CN111009255B

CN111009255B - Method, apparatus, computer device and storage medium for eliminating internal noise interference

Info

Publication number: CN111009255B
Application number: CN201911206809.9A
Authority: CN
Inventors: 简华; 郑勇; 许仕哲; 戴志涛; 潘濛濛; 王艳芬
Original assignee: Shenzhen Water World Co Ltd
Current assignee: Shenzhen Water World Co Ltd
Priority date: 2019-11-29
Filing date: 2019-11-29
Publication date: 2022-04-22
Anticipated expiration: 2039-11-29
Also published as: CN111009255A

Abstract

The application discloses eliminate internal noise interference's method for smart machine, smart machine is inside including internal noise source, microphone array and reference microphone, sets up the reference microphone in the first distance scope apart from internal noise source, and the microphone array sets up in the second distance department apart from internal noise source, and first distance is less than the second distance, and the method includes: acquiring noise data corresponding to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation; acquiring noise curves of corresponding internal noise sources received by each reference microphone in real time; calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation amount; and eliminating noise signals corresponding to the actual noise curves from the sound signals acquired by the microphone array. The internal noise signal is determined and eliminated by the reference microphone in cooperation with the microphone array.

Description

Method, apparatus, computer device and storage medium for eliminating internal noise interference

Technical Field

The present application relates to the field of computers, and more particularly, to a method, an apparatus, a computer device and a storage medium for eliminating internal noise interference.

Background

The intelligent device gradually walks into family life, the interaction with the user is realized through voice interaction to the majority intelligent device, but some intelligent devices are because there are the machine part that produces high strength internal noise such as motor, when leading to acquireing user speech information, the internal noise has been picked up, the discernment that leads to sound source location or speech information is inaccurate, influence human-computer interaction's effect, but the existence condition of internal noise is complicated, unable specific discernment and confirmation, lead to the influence that can't get rid of inside strong noise yet today.

Disclosure of Invention

The main purpose of the application is to provide a method for eliminating internal noise interference, and the method aims to solve the technical problem that the influence of internal strong noise cannot be eliminated in the voice interaction of the existing intelligent equipment.

An embodiment of the application provides a method for eliminating internal noise interference, is used for smart machine, smart machine includes internal noise source, microphone array and reference microphone, is in the distance internal noise source first distance within range sets up the reference microphone, the microphone array sets up at the distance the second distance department of internal noise source, first distance is less than the second distance, and the method includes:

acquiring noise data corresponding to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation;

acquiring a noise curve of each reference microphone in real time, wherein the noise curve corresponds to an internal noise source and is received by each reference microphone respectively;

calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation amount;

and eliminating noise signals corresponding to the actual noise curves from the sound signals acquired by the microphone array.

Preferably, the step of calculating, according to the noise curve, the deviation estimation curve and the noise attenuation amount, an actual noise curve corresponding to each of the internal noise sources received by the microphone array includes:

acquiring noise characteristics carried in specified noise, wherein the noise characteristics comprise noise frequency and/or noise loudness amplitude, the specified noise corresponds to a first internal noise source, and the first internal noise source is any one of all internal noise sources emitting noise;

according to the noise characteristics carried in the specified noise, a specified noise attenuation amount and a specified deviation estimation curve corresponding to the first internal noise source are obtained;

calculating to obtain the deviation amount corresponding to each noise loudness value in the specified noise curve according to the specified deviation estimation curve;

subtracting the specified noise attenuation amount corresponding to each noise loudness value and the deviation amount corresponding to each noise loudness value respectively from each noise loudness value in the specified noise curve in a one-to-one correspondence manner to obtain each actual noise loudness value;

connecting the actual noise loudness values to form an actual noise curve corresponding to the first internal noise source;

and according to the calculation process of the actual noise curve corresponding to the first internal noise source, obtaining the actual noise curve corresponding to each internal noise source received by the microphone array.

Preferably, the intelligent device at least includes two dispersedly distributed internal noise sources, the internal noise sources are arranged in one-to-one correspondence with the reference microphones, the internal noise sources include motors, and before the step of obtaining pre-stored noise data corresponding to each of the internal noise sources, the method includes:

respectively acquiring a first noise loudness value of a specified motor at a first rotating speed and a second noise loudness value of the specified motor at a second rotating speed, which are acquired by a specified reference microphone corresponding to the specified motor in the no-load state of the intelligent equipment, wherein the specified motor is any one of all motors in the intelligent equipment;

acquiring a linear distance from the specified motor to the microphone array;

according to the noise attenuation

Respectively calculating a first attenuation loudness value when the first noise loudness value reaches the microphone array and a second attenuation loudness value when the second noise loudness value reaches the microphone array, wherein r represents a linear distance of the specified motor from the microphone array;

acquiring a first actual noise loudness value of the specified motor at the first rotating speed and a second actual noise loudness value of the specified motor at the second rotating speed, which are acquired by the microphone array respectively in an idle state of the intelligent device;

obtaining a first deviation according to the first attenuation loudness value, the first noise loudness value and the first actual noise loudness value, and obtaining a second deviation according to the second attenuation loudness value, the second noise loudness value and the second actual noise loudness value;

calculating a section deviation of a designated section according to the first deviation and the second deviation, wherein the designated section is a section formed by the first noise degree value and the second noise degree value, and the designated section is one of response sections formed by receiving the noise loudness of the designated motor according to the designated reference microphone;

and obtaining the noise sound values of all the motors in the intelligent equipment corresponding to the rotating speeds according to the acquisition process of the first noise sound value and the second noise sound value of the specified motor, and obtaining a specified deviation estimation curve corresponding to a response interval formed by the specified reference microphone receiving the noise loudness of the specified motor according to the calculation process of the interval deviation corresponding to the specified interval.

