Additionally, a computer-readable recording medium recording the noise reduction program records the noise reduction program for causing the computer to execute the above-described steps.

According to the present invention, the noise reduction apparatus can be retrofitted in the existing automobile by using the functions of the digital audio device.

FIG. 1 is a plan view of an exemplary cabin layout of the truck. FIG. 2 is a right side view of an exemplary cabin layout of the truck. FIG. 3 is a flowchart of exemplary noise reduction processing. FIG. 4 is a flowchart of exemplary implementation procedure of the noise reduction program. FIG. 5 is a control block diagram of noise reduction processing. FIG. 6 is an illustration of a transfer function of the sound field. FIG. 7 is an illustration of the position relationship between a driver and each speaker. FIG. 8 is an illustration of the position relationship between a microphone and each speaker. FIG. 9 is a result of fast Fourier transform indicating noise at the head of the driver when the noise reduction program is not executed. FIG. 10 is a result of fast Fourier transform indicating noise at the head of the driver when the noise reduction program is executed. FIG. 11 is an illustration of another embodiment of the microphone.

Here in below, embodiments for implementing the present invention will be described in detail with reference to the accompanying drawings.
FIGS. 1 and 2 illustrate an example of a cabin layout of a truck, which can be an example to which the present embodiment is applied. The present invention is not limited to trucks, but can be applied to cabins of passenger cars, construction machines, and the like.

The noise reduction apparatus has a microphone 100 for detecting noise, digital audio device 120, and four speakers 140 connected to the digital audio device 120. The number of speakers 140 is not limited to four, but is at least one speaker.

The microphone 100 is attached to the front face of the casing of the digital audio device 120 so as to be incorporated into the digital audio device 120, and detects noise inside the cabin 200, for example, engine noise, brake noise, wind noise at the time of traveling and road noise of the truck. The microphone 100 is not limited to the casing of the digital audio device 120 but can also be attached to any position inside the cabin 200.

The digital audio device 120 is connected, for example, wirelessly via for example Bluetooth (Registered Trademark) and WiFi (Registered Trademark), or wired via for example a USB (Universal Serial Bus) port, to a digital device such as a smartphone, and plays back a music file, or the like, which is stored in the digital device. The digital audio device 120 is provided with a digital signal processor 122 for processing digital data such as a music file, a flash memory (not shown) for storing a music playback program and a noise reduction program, an SD (Secure Digital) memory slot (not shown), and the like.

Additionally, the digital audio device 120 is attached to a predetermined position of the dashboard 220 of the cabin 200; specifically, the position operable by the driver sitting on the driver seat 240. The digital audio device 120 may be provided with the display for displaying the music file, or the like, during playback and operation buttons for selecting and playing back the music file, or the like. The digital audio device 120 may also be provided with radio functions capable of receiving and outputting radio signals.

The speakers 140 are attached to predetermined positions in the cabin 200, specifically, at the top of left and right body side trims, in a state of being spaced apart by a predetermined distance. Each speaker 140 can be connected wirelessly via for example Bluetooth and WiFi, or wired via a dedicated port, to the digital audio device 120. The attachment positions of the speakers 140 are not limited to the top of the body side trims 260 but can be any positions, for example, the dashboard 220, door trims, front pillars, rear pillars, and rear trims.

Although no explanation has been given, FIGS. 1 and 2 show the passenger seat 280, the steering wheel 300, the accelerator pedal 320 and the door glass 340.

The noise reduction program according to the present embodiment is stored in advance in the flash memory of the digital audio device 120 wirelessly via such as Bluetooth and WiFi, wired via a USB port, or via an SD memory slot. Then, triggered by the activation of the digital audio device 120, the digital signal processor 122 executes the noise reduction program and reduces noise at the head of a driver, as will be described in detail below.

FIG. 3 shows an example of noise reduction processing, in which, triggered by activation of the digital audio device 120, repeatedly executed at predetermined time intervals by the digital signal processor 122. The noise reduction processing may only be executed while the engine is operating, if the main object is to reduce engine noise of the truck.

In step 1 (abbreviated as “S1” in FIG. 3; the same applies below), the digital signal processor 122 reads the output signal of the microphone 100 to measure noise at a predetermined position inside the cabin 200.

In step 2, the digital signal processor 122 applies the Kalman filter to the noise detected by the microphone 100 to estimate the noise at the head of the driver. That is, the digital signal processor 122 uses the Kalman filter, which is a type of Infinite Impulse Response Filters for estimating the state of a dynamic system by means of an observed value with error, and estimates the noise at the head of the driver which is distant from the noise measurement point. The Kalman filter will later be described in detail. Here, step 2 is an example for applying the Kalman filter to the output signal of the microphone for detecting noise to estimate the noise at the head of the driver.

