Method for implementing automatic adjustment of audio output of a sound box based on a doppler radar sensor

Abstract

The method for realizing automatic adjustment of audio output of a sound box based on a Doppler radar sensor relates to the technical field of intelligent control of sound boxes. The technical defects that the existing sound boxes have complex structure and algorithm, poor privacy, short detection distance and difficulty in continuous tracking in automatically switching audio adjustment strategies under different life states are solved. First, a Doppler radar sensor module is used to track the distance and azimuth angle of the sound box in the room, and discretization is performed to form an index number. Then, a second digital signal processor of the sound box calls a preset audio adjustment scheme according to the index number to execute automatic adjustment of the audio output of the sound box. According to the equal loudness curve, a statistical analysis method is used to compensate for the frequency response loss, thereby realizing the function of automatic adjustment of the audio of the sound box, good privacy, continuous tracking, and precise and reliable control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of sound boxes. BACKGROUND

[0002] When the sound box plays music, the frequency response changes greatly at different distances of the listener, especially in a large room, under the condition of a certain volume, resulting in large and small loudness and great changes in sound quality. In addition, the listener has different requirements for music performance in different life states. For example, when walking leisurely in the living room, people generally prefer lively sound effects, while lying on the sofa to rest requires relatively flat sound effects and small volume, or directly turning off the power.

[0003] To achieve the automatic adjustment function of the above-mentioned audio: the listener can hear music with moderate loudness and basically unchanged sound quality wherever he goes; and in different life states, the corresponding performance effect can be automatically switched, the relative distance between the listener and the sound box or the specific position of the listener in the room must be obtained in real time, and the current life state must be inferred from this, and the corresponding strategy of audio adjustment must be called. The current scheme for monitoring a moving target generally includes a camera and a passive infrared device. However, there are obvious disadvantages in this application. The traditional camera scheme has complex structure and algorithm, and poor privacy; the passive infrared scheme has limited action distance and is difficult to continuously track.

[0004] The Doppler radar sensor has the capabilities of long action distance (more than 5m), large enough detection angle (up to 90 degrees), and continuous detection, which can continuously track the distance and direction of the listener, and there is no privacy problem. SUMMARY

[0005] In summary, the purpose of the present application is to solve the technical deficiencies of the existing sound box in automatically switching the audio adjustment strategy in different life states, such as complex structure and algorithm, poor privacy, short detection distance, and difficulty in continuous tracking, and to propose a method for realizing automatic adjustment of sound box audio output based on a Doppler radar sensor.

[0006] To solve the technical deficiencies proposed in the present application, the technical scheme adopted is:

[0007] The method for realizing automatic adjustment of sound box audio output based on a Doppler radar sensor, characterized in that the method first uses a Doppler radar sensor module to track the distance and azimuth angle of the listener in the room, and performs discretization to form an index number; then, a second digital signal processor of the sound box calls a preset audio adjustment scheme according to the index number to execute automatic adjustment of the sound box audio output.

[0008] The further defined technical solutions of the present application include:

[0009] The Doppler radar sensor module comprises a transmitting antenna, a first receiving antenna, a second receiving antenna and a first digital signal processor; the first digital signal processor mixes the transmitting signal of the transmitting antenna with the first receiving signal of the first receiving antenna to obtain an intermediate frequency signal IF, and performs a discrete Fourier transform after a low-pass filter to calculate the relative distance between the listener and the sound box; the first digital signal processor further calculates the time delay τ of the listener to the two receiving antennas based on the generalized cross-correlation-phase transform method according to the first receiving signal received by the first receiving antenna and the second receiving signal received by the second receiving antenna, so as to solve the azimuth angle θ.

[0010] The second digital signal processor of the sound box divides the room where the sound box is located into regions and assigns corresponding index values, detects and tracks the residence time of the listener in the corresponding region by the Doppler radar sensor module, and estimates the three life states of active, rest and sleep; when the active life state is judged, the bright melody is automatically switched; when the rest life state is judged, the gentle and soft melody is automatically switched; when the sleep life state is judged, the machine is automatically switched off.

[0011] The method for obtaining the azimuth angle θ comprises the following steps: , The frequency domain generalized cross-correlation (GCC) is as shown in formula 1:

[0012] (Formula 1)

[0013] Among them:

[0014] , are the frequency domain representations of the signals , after the discrete Fourier transform, the variable is the frequency point index, is the integral upper limit, and , is the number of sampling points for the first time discrete Fourier transform; is the frequency resolution, and , is the sampling rate; is the discrete time sequence number, ;

[0015] The time delay estimation is: (Formula 2)

[0016] The weighted function , formula 1 is converted into the following formula:

[0017] (formula 3)

[0018]

[0019] Wherein: is Conjugate of formula (2), the listener can obtain the time delay estimation of two received signals , echo distance difference : , =340m / s, is the sound speed; horizontal azimuth , is the receiving antenna spacing.