Preferably, after the step of obtaining the noise degree values corresponding to the respective rotation speeds of all the motors in the smart device according to the obtaining process of the first noise degree value and the second noise degree value of the designated motor, and obtaining the designated deviation estimation curve corresponding to the response interval formed by the designated reference microphone receiving the noise loudness of the designated motor according to the calculation process of the interval deviation corresponding to the designated interval, the method includes:

acquiring coding information and position information of a motor in a rotating state in the intelligent equipment in a current working mode;

acquiring the rotation rate respectively corresponding to each motor in the rotation state in the current working mode;

forming noise data corresponding to the current working mode according to the rotation rate respectively corresponding to each motor in the rotating state and the coding information and the position information of each motor in the rotating state;

and according to the forming process of the noise data corresponding to the current working mode, obtaining the noise data corresponding to each working mode respectively, and forming a noise database.

Preferably, the step of obtaining noise data corresponding to each of the pre-stored internal noise sources includes, before the step of obtaining noise data corresponding to each of the pre-stored internal noise sources:

judging whether the current time is more than a preset threshold value from the latest recorded storage time of the noise database;

if yes, sending out an information prompt for updating the noise database;

judging whether the intelligent equipment is in a no-load state at the current time or not, and the motor is in a rotating state;

and if so, carrying out updating test on the intelligent equipment so as to update the noise data in the noise database.

Preferably, the step of calculating a section deviation of a designated section based on the first deviation and the second deviation includes:

calculating an arithmetic mean of the first deviation and the second deviation;

and taking the arithmetic mean of the first deviation and the second deviation as the interval deviation of the designated interval.

obtaining a first theoretical noise sound degree value according to the first attenuation sound degree value and the first noise sound degree value, and obtaining a second theoretical noise sound degree value according to the second attenuation sound degree value and the second noise sound degree value;

performing linear fitting according to the corresponding relation between the first deviation and the first theoretical noise sound degree value and the corresponding relation between the second deviation and the second theoretical noise sound degree value to obtain a linear equation;

taking the linear equation as a formula for calculating the interval deviation of the specified interval;

and calculating deviation values corresponding to the theoretical noise sound values in the specified interval according to the linear equation.

The embodiment of this application still provides a device of eliminating internal noise interference, assembles in smart machine, smart machine includes internal noise source, microphone array and reference microphone, is in the distance set up in the first distance range of internal noise source the reference microphone, the microphone array sets up at the distance the second distance department of internal noise source, first distance is less than the second distance, and the device includes:

the first acquisition module is used for acquiring noise data which respectively correspond to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation;

the second acquisition module is used for acquiring noise curves of corresponding internal noise sources received by the reference microphones respectively in real time;

the first calculation module is used for calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation amount;

and the elimination module is used for eliminating the noise signals corresponding to the actual noise curves from the sound signals acquired by the microphone array.

Embodiments of the present application further provide a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

Embodiments of the present application also provide a computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the method as described above.

The embodiment of the application is provided with the reference microphone near the equipment internal noise source generating certain intensity noise to acquire the noise loudness value of the internal noise source adjacent to the reference microphone, form a noise curve, and determine the noise attenuation amount and deviation estimation curve of the noise reaching the microphone array to determine the noise signal of the internal noise included in the sound signal collected by the microphone array, thereby achieving the effect of accurately eliminating the influence of the internal noise.

Drawings

FIG. 1 is a flow chart illustrating a method for eliminating internal noise interference according to an embodiment of the present application;

FIG. 2 is a schematic structural diagram of an apparatus for eliminating internal noise interference according to an embodiment of the present application;

FIG. 3 is a schematic structural diagram of a first computing module according to an embodiment of the present application;

FIG. 4 is a schematic structural diagram of an apparatus for eliminating internal noise interference according to another embodiment of the present application;

FIG. 5 is a schematic block diagram of a third computing module according to an embodiment of the present application;

FIG. 6 is a schematic block diagram of a third computing module according to another embodiment of the present application;

FIG. 7 is a schematic structural diagram of an apparatus for eliminating internal noise interference according to yet another embodiment of the present application;

FIG. 8 is a schematic structural diagram of an apparatus for eliminating internal noise interference according to yet another embodiment of the present application;

FIG. 9 is a schematic structural diagram of an apparatus for eliminating internal noise interference according to yet another embodiment of the present application;

FIG. 10 is a block diagram of an embodiment of a storage medium provided in the present application;

FIG. 11 is a block diagram illustrating an embodiment of a computer device provided herein;

fig. 1A is a schematic structural diagram of an intelligent sweeping robot according to an embodiment of the present application without a reference microphone;

fig. 1B is a schematic structural diagram of an intelligent sweeping robot according to an embodiment of the present application, with a reference microphone added.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

Referring to fig. 1, 1A and 1B, a method for canceling internal noise interference according to an embodiment of the present application is applied to a smart device, where the smart device includes an internal noise source, a microphone array and a reference microphone, the reference microphone is disposed within a first distance range from the internal noise source, the microphone array is disposed at a second distance from the internal noise source, and the first distance is smaller than the second distance, and the method includes:

s1: and acquiring noise data corresponding to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation.