In step 3, the digital signal processor 122 calculates the noise cancellation signal for cancelling the noise at the head of the driver. The noise cancellation signal is, for example, a signal having a phase opposite to the noise at the head of the driver. Here, if multiple speakers 140 are connected to the digital audio device 120, the digital signal processor 122 can calculate the noise cancellation signal for each speaker 140 according to the distance between the microphone 100 and each speaker 140 and the distance between the head of the driver and each speaker 140.

In step 4, the digital signal processor 122 superimposes the noise cancellation signal on the audio signal of music or the like, and drives each speaker 140 with the audio signal on which the noise cancellation signal is superimposed. Here, step 4 is an example for superimposing the signal for cancelling the estimated noise on the audio signal of the digital audio device, and for driving at least one speaker connected to the digital audio device with the audio signal.

According to such noise reduction apparatus, the Kalman filter is applied to the noise detected by the microphone 100 so that the noise at the head of the driver, which is located distant from the microphone 100, is estimated. Then, the noise cancellation signal for cancelling the noise at the head of the driver is calculated, and the noise cancellation signal is superimposed on the audio signal to drive the speakers 140. Accordingly, the noise at the head of the driver is reduced, and the noise the driver feels can be small.

At this time, the noise reduction apparatus can be equipped by installing the noise reduction program to the digital audio device 120, and thus, the noise reduction apparatus requires no special device, and can be retrofitted in existing trucks while reducing costs. The noise reduction apparatus can also be retrofitted, for example, by replacing the existing audio device with the digital audio device to install the noise reduction program, and by replacing the existing audio device with the digital audio device in which the noise reduction program is installed.

Next, FIG. 4 is referred to explain the implementation procedure of the noise reduction program using the Kalman filter, which is performed, for example, by the control system designer. The noise reduction program, as shown in FIG. 5, uses the transfer function G(s) of the sound field in the cabin 200 to calculate a signal indicating how the output signal (sound) of the speaker 140 is transmitted to the microphone 100, and superimposes external noise such as engine noise on the signal. Additionally, the noise reduction program applies the Kalman filter to the signal on which the external noise is superimposed, and estimates the noise at the head of the driver. Then, the noise reduction program generates the noise cancellation signal for cancelling the noise at the head of the driver to drive the speaker 140. At this time, the noise reduction program can superimpose the noise cancellation signal on the audio signal of the digital audio device 120.

In step 11, the control system designer defines, for each speaker 140, the transfer function Gd(s) of the sound field regarding the driver and the transfer function Gs(s) of the sound field regarding the microphone 100, as the transfer function G(s) of the sound field (sound space model) in the cabin 200. The transfer function Gd(s) of the sound field regarding the driver, as shown in FIG. 6, can be expressed by using the gain K1, the distance L1 between the speaker 140 and the head of the driver, the sound velocity c and the Laplace operator s, with respect to the output signal u of the speaker 140. Additionally, the transfer function Gs(s) of the sound field regarding the microphone 100 can be expressed by using the gain K2, the distance L2 between the speaker 140 and the microphone 100, the sound velocity c and the Laplace operator s, with respect to the output signal u of the speaker 140. It is supposed that the disturbance d such as engine noise is superimposed on the sound signal xd (noise) at the head of the driver and the output signal xs of the microphone 100, which are calculated from the transfer function Gd(s) of the sound field regarding the driver and the transfer function Gs(s) of the sound field regarding the microphone 100. Here, the sound velocity c can be a fixed value such as 340m/s, for example.

The transfer function Gd(s) of the sound field regarding the driver, which is calculated in step 11, can be the cross spectrum, which is calculated from the ratio of the cross correlation function of the sound signal of the speaker 140 measured at the head of the driver and the driving signal of the speaker 140 to the autocorrelation function of the driving signal of the speaker 140. Additionally, the transfer function Gs(s) of the sound field regarding the microphone 100, which is calculated in step 11, can be the cross spectrum, which is calculated from the ratio of the cross correlation function of the sound signal of the speaker 140 measured near the microphone 100 and the driving signal of the speaker 140 to the autocorrelation function of the driving signal of the speaker 140.

In step 12, the control system designer identifies the gains K1 and K2 as the optimum model parameters through experiments and simulations, for example.