[0020] The beneficial effects of the present application are: the present application provides a location solution based on Doppler radar sensor (Doppler Radar Sensor), compensates the frequency response loss by using statistical analysis method according to the equal loudness curve, realizes the function of automatic adjustment of the sound box audio, has good privacy, can continuously track, and the control is accurate and reliable. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 It is the sound box structure principle schematic diagram of the present application;

[0022] Figure 2 It is the implementation method flow chart of the present application;

[0023] Figure 3 It is the distance, horizontal azimuth definition schematic diagram;

[0024] Figure 4 It is the room nine palace diagram schematic diagram;

[0025] Figure 5 It is the equal loudness curve diagram of the present application. DETAILED DESCRIPTION

[0026] The method of the present application is further explained in combination with the drawings and the preferred specific embodiments of the present application.

[0027] Refer to Figures 1 to 2As shown, the method for automatically adjusting the audio output of a sound box based on a Doppler radar sensor disclosed in the present application first uses a Doppler radar sensor module to track the distance and azimuth angle of the listener in the room, and to discretize, forming an index number; then, the second digital signal processor of the sound box calls the preset audio adjustment scheme according to the index number, and executes the automatic adjustment of the audio output of the sound box. The main functions to be achieved by the present application are: ① in the application scene of a home room, the distance between the sound box and the listener is detected to automatically adjust the loudness and volume of the sound box, so that the volume is moderate and the sound quality is basically the same wherever the listener goes. ② The listener's living state is divided into different living states such as active, rest or sleep. The area of the room where the listener stays and the staying time are used to estimate the possible living state, and the automatic switching of the corresponding audio performance effect is realized. For example, keep the melody bright and lively when walking, keep the melody gentle and soft when sitting on the sofa, and turn off when sleeping.

[0028] In order to realize the above two functions of the present application, in the specific implementation process, the Doppler radar sensor module comprises a transmitting antenna, a first receiving antenna, a second receiving antenna and a first digital signal processor DSP1; the first digital signal processor DSP1 mixes the transmitting signal of the transmitting antenna with the first receiving signal of the first receiving antenna to obtain an intermediate frequency signal IF, and performs discrete Fourier transform after low-pass filtering to calculate the relative distance between the listener and the sound box; the first digital signal processor also calculates the delay τ of the listener to the two receiving antennas based on the generalized cross-correlation-phase transform method according to the first receiving signal received by the first receiving antenna and the second receiving signal received by the second receiving antenna, so as to solve the azimuth angle θ. That is, the Doppler radar sensor module is a front-end data acquisition and operation component for realizing distance and azimuth tracking, and the integrated ADC unit inside is used to realize analog-to-digital conversion and transmit to the first digital signal processor DSP1 for subsequent processing. The first digital signal processor DSP1 is the core component for realizing ranging and orientation algorithms, and the measured distance or specific position is discretized to form an index number, which is then transmitted to the main control module of the sound box through a serial port to execute the automatic adjustment of the audio output of the sound box and call the corresponding audio adjustment scheme. The MCU of the sound box is responsible for human-computer interaction, user interface control, initialization of various main devices, power on / off, etc.

[0029] The second digital signal processor of the sound box divides the room into areas, such as Figure 4 As shown in the middle, the room is divided into nine grids, and corresponding index values are assigned, and the staying time of the listener in the corresponding area is detected by the Doppler radar sensor module to estimate the active, rest and sleep living states; when it is judged to be an active living state, the melody is automatically switched to a bright and lively one; when it is judged to be a rest living state, the melody is automatically switched to a gentle and soft one; when it is judged to be a sleep living state, the machine is automatically switched off.

[0030] As Figure 3 shown in the figure, when the Doppler radar sensor module detects the azimuth angle θ and the distance d of the listener relative to the front of the sound box, it can be determined that the listener is located in Figure 4 the square shown in the nine-square diagram, the approximate area where the sofa and bed are located is determined, and the corresponding index value δ_i is assigned, and when it is determined that the rest reaches a certain length of time, the preset audio rendering scheme is called by the first digital signal processor DSP1 for corresponding processing to generate the corresponding music melody; otherwise, the filter l_i is called for automatic adjustment of the loudness. Where ∅_max represents the maximum angle that the radar can detect, and the typical value is 90 degrees.

[0031] The specific method for obtaining the azimuth angle θ is to first make the first receiving antenna and the second receiving antenna receive discrete signals , The generalized cross-correlation (GCC) in the frequency domain is based on the generalized cross-correlation-phase transform method (GCC-PHAT) to solve the time delay τ of the listener to the two receiving antennas to solve the azimuth angle θ. This method is simple in calculation, small in time delay, strong in tracking ability, and also suitable for indoor environments with certain reverberation. As shown in formula 1:

[0032] (Formula 1)

[0033] Wherein:

[0034] , are the frequency domain representations of the signals , after discrete Fourier transform, the variable is the frequency point index, is the upper limit of integration, and , is the number of sampling points for the first discrete Fourier transform; is the frequency resolution, and , is the sampling rate; is the discrete time number, ;

[0035] Formula 1 is actually equivalent to the inverse fast Fourier transform of the frequency domain product of the two receiving signals.