The intelligent equipment comprises but is not limited to intelligent sweeping robots, intelligent washing machines, intelligent range hoods and other equipment which have strong internal noise and need to receive external user voice for voice interaction. The internal noise source includes a motor. The internal strong noise includes but is not limited to strong noise generated by a motor, and the microphone array is an array formed by arranging a plurality of microphones according to a certain regular layout, and is used for receiving external user voice so as to perform voice interaction. The reference microphone is located within a first distance from the motor, for example within 10 cm of the first distance, to avoid a large error between the loudness of the noise picked up by the reference microphone and the loudness of the noise actually emitted by the motor, preferably the reference microphone is located close to the motor. The pre-stored noise data is related to internal noise obtained by testing in a testing environment, and the testing environment comprises the following elements: the environmental sound is lower than 40dB, and meanwhile, the voltage of the motor is stabilized, and the no-load running state of the motor is kept. Noise collection is carried out to each motor under above-mentioned test environment to combine the different mode of intelligence robot of sweeping the floor to carry out the noise collection in order to form noise data. The noise database includes noise data of the motor in each operating mode, and the noise data includes an attenuation difference (i.e., a noise attenuation amount) between an actual loudness of the motor noise collected by the microphone array and the motor noise collected by the reference microphone, a deviation estimation curve, and the like. The application describes in detail by taking the elimination of strong noise generated by an internal motor of the intelligent sweeping robot as an example, the influence of the strong noise in the intelligent sweeping robot is removed in an auxiliary mode by arranging a reference microphone at the motor, a microphone array of the intelligent sweeping robot is arranged in the center of the internal space of the intelligent sweeping robot, and the reference microphone is arranged close to the motor so as to collect noise generated by the movement of a part driven by the motor and the noise generated by the rotation of the motor through the reference microphone. The above-mentioned working modes include, but are not limited to, a mopping mode, a sweeping mode, etc. The running states of the motors in different working modes are different, for example, all the motors in the intelligent sweeping robot are almost in a rotating state in the sweeping mode, the sweeping mode comprises a powerful mode, a standard mode and an energy-saving mode, and each motor of the intelligent sweeping robot in the powerful mode rotates at a high speed for about 70 min; in the standard mode, the intelligent sweeping robot works for about 100min when each motor rotates at a medium speed; and under the energy-saving mode, each motor of the intelligent sweeping robot works for 110min at a low speed.

S2: and acquiring a noise curve of the corresponding internal noise source received by each reference microphone in real time.

And the noise loudness values of all adjacent time points at the same rotating speed are connected through a smooth curve to form a noise curve of a reference microphone for receiving the corresponding internal noise source. The correspondence refers to the association of the internal noise source with the reference microphone to which it is next to. The noise curve obtained by real-time acquisition includes the dispersion points on the curve and also includes a curve composed of the dispersion points. In fact, the noise curve can be seen as a further point at a time/short period of time, but as a curve of points over a longer time span.

S3: and calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation.

The noise attenuation amount and the deviation estimation curve of the noise sound degree value of the same motor at different rotating speeds are obviously different, and the noise attenuation amount and the deviation estimation curve of the motors with different types of structures at the same rotating speed are also different. Therefore, each motor in the corresponding intelligent sweeping robot respectively has a pre-stored deviation estimation curve. The deviation estimation curve is obtained by testing corresponding deviation values under various theoretical loudness values under a test environment, the deviation values are obtained by subtracting theoretical noise loudness values of the microphone array from actual noise loudness values collected by the microphone array, and the theoretical noise loudness values are obtained by subtracting noise attenuation values from noise loudness values collected by reference microphones. Under the working state, actual noise sound values acquired by the microphone array are obtained by subtracting the deviation amount and the noise attenuation amount estimated through the deviation estimation curve from the noise sound values on the noise curve acquired by the reference microphone, and the curve corresponding to each actual noise sound value is the actual noise curve. In other embodiments of the present application, noise degree values within a reference microphone loudness interval and actual noise degree values actually acquired by a microphone array corresponding to the noise degree values may also be obtained by intensive collection in a test environment, deviation amounts corresponding to the noise degree values are obtained by calculation, noise attenuation amounts corresponding to the noise degree values are obtained by a noise attenuation model, and the noise attenuation amounts are arranged in a list according to an association correspondence relationship between motors through the noise degree values, the actual noise degree values, the deviation amounts, and the noise attenuation amounts to form a noise data list, and noise data lists corresponding to a plurality of motors form a noise database. Under the working state, the estimation of the actual noise sound degree value actually received by the microphone array can be carried out by directly calling the related data recorded in the noise data list.

S4: and eliminating noise signals corresponding to the actual noise curves from the sound signals acquired by the microphone array.

The method includes the steps that a reference microphone is assembled near an internal noise source generating internal noise with intensity, the noise loudness value of the internal noise source close to the reference microphone is obtained, the noise attenuation is calculated according to the distance between the internal noise source and a microphone array, and the noise attenuation is subtracted from the noise loudness value collected by the reference microphone to obtain the theoretical noise sound degree value when the noise of the internal noise source reaches the microphone array. The noise attenuation amount and the deviation estimation curve of the noise reaching the microphone array are determined, so that the noise signal of the internal noise included in the sound signal collected by the microphone array, namely the curve signal corresponding to the actual noise sound degree value is determined, and the noise signal of the internal noise is eliminated from the sound signal in a reverse wave mode (for example), so that the effect of accurately eliminating the influence of the internal noise is achieved. Since the noise attenuation amount between the reference microphone and the microphone array for the same noise level theoretically satisfies a noise attenuation model in which the attenuation amount is proportional to the distance, the noise attenuation amount is a fixed value in the case where the distance between the reference microphone and the microphone array is fixed. However, in practice, noise is transmitted from the motor position to the microphone array, and there is not only a noise attenuation amount due to the transmission distance, but also a deviation amount due to factors such as reflection and absorption of noise by the internal structure material of the smart device, which is expressed as a difference between the actual noise loudness value received by the microphone array and the theoretical noise loudness value. The theoretical noise sound degree value refers to a sound degree value which should be acquired by the microphone array theory under the condition of normal attenuation, and the sound degree value is a sound degree value obtained by subtracting attenuation from a noise degree value acquired by a reference microphone. Actual factors such as structural design in the intelligent device can interfere with the noise sound degree value to generate a deviation epsilon. For example, the deviation between the actual and theoretical noise loudness values is approximately 1-2 db. The theoretical noise value and the deviation value of the microphone array may have a fixed correlation, such as a straight-line correlation or a curve correlation, and the specific correlation depends on the fitting algorithm used. Taking noise data of theoretical noise sound degree values of a microphone array as a curve and a fitting algorithm as straight line correlation as an example, dividing the whole section of the curve into 10 sections, randomly selecting loudness sections (55, 60) as an example, wherein the section interval is 5, solving deviation amounts between the two endpoints of 55 and 60 and actual noise loudness values respectively to obtain numerical values of the two deviation amounts, and taking an arithmetic mean as the deviation amounts of all points in the section (55, 60). Therefore, the method is popularized to the loudness range corresponding to the whole curve, all interval deviations are calculated and added into a noise database. In the practical use of the intelligent sweeping robot, the corresponding deviation amount and the noise attenuation amount are subtracted from the noise loudness value collected by the reference microphone in the corresponding interval, so that the estimation of the actual noise loudness value of the motor noise actually received by the microphone array is realized, the noise signal eliminated from the sound signal collected by the microphone array is obtained, and the internal noise is eliminated. The deviation estimation curve and the noise data are both logarithmic curves, the logarithmic curves are divided into a plurality of sections by a straight line fitting method, each section can be approximately fitted into a straight line, a straight line equation corresponding to the section is obtained by calculating two endpoint values of the section, the deviation amount corresponding to the section can be basically determined by straight line fitting, the calculated amount is small, and quick response is easy. And then obtaining a deviation amount corresponding to the theoretical noise sound degree value according to any theoretical noise sound degree value in the interval. According to the other embodiments of the application, the polynomial coefficients can be obtained through derivation according to a least square curve fitting method, if the polynomial coefficients corresponding to the two curves are the same or similar, the two curves are considered to be the same, the calculation deviation amount is more accurate, but the calculation amount is large, and quick response is not facilitated.