In step 13, the control system designer calculates the system matrix A and control matrix B of each speaker 140. As preconditions for calculating the system matrix A and control matrix B, FIG. 7 shows the distance L_1rf between the driver 300 and the speaker 140 located front right, the distance L_1lf between the driver 300 and the speaker 140 located front left, the distance L_1rr between the driver 300 and the speaker 140 located rear right, and the distance L_1lr between the driver 300 and the speaker 140 located rear left. In addition to such preconditions, FIG. 8 shows the distance L_2rf between the microphone 100 and the speaker 140 located front right, the distance L_2lf between the microphone 100 and the speaker 140 located front left, the distance L_2rr between the microphone 100 and the speaker 140 located rear right, and the distance L_2lr between the microphone 100 and the speaker 140 located rear left.

Under these preconditions, the system matrix A and control matrix B of the speaker 140 located front right in the cabin 200 can be calculated as follows.

Additionally, under these preconditions, the system matrix A and control matrix B of the speaker 140 located front left in the cabin 200 can be calculated as follows.

Furthermore, under these preconditions, the system matrix A and control matrix B of the speaker 140 located rear right in the cabin 200 can be calculated as follows.

Still furthermore, under these preconditions, the system matrix A and control matrix B of the speaker 140 located rear left in the cabin 200 can be calculated as follows.

In step 14, the control system designer calculates the continuous-time system state equation for each speaker 140 based on the sound signal xd and output signal xs as well as the system matrix A and control matrix B as follows:

, where D represents a constant matrix for multiplying the disturbance noise d by constant, which can be set to any value.

In step 15, the control system designer transforms the continuous-time system state equation to the following discrete-time system state equation which can be operated by the digital signal processor 122.

, where C represents a constant matrix, I represents an identity matrix, and T represents sampling time.

When the Kalman filter gain is represented by K, the Kalman filter of the discrete-time system state equation can be expressed as follows:

In step 16, the control system designer calculates the Kalman filter gain K. When the variance matrix of the disturbance is represented by Q and the variance matrix of the observation noise is represented by R, the Kalman filter gain K can be calculated by the following solution P of the Riccati equation.

In step 17, the control system designer designs the optimum feedback control system according to the procedure described below. First, the control system designer considers a system which is expressed by the following difference equation.

The optimum control law of such system is defined by the following formula when F is represented by the feedback gain.

Then, the feedback gain F can be determined by calculating variance matrix Q of the disturbance and variance matrix R of the observation noise which minimizes the following equation:

Wherein, the feedback gain F is expressed as follows.

Here, in the above formula expressing the feedback gain F, P is the solution of the Riccati equation which can be expressed as follows:

In step 18, the control system designer creates the noise reduction program for realizing such designed Kalman filter and the optimum control law, and stores the noise reduction program in a flash memory wirelessly via Bluetooth and WiFi, wired via USB port, or via a network such as CAN (Controller Area Network), for example. The noise reduction program can be successively updated such as by a user and a worker of the dealer.

In short, the control system designer derives the Kalman filter from the transfer function G(s) of the sound field in the cabin 200 of the truck, more specifically, the time of arrival of sound in the cabin 200 is defined by the first order lag system model to derive the Kalman filter. At this time, the control system designer and the like can actually measure the first noise detected by the microphone 100, the second noise at the head of the driver, and the driving signal of the speaker 140, and calculate the transfer function G(s) of the sound field from the correlation of the first noise, second noise and driving signal. Additionally, the control system designer derives the feedback gain F of the system expressed by the difference equation, and uses it to calculate the optimum control law.

Next, the noise reduction effect of the noise reduction program will be explained. If the noise reduction program is not executed, as shown in FIG. 9, the noise at the head of the driver is large particularly in the range of 10Hz to 1000Hz, and the maximum value is 84.4dB. On the other hand, if the noise reduction program is executed, as shown in FIG. 10, the noise at the head of the driver is small in the range of 10Hz to 1000Hz, and the maximum value is 78.9dB. Therefore, as a result of executing the noise reduction program, the maximum value of the noise at the head of the driver is reduced by 5.5dB from 84.4dB to 78.9dB.

The noise caused during idling of a heavy duty truck is in the very low frequency range between 10Hz and 400Hz, and the sound insulation effect of glass cannot be exerted. However, implementing the noise reduction program of the present embodiment reduces the noise of 10Hz to 1000Hz so that the engine noise can be reduced without using thick glass, for example, and thus, contributing to cost reduction. Additionally, the noise reduction effect is exerted by simply installing the noise reduction program in the digital audio device 120, and thus, no cost is necessary for attaching a special device. Furthermore, since there is little influence on other electronic control units, there is no problem in controlling in-vehicle device.