[0036] Then the time delay estimation is: (Formula 2)

[0037] In order to improve the anti-noise and anti-reverberation performance and the time delay accuracy, the phase transform method (PHAT) is introduced, that is, the weighting function is used, and formula 1 is converted into the following formula:

[0038] (F3)

[0039]

[0040] wherein: is the conjugate of ; from equation (2), the time delay estimate of the listener to the two received signals , the echo distance difference : , =340m / s, the speed of sound; since the antenna spacing is much smaller than the distance of the listener to the antenna , the echo signal can be considered as a plane wave. From Figure 3 , the horizontal azimuth angle , is the spacing between the receiving antennas.

[0041] The present application is based on the equal loudness curve, and makes frequency response compensation to realize audio equal loudness adjustment. The listener's feeling of music is affected by both the objective change of frequency response caused by distance change and the subjective feeling of the human ear's "equal loudness effect", i.e. the human ear's feeling of loudness of each frequency band is not the same under the same sound pressure. The equal loudness curve is used to describe this psychoacoustic characteristic of the human ear.

[0042] As shown in the equal loudness curve in Figure 5 , the sound pressure level-loudness level of the signal near 1kHz is taken as a reference to form the corresponding relationship of loudness level-frequency-sound pressure level. If the frequency response of the listener at a certain place in the room of the full frequency band (20Hz~20KHz) can be measured, and a certain fixed loudness level near the frequency of 1kHz is taken as the desired adjustment target, according to the equal loudness curve shown in Figure 5 , the automatic adjustment of audio equal loudness can be realized by making frequency response compensation to other frequency bands. The specific process is as follows:

[0043] ①Laboratory measurement of amplitude-frequency response at different distances (discretization) to obtain the distance-amplitude-frequency response vector:

[0044] , i is the index number of discrete distance; j is the frequency point index number.

[0045] ②Calculate the compensation vector at each distance

[0046] , which is the equal loudness amplitude-frequency characteristic vector at the desired loudness level . j is the frequency point index number.

[0047] is the frequency response at 1 kHz. The distance r to the listener is given by the Doppler radar sensor module in real time, and the amplitude-frequency response vector mapped to the distance index number i in step ① above is:

[0048] is the frequency response at 1 kHz.

[0049] The difference between the two amplitude-frequency responses is:

[0050]

[0051] The loudness levels are given by Figure 5 When the loudness curves fall above the expected equal-loudness curves, When the loudness curves fall below the expected equal-loudness curves.

[0052] The compensation vector is given by

[0053]

[0054] The final audio signal output:

[0055] (k)

[0056] is the frequency-domain representation of the input signal,

[0057] is the frequency-domain representation of the output signal,

[0058] is the time series, is the frequency point series.

[0059] The inverse Fourier transform reconstructs the time-domain signal:

[0060] ​​​

Claims

1. A method for automatically adjusting the audio output of a speaker based on a Doppler radar sensor, characterized in that... The method first uses a Doppler radar sensor module to track the listener's distance and horizontal azimuth angle indoors, and discretizes them to form an index number. Then, the speaker's second digital signal processor calls a preset audio adjustment scheme according to the index number to automatically adjust the speaker's audio output. The Doppler radar sensor module includes a transmitting antenna, a first receiving antenna, a second receiving antenna, and a first digital signal processor. The first digital signal processor mixes the transmitted signal from the transmitting antenna with the first received signal from the first receiving antenna to obtain an intermediate frequency (IF) signal, and performs a discrete Fourier transform after passing through a low-pass filter to calculate the relative distance between the listener and the speaker. The processor also calculates the delay τ from the listener to the two receiving antennas based on the first received signal received by the first receiving antenna and the second received signal received by the second receiving antenna, using the generalized cross-correlation-phase transform method to solve for the horizontal azimuth angle θ. The second digital signal processor of the speaker divides the room where the speaker is located into areas and assigns corresponding index values. Based on the Doppler radar sensor module detecting and tracking the listener's stay time in the corresponding area, it estimates three life states: active, resting, and sleeping. When the speaker is judged to be in an active life state, it automatically switches to a lively melody; when it is judged to be in a resting life state, it automatically switches to a gentle and soothing melody; and when it is judged to be in a sleeping life state, it automatically switches to power off.

2. The method for automatically adjusting speaker audio output based on a Doppler radar sensor according to claim 1, characterized in that: The method for obtaining the horizontal azimuth angle θ is as follows: First, construct the discrete signals received by the first and second receiving antennas. , Frequency domain generalized cross-correlation (GCC), as shown in Equation 1: (Equation 1) in: , These are signals , Frequency domain representation after discrete Fourier transform, variables For frequency point index, This is the maximum score, and , The number of sampling points for one discrete Fourier transform; For frequency resolution, and , Sampling rate; For discrete time sequence numbers. ; Then, the time delay is estimated as follows: (Equation 2) From Equation 2, we can obtain the time delay estimate of the listener receiving the two signals. echo distance difference : , =340m / s is the speed of sound; horizontal azimuth angle , This refers to the spacing between the receiving antennas.

3. The method for automatically adjusting speaker audio output based on a Doppler radar sensor according to claim 2, characterized in that: Using weighted functions Equation 1 can be transformed into the following equation: (Equation 3) in: for . conjugate.

Images