Further, the step S3 of calculating an actual noise curve corresponding to each of the internal noise sources received by the microphone array according to the noise curve, the deviation estimation curve and the noise attenuation amount includes:

s31: acquiring noise characteristics carried in specified noise, wherein the noise characteristics comprise noise frequency and/or noise loudness amplitude, the specified noise corresponds to a first internal noise source, and the first internal noise source is any one of all internal noise sources emitting noise.

S32: and calling a specified noise attenuation amount and a specified deviation estimation curve corresponding to the first internal noise source according to the noise characteristics carried in the specified noise.

S33: and calculating to obtain the deviation amount corresponding to each noise loudness value in the specified noise curve according to the specified deviation estimation curve.

S34: and subtracting the specified noise attenuation amount corresponding to each noise loudness value and the deviation amount corresponding to each noise loudness value respectively from each noise loudness value in the specified noise curve in a one-to-one correspondence manner to obtain each actual noise loudness value.

S35: and connecting the actual noise loudness values to form an actual noise curve corresponding to the first internal noise source.

S36: and according to the calculation process of the actual noise curve corresponding to the first internal noise source, obtaining the actual noise curve corresponding to each internal noise source received by the microphone array.

The noise characteristics are used to distinguish the internal noise sources generating noise, for example, if the internal noise sources are motors, the internal structure, material composition, and the like corresponding to each motor are different, and the corresponding noise characteristics are different. By determining which motor the noise comes from, the corresponding deviation estimation curve and the noise attenuation amount of the motor can be selected, and the noise estimation accuracy is improved. And respectively subtracting the noise attenuation amount and the deviation amount estimated according to the selected deviation estimation curve from each noise degree value on the noise curve acquired by the reference microphone corresponding to the motor to obtain the actual noise degree value acquired by the microphone array, obtaining the actual noise curve according to each actual noise degree value, and realizing the association between the noise curve acquired by the reference microphone and the actual noise curve acquired by the microphone to determine the noise signals contained in the sound signals received by the microphone array and accurately eliminate the internal noise in a reverse wave mode.

The embodiment provided by the application matches the motor noise acquired by the reference microphone to a linear equation or an arithmetic mean for deviation estimation corresponding to the motor noise interval prestored in a noise database under the test condition. And calculating and removing the actually collected noise signals of the microphone array in the noise interval according to a known linear equation or arithmetic mean and a known noise loudness value collected by the reference microphone so as to purify the sound signals collected by the microphone array. Under the working state, the reference microphone and the microphone array are connected through communication signals to carry out information transmission, and the motor noise collected by the reference microphone reaches the microphone array through signal transmission so as to ensure the real-time transmission of the noise without attenuation and deviation, so that an accurate noise reference line is provided for the noise analysis in the microphone array. Such signaling includes, but is not limited to, current or data lines. The embodiment that this application provided uses adaptive filtering algorithm to carry out equipment internal noise elimination through DSP digital processing unit, and adaptive filtering algorithm can solve the above-mentioned similar motor and use the variable factor that causes, for example along with the live time, motor noise has been the factor such as bigger and bigger. The method can well operate in an unknown environment, track the capability of input statistics changing along with time, and adaptively update a performance function, so that the filtering effect is optimal. The generated internal noise is offset, so that the pretreatment of strong noise in equipment by human-computer interaction is completed, the human-computer interaction can be better completed, and the interaction accuracy and the interaction quality are improved. After internal noise of the sound signal is eliminated, the sound signal is input to a voice recognition system, sound source positioning and voice translation and interaction are carried out, and the accuracy of man-machine interaction is improved.

Further, the intelligent device at least includes two dispersedly distributed internal noise sources, the internal noise sources are arranged in one-to-one correspondence with the reference microphones, the internal noise sources include motors, and before the step S1 of obtaining pre-stored noise data corresponding to each of the internal noise sources, the method includes:

s11: respectively acquiring a first noise loudness value of a specified motor at a first rotating speed and a second noise loudness value of the specified motor at a second rotating speed, which are acquired by a specified reference microphone corresponding to the specified motor in the no-load state of the intelligent device, wherein the specified motor is any one of all motors in the intelligent device.

S12: and acquiring the linear distance between the specified motor and the microphone array.

S13: according to the noise attenuation

Calculating a first attenuation loudness value when the first noise loudness value reaches the microphone array and a second noise loudness value when the second noise loudness value reaches the microphone array, respectivelyA second attenuated loudness value with an array of microphones, where r represents a linear distance of the specified motor from the array of microphones.