The microphone is not limited to the microphone 100 attached to the digital audio device 120, but may be the microphone 322, which is incorporated in an information terminal 320 such as a smartphone carried by the driver 300, as shown in FIG. 11. In this case, the digital audio device 120 may read the noise detected by the microphone 322 of the information terminal 320 wirelessly via such as Bluetooth and WiFi.

Additionally, the noise reduction program may be installed in the information terminal 320 to generate the noise cancellation signal for cancelling noise at the head of the driver in the information terminal 320 and send it to the digital audio device 120 wirelessly. In this case, the digital audio device 120 may superimpose the noise cancellation signal on the audio signal to drive the speaker 140.

The noise reduction program can be stored in a computer-readable recording medium, for example, an SD card and a USB memory, and distributed in the market. Additionally, the noise reduction program can be stored in the storage in a node or the like connected to the Internet or the like, and can be distributed from the node. In this case, the storage of the node is an example of the computer-readable recording medium.

Furthermore, the noise reduction apparatus is not limited to automobiles such as trucks and passenger cars but can also be used for environmental improvement of rental offices and construction machines. Additionally, self-tuning functions using deep learning techniques and the like can be incorporated into the noise reduction apparatus.

One skilled in the art would readily understand that a new embodiment can be made by omitting a part of the technical idea of the various embodiments, freely combining parts of the technical idea of the various embodiments, and substituting a part of the technical idea of the various embodiments.

100 microphone
120 digital audio device
122 digital signal processor
140 speaker
200 cabin
300 driver
320 information terminal
322 microphone

Claims

A noise reduction apparatus comprising:
a microphone for detecting noise;
a digital audio device including a digital signal processor, and
at least one speaker connected to the digital audio device;
wherein the digital signal processor is configured to apply a Kalman filter to the noise detected by the microphone to estimate noise at the head of a driver, superimposes a signal for cancelling the estimated noise on an audio signal of the digital audio device, and drives the at least one speaker with the audio signal.
The noise reduction apparatus according to claim 1, wherein the Kalman filter is derived from a transfer function of a sound field in a cabin of an automobile.
The noise reduction apparatus according to claim 2, wherein the transfer function is calculated by defining a time of arrival of sound in a first order lag system model.
The noise reduction apparatus according to claim 2, wherein a first noise detected by the microphone, a second noise at the head of the driver, and a driving signal of the speaker are measured, and the transfer function is calculated from correlation of the first noise, the second noise and the driving signal.
The noise reduction apparatus according to any one of claims 1 to 4, wherein the microphone is incorporated into the digital audio device.
The noise reduction apparatus according to any one of claims 1 to 4, wherein the microphone is incorporated into an information terminal carried by the driver.
A noise reduction method comprising, by a digital signal processor capable of reading an output signal of a microphone for detecting noise and capable of outputting a driving signal to at least one speaker connected to digital audio device:
applying a Kalman filter to the noise detected by the microphone to estimate noise at the head of a driver,
superimposing a signal for cancelling the estimated noise on an audio signal of the digital audio device, and
driving the at least one speaker with the audio signal.
The noise reduction method according to claim 7, wherein the Kalman filter is derived from a transfer function of a sound field in a cabin of an automobile.
The noise reduction method according to claim 8, wherein the transfer function is calculated by defining a time of arrival of sound in a first order lag system model.
The noise reduction method according to claim 8, wherein a first noise detected by the microphone, a second noise at the head of the driver, and a driving signal of the speaker are measured, and the transfer function is calculated from correlation of the first noise, the second noise and the driving signal.
The noise reduction method according to any one of claims 7 to 10, wherein the microphone is incorporated into the digital audio device.
The noise reduction method according to any one of claims 7 to 10, wherein the microphone is incorporated into an information terminal carried by the driver.
A noise reduction program for causing a computer to execute the steps of:
applying a Kalman filter to an output signal of a microphone for detecting noise to estimate noise at the head of a driver;
superimposing a signal for cancelling the estimated noise on an audio signal of the digital audio device, and
driving at least one speaker connected to the digital audio device with the audio signal.
A computer-readable recording medium recording a noise reduction program for causing a computer to execute the steps of:
applying a Kalman filter to an output signal of a microphone for detecting noise to estimate noise at the head of a driver;
superimposing a signal for cancelling the estimated noise on an audio signal of the digital audio device, and
driving at least one speaker connected to the digital audio device with the audio signal.