S14: and acquiring a first actual noise loudness value of the specified motor at the first rotating speed and a second actual noise loudness value of the specified motor at the second rotating speed, which are acquired by the microphone array respectively in the no-load state of the intelligent device.

S15: and obtaining a first deviation according to the first attenuation loudness value, the first noise loudness value and the first actual noise loudness value, and obtaining a second deviation according to the second attenuation loudness value, the second noise loudness value and the second actual noise loudness value.

S16: and calculating a section deviation of a designated section according to the first deviation and the second deviation, wherein the designated section is a section formed by the first noise degree value and the second noise degree value, and the designated section is one of response sections formed by receiving the noise loudness of the designated motor according to the designated reference microphone.

S17: and obtaining the noise sound values of all the motors in the intelligent equipment corresponding to the rotating speeds according to the acquisition process of the first noise sound value and the second noise sound value of the specified motor, and obtaining a specified deviation estimation curve corresponding to a response interval formed by the specified reference microphone receiving the noise loudness of the specified motor according to the calculation process of the interval deviation corresponding to the specified interval.

The microphone array is a disc formed by arranging six microphones, the distances among the six microphones are greatly different from the distances from a reference microphone to the microphone array, so that the attenuation difference among the six microphones can be ignored, and the arithmetic average value of the loudness of the motor noise received by the six microphones respectively is used as the loudness value of the motor noise actually received by the microphone array. Only between the motor and the microphone array is considered, the propagation of noise data being direct waves. The noise attenuation is calculated by the algebraic relation of the noise attenuation model: amount of noise attenuation

r represents distance in decibels (db). The distance between the geometric center position of the microphone array and the motor is known and fixed, the mutual distance between a plurality of microphones of the microphone array is relatively small with the distance from a reference microphone, the loudness of the plurality of microphones of the microphone array can be directly subjected to arithmetic averaging and can be used as the loudness of the microphone array, and therefore the actual loudness value of the noise collected by the microphone array can be obtained by calculating the noise attenuation amount and deviation amount of each motor. The noise loudness is collected by the reference microphone, and meanwhile, the distance between the reference microphone and the motor is small, so that the motor noise loudness collected by the reference microphone is approximately considered to be not attenuated, and therefore when the noise database is built, the difference between the motor noise loudness alpha (db) collected by the reference microphone and the noise loudness beta (db) collected by the microphone array is (alpha-beta) db. Noise characteristics generated by the motor have a range, so that noise loudness acquired by the reference microphone has a loudness range (D1, D2), the microphone array corresponding to the linear relation also acquires a loudness range (D3, D4), shielding exists in equipment, and distance difference exists between the reference microphone and the microphone array, so that physical attenuation of direct waves is compared with noise attenuation obtained by a noise attenuation model to generate certain deviation. Such as: when the reference microphone acquires the loudness of D1, the theoretical value of the loudness acquired by the microphone array is D3, and when the motor noise loudness is actually acquired, the loudness relation value between the reference microphone and the microphone array is distributed around a straight line to form a curve, namely, under the actual condition (beta') is the actual loudness): and forming a deviation estimation curve according to the deviation amount of the same motor corresponding to different noise loudness, wherein epsilon is the deviation amount. Calculating the noise attenuation and the theoretical noise sound degree value received by the microphone array through a motor noise attenuation model, and forming the noise attenuation, deviation estimation curve, noise collection mode, motor position information, coding information, distance between the motor and the microphone array and other data into noise data so as to facilitate the motor to work in a working stateThe noise signal collected in the microphone array is estimated by retrieving the noise data.

Further, the step S16 of calculating a section deviation of the designated section according to the first deviation and the second deviation includes:

s161: calculating an arithmetic mean of the first deviation and the second deviation;

s162: and taking the arithmetic mean of the first deviation and the second deviation as the interval deviation of the designated interval.

And comparing the noise loudness curve actually received by the microphone array with the theoretical noise curve, selecting a relevant interval, and calculating the deviation amount of the theoretical noise degree value and the actual noise degree value. The deviation amount corresponding to the whole interval can be calculated according to the two deviation amounts corresponding to the interval endpoints. The deviation amount in the interval is not too large because of fluctuation, and the arithmetic mean of the deviation amounts of two end points of the interval is used as the corresponding deviation amount of the whole interval, so that the deviation amount between each noise sound value on a theoretical curve and the actual noise sound value on an actual curve can be approximately represented.

Further, the step S16 of calculating the section deviation of the designated section according to the first deviation and the second deviation may specifically include:

s16 a: obtaining a first theoretical noise sound degree value according to the first attenuation sound degree value and the first noise sound degree value, and obtaining a second theoretical noise sound degree value according to the second attenuation sound degree value and the second noise sound degree value;

s16 b: performing linear fitting according to the corresponding relation between the first deviation and the first theoretical noise sound degree value and the corresponding relation between the second deviation and the second theoretical noise sound degree value to obtain a linear equation;

s16 c: taking the linear equation as a formula for calculating the interval deviation of the specified interval;

s16 d: and calculating deviation values corresponding to the theoretical noise sound values in the specified interval according to the linear equation.

And comparing the noise loudness curve actually received by the microphone array with a theoretical noise curve, selecting an interval related to the theoretical noise curve, and performing linear fitting on the deviation estimation in the interval according to the deviation values respectively corresponding to two end points of the interval. For example, the equation of the fitted straight line is y-kx + b, where y represents the theoretical noise sound value, x represents the deviation, k represents the slope of the straight line, and b represents the intercept of the straight line on the y-axis. And substituting the theoretical noise sound degree values and the deviation amounts respectively corresponding to the two end points of the interval into the fitting straight line equation, solving and determining k and b, calculating the deviation amount corresponding to the theoretical noise sound degree value when any theoretical noise sound degree value in the known interval is obtained, and further calculating the noise signal in the sound signal actually acquired by the microphone array by combining the noise attenuation amount.

Further, after the step S17 of obtaining the noise degree values corresponding to the respective rotation speeds of all the motors in the smart device according to the obtaining process of the first noise degree value and the second noise degree value of the designated motor, and obtaining the designated deviation estimation curve corresponding to the response interval formed by the designated reference microphone receiving the noise loudness of the designated motor according to the calculation process of the interval deviation corresponding to the designated interval, the method includes:

s171: acquiring coding information and position information of a motor in a rotating state in the intelligent equipment in a current working mode;

s172: acquiring the rotation rate respectively corresponding to each motor in the rotation state in the current working mode;

s173: forming noise data corresponding to the current working mode according to the rotation rate respectively corresponding to each motor in the rotating state and the coding information and the position information of each motor in the rotating state;

s174: and according to the forming process of the noise data corresponding to the current working mode, obtaining the noise data corresponding to each working mode respectively, and forming a noise database.

The coded information includes a motor number, including but not limited to a unique code when the equipment leaves a factory, and is stored together with the position information of the equipment in the intelligent sweeping robot. The position information takes the plane of the microphone array as the plane of the coordinate xy, the center of the microphone array as the origin of coordinates, and a straight line which is perpendicular to the plane of the coordinate xy and passes through the origin of coordinates as the z-axis of coordinates to form xyz three-dimensional coordinates, and the azimuth and the distance relative to the origin of coordinates are obtained according to the actual layout position of each motor. The noise loudness of each motor at different rotating speeds is stored in the same row, and the noise loudness deviation amount of different motors at the same rotating speed is stored in the same column to form the table-type pre-stored data. And then according to the rotating speed corresponding to each motor in the current working mode, noise data such as noise loudness deviation amount of each motor in the working mode are obtained, noise data taking the working mode as a unit are formed, and the noise data of various working modes are collected to form a noise database.

Further, before the step S1 of obtaining noise data corresponding to each of the pre-stored internal noise sources, the method includes:

s1 a: judging whether the current time is more than a preset threshold value from the latest recorded storage time of the noise database;

s1 b: if yes, sending out an information prompt for updating the noise database;

s1 c: judging whether the intelligent equipment is in a no-load state at the current time and the motor is in a rotating state;

s1 d: and if so, carrying out updating test on the intelligent equipment so as to update the noise data in the noise database.

Factors influencing the noise of the motor are many, for example, the internal structure of the intelligent floor sweeping robot is provided with a shielding object, and the noise of the motor is also faded. The loudness of noise generated by the motor increases along with the lapse of service time under the same rotating speed, so after a time period exceeding a preset threshold value, the pre-stored noise data needs to be tested and updated again to improve the estimation accuracy. The preset threshold is time data such as one month later, half year later, and the like. The motors are in rotation, including any one or more of the motors rotating.

Further, after the step S4 of eliminating the noise signal corresponding to each of the actual noise curves from the sound signal obtained by the microphone array, the noise data includes a noise frequency characteristic and a noise loudness characteristic, and the method includes:

s4a, judging whether the noise data is motor noise according to the noise frequency characteristics;

s4b, if yes, judging that the motor has a fault;

and S4c, giving out a fault alarm and stopping running the intelligent equipment.

The embodiment provided by the application can also judge the fault by identifying the noise data, for example, by identifying the noise frequency in the noise data, and judge whether the data is the noise data. The frequency of motor noise is unchanged in the transmission process, the rated frequency range of a common motor is approximately 50-60HZ, the noise is collected by a microphone, the frequency is subjected to feature extraction from noise audio, then Fast Fourier Transform (FFT) is carried out to obtain a related spectrogram of the motor noise, the motor noise is transmitted to the microphone array, the frequency domain feature of a noise signal is not changed, and the noise generated by the motor can be judged by comparing the frequency domain feature of a noise database with the frequency domain feature of real-time noise if the frequency domain features can be the same, so that the external environment sound and the internal motor noise can be judged and distinguished. However, if the pre-stored deviation estimation curve corresponding to the noise data is not matched with the noise curve, for example, the deviation is far from the original fitting straight line, it indicates that the noise data of the motor is the noise generated by abnormal rotation, and is in a fault state. The faults include faults of internal components of the motor and faults of external environments such as winding or friction.

Referring to fig. 2, an apparatus for canceling internal noise interference according to an embodiment of the present application is mounted on a smart device, the smart device includes an internal noise source, a microphone array and a reference microphone, the reference microphone is disposed within a first distance range from the internal noise source, the microphone array is disposed at a second distance from the internal noise source, the first distance is smaller than the second distance, and the method includes:

the first obtaining module 1 is configured to obtain noise data corresponding to each of the pre-stored internal noise sources, where the noise data includes a deviation estimation curve and a noise attenuation.

And the second obtaining module 2 is configured to obtain, in real time, noise curves of corresponding internal noise sources received by each of the reference microphones.

And the first calculating module 3 is used for calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation amount.

And the eliminating module 4 is used for eliminating the noise signals corresponding to the actual noise curves from the sound signals acquired by the microphone array.

Referring to fig. 3, the first calculation module 3 includes:

an obtaining unit 31, configured to obtain a noise feature carried in a specified noise, where the noise feature includes a noise frequency and/or a noise loudness amplitude, the specified noise corresponds to a first internal noise source, and the first internal noise source is any one of all internal noise sources emitting noise;

the retrieval unit 32 is configured to retrieve a specified noise attenuation amount and a specified deviation estimation curve corresponding to the first internal noise source according to a noise characteristic carried in the specified noise;

the first calculating unit 33 is configured to calculate, according to the specified deviation estimation curve, deviation amounts corresponding to noise loudness values in the specified noise curve;

a first obtaining unit 34, configured to obtain actual noise loudness values by subtracting specified noise attenuation amounts corresponding to the noise loudness values and deviation amounts corresponding to the noise loudness values in the specified noise curve one by one;

a forming unit 35, configured to connect the actual noise loudness values to form an actual noise curve corresponding to the first internal noise source;

the second obtaining unit 36 is configured to obtain, according to a calculation process of an actual noise curve corresponding to the first internal noise source, an actual noise curve corresponding to each internal noise source received by the microphone array.

Referring to fig. 4, the smart device at least includes two dispersedly distributed internal noise sources, the internal noise sources are disposed in one-to-one correspondence with the reference microphones, the internal noise sources include motors, and the apparatus for eliminating internal noise interference includes:

a third obtaining module 11, configured to obtain, in an idle state of the intelligent device, a first noise loudness value of the specified motor at a first rotation speed and a second noise loudness value of the specified motor at a second rotation speed, where the first noise loudness value is collected by a specified reference microphone corresponding to the specified motor, and the specified motor is any one of all motors in the intelligent device;

a fourth obtaining module 12, configured to obtain a linear distance between the specified motor and the microphone array;

a second calculation module 13 for calculating the amount of noise attenuation according to the amount of noise attenuation

a fifth obtaining module 14, configured to obtain, in an idle state of the smart device, a first actual noise loudness value of the specified motor at the first rotation speed and a second actual noise loudness value of the specified motor at the second rotation speed, which are respectively collected by the microphone array;

a first obtaining module 15, configured to obtain a first deviation according to the first attenuation loudness value, the first noise loudness value, and the first actual noise loudness value, and obtain a second deviation according to the second attenuation loudness value, the second noise loudness value, and the second actual noise loudness value;

a third calculating module 16, configured to calculate a section deviation of a specified section according to the first deviation and the second deviation, where the specified section is a section formed by the first noise degree value and the second noise degree value, and the specified section is one of response sections formed by receiving the noise loudness of the specified motor according to the specified reference microphone;

a second obtaining module 17, configured to obtain noise values of all motors in the intelligent device corresponding to the respective rotation speeds according to an obtaining process of the first noise value and the second noise value of the specified motor, and obtain a specified deviation estimation curve corresponding to a response interval formed by receiving the noise loudness of the specified motor by the specified reference microphone according to a calculation process of an interval deviation corresponding to the specified interval.

r represents distance in decibels (db). The distance between the geometric center position of the microphone array and the motor is known and fixed, the mutual distance between a plurality of microphones of the microphone array is relatively small with the distance from the reference microphone, the loudness of the plurality of microphones of the microphone array can be directly subjected to arithmetic averaging to be used as the loudness of the microphone array, and therefore, the noise attenuation of each motor is calculatedAnd the amount and the deviation amount can be obtained, and the actual noise loudness value collected by the microphone array can be obtained. The noise loudness is collected by the reference microphone, and meanwhile, the distance between the reference microphone and the motor is small, so that the motor noise loudness collected by the reference microphone is approximately considered to be not attenuated, and therefore when the noise database is built, the difference between the motor noise loudness alpha (db) collected by the reference microphone and the noise loudness beta (db) collected by the microphone array is (alpha-beta) db. Noise characteristics generated by the motor have a range, so that noise loudness acquired by the reference microphone has a loudness range (D1, D2), the microphone array corresponding to the linear relation also acquires a loudness range (D3, D4), shielding exists in equipment, and distance difference exists between the reference microphone and the microphone array, so that physical attenuation of direct waves is compared with noise attenuation obtained by a noise attenuation model to generate certain deviation. Such as: when the reference microphone acquires the loudness of D1, the theoretical value of the loudness acquired by the microphone array is D3, and when the motor noise loudness is actually acquired, the loudness relation value between the reference microphone and the microphone array is distributed around a straight line to form a curve, namely, under the actual condition (beta') is the actual loudness): and forming a deviation estimation curve according to the deviation amount of the same motor corresponding to different noise loudness, wherein epsilon is the deviation amount. And calculating a noise attenuation amount and a theoretical noise degree value received by the microphone array through a motor noise attenuation model, and forming noise data by using the noise attenuation amount, a deviation estimation curve, a noise acquisition mode, motor position information, coding information, the distance between the motor and the microphone array and other data so as to estimate a noise signal acquired in the microphone array by calling the noise data in a working state.

Referring to fig. 5, the third calculation module 16 includes:

a second calculating unit 161 for calculating an arithmetic mean of the first deviation and the second deviation;

a first operation unit 162 configured to use an arithmetic mean of the first deviation and the second deviation as the section deviation of the designated section.

Referring to fig. 6, the third computing module 16 may also operate as follows, specifically including:

a third obtaining unit 16a, configured to obtain a first theoretical noise sound value according to the first attenuation loudness value and the first noise sound value, and obtain a second theoretical noise sound value according to the second attenuation loudness value and the second noise sound value;

the fitting unit 16b is configured to perform linear fitting according to the corresponding relationship between the first deviation and the first theoretical noise sound degree value and the corresponding relationship between the second deviation and the second theoretical noise sound degree value to obtain a linear equation;

a second determining unit 16c for determining the linear equation as a formula for calculating the section deviation of the designated section;

and the third calculating unit 16d is configured to calculate, according to the linear equation, deviation amounts corresponding to the theoretical noise sound values in the specified interval respectively.

Referring to fig. 7, the apparatus for canceling internal noise interference includes:

the sixth obtaining module 171 is configured to obtain coding information and position information of a motor in a rotating state in the smart device in the current operating mode.

And a seventh obtaining module 172, configured to obtain rotation rates corresponding to the motors in the rotation states in the current working mode.

And a forming module 173 for forming noise data corresponding to the current operating mode according to the rotation rate corresponding to each motor in the rotation state, and the coding information and the position information of each motor in the rotation state.

And a third obtaining module 174, configured to obtain noise data corresponding to each working mode according to a forming process of the noise data corresponding to the current working mode, and form a noise database.

Referring to fig. 8, in still another embodiment, an apparatus for canceling internal noise interference includes:

the first judging module 1a is used for judging whether the storage time of the latest record of the current time and the noise database exceeds a preset threshold value.

And the prompt module 1b is used for sending an information prompt for updating the noise database if the noise database is updated.

And the second judging module 1c is used for judging whether the intelligent equipment is in a no-load state at the current time and the motor is in a rotating state.

And the updating module 1d is used for updating and testing the intelligent equipment if the noise data in the noise database is correct, so that the noise data in the noise database can be updated.

Referring to fig. 9, the noise data includes a noise frequency characteristic and a noise loudness characteristic, and in still another embodiment, the apparatus for eliminating internal noise interference includes:

and the third judging module 4a is used for judging whether the noise data is motor noise according to the noise frequency characteristic.

And the judging module 4b is used for judging that the motor has a fault if the motor has the fault.

And the alarm module 4c is used for giving out a fault alarm and stopping the intelligent equipment.

Referring to fig. 10, the present application further provides a storage medium 100, in which a computer program 200 is stored in the storage medium 100, and when the computer program runs on a computer, the computer is enabled to execute the method for eliminating the internal noise interference described in the above embodiments.

Referring to fig. 11, the present application further provides a computer device 300 containing instructions, which when run on the computer device 300, causes the computer device 300 to execute the method for eliminating internal noise interference described in the above embodiments through a processor 400 disposed inside the computer device 300.

Those skilled in the art will appreciate that the apparatus for canceling internal noise interference of embodiments of the present application and the apparatus referred to above for performing one or more of the methods described in the present application. These devices may be specially designed and manufactured for the required purposes, or they may comprise known devices in general-purpose computers. These devices have stored therein computer programs or applications that are selectively activated or reconfigured. Such a computer program may be stored in a device (e.g., computer) readable medium, including, but not limited to, any type of disk including floppy disks, hard disks, optical disks, CD-ROMs, and magnetic-optical disks, ROMs (Read-Only memories), RAMs (Random Access memories), EPROMs (Erasable Programmable Read-Only memories), EEPROMs (Electrically Erasable Programmable Read-Only memories), flash memories, magnetic cards, or optical cards, or any type of media suitable for storing electronic instructions, and each coupled to a bus. That is, a readable medium includes any medium that stores or transmits information in a form readable by a device (e.g., a computer).

The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.

Claims

1. A method of canceling internal noise interference for a smart device comprising an internal noise source, an array of microphones, and a reference microphone, the reference microphone being positioned within a first distance from the internal noise source, the array of microphones being positioned at a second distance from the internal noise source, the first distance being less than the second distance, the method comprising:

acquiring noise data corresponding to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation; wherein, the deviation estimation curve is a designated deviation estimation curve corresponding to a response interval formed by the noise loudness; the deviation estimation curve is obtained by testing corresponding deviation values under various theoretical loudness values in a test environment, the deviation values are obtained by subtracting theoretical noise loudness values of the microphone array from actual noise loudness values collected by the microphone array, and the theoretical noise loudness values are obtained by subtracting noise attenuation from noise loudness values collected by a reference microphone;

2. The method of claim 1, wherein the step of calculating an actual noise curve corresponding to each of the internal noise sources received by the microphone array according to the noise curve, the deviation estimation curve and the noise attenuation amount comprises:

3. The method according to claim 1, wherein the intelligent device comprises at least two dispersedly distributed internal noise sources, the internal noise sources are disposed in a one-to-one correspondence with the reference microphones, the internal noise sources comprise motors, and the step of obtaining pre-stored noise data corresponding to each of the internal noise sources comprises:

acquiring a linear distance from the specified motor to the microphone array;

according to the noise attenuation

) Calculating a first attenuation loudness value when the first noise loudness value reaches the microphone array and a second attenuation loudness value when the second noise loudness value reaches the microphone array, wherein r represents a linear distance of the specified motor from the microphone array;

4. The method according to claim 3, wherein the step of obtaining the noise degree values corresponding to the rotation speeds of all the motors in the smart device according to the obtaining process of the first noise degree value and the second noise degree value of the designated motor and obtaining the designated deviation estimation curve corresponding to the response interval formed by the designated reference microphone receiving the noise loudness of the designated motor according to the calculation process of the interval deviation corresponding to the designated interval is followed by:

5. The method of claim 4, wherein the step of obtaining pre-stored noise data corresponding to each of the internal noise sources is preceded by the step of:

if yes, sending out an information prompt for updating the noise database;

6. The method of claim 3, wherein the step of calculating a section deviation of a designated section according to the first deviation and the second deviation comprises:

calculating an arithmetic mean of the first deviation and the second deviation;

7. The method of claim 3, wherein the step of calculating a section deviation of a designated section according to the first deviation and the second deviation comprises:

8. An apparatus for canceling internal noise interference, assembled in a smart device comprising an internal noise source, an array of microphones, and a reference microphone, the reference microphone disposed within a first distance range from the internal noise source, the array of microphones disposed at a second distance from the internal noise source, the first distance being less than the second distance, the apparatus comprising:

the first acquisition module is used for acquiring noise data which respectively correspond to each pre-stored internal noise source, wherein the noise data comprises a deviation estimation curve and noise attenuation; wherein, the deviation estimation curve is a designated deviation estimation curve corresponding to a response interval formed by the noise loudness; the deviation estimation curve is obtained by testing corresponding deviation values under various theoretical loudness values in a test environment, the deviation values are obtained by subtracting theoretical noise loudness values of the microphone array from actual noise loudness values collected by the microphone array, and the theoretical noise loudness values are obtained by subtracting noise attenuation from noise loudness values collected by a reference microphone;

the calculation module is used for calculating actual noise curves corresponding to the internal noise sources received by the microphone array according to the noise curves, the deviation estimation curves and the noise attenuation amount;

9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.