DE102017203469A1 - A method and a device for noise removal of audio signals and a voice control of devices with this Störfreireiung - Google Patents

Abstract

Noise suppression device comprising
a detection module (55) configured to detect an operating state of at least one component (11, 12, 13, 14, 21, 22, 23, 24, 25) generating noise in at least one operating state,
a control module (54) configured to generate control parameters (540) as a function of the detected operating state,
a signal processing module (53) configured to process a supplied audio signal (510) in accordance with a processing mode controllable via the control parameters (540) and to output an audio user signal (530) obtained by the processing
and methods for doing so.

Description

State of the art

The invention relates to a method and a device for interference-freeing of audio signals and a voice control of devices with this Störergreiung according to the independent claims.

Speech control methods, ie for controlling devices by means of spoken voice commands, are used, for example, in the field of automotive electronics, for example for operating driver information systems operated in vehicles for selecting functions and for further operation within selected functions, for example Destination input in vehicle navigation systems used.

In disturbed environments, the device-side interpretation of the speech commands recorded via a microphone is hampered by noise components contained in the microphone signals.

Out EP 0 411 360 A1 For example, there are known a method and an arrangement for eliminating interference from voice signals recorded in a motor vehicle for voice control of a device. For this purpose, on the one hand recorded an audio signal, which also includes sounds from the interior of the vehicle in addition to the spoken voice commands. Furthermore, noises of a noise source are recorded by means of a separate microphone. The audio and the separately recorded noise signal are supplied to a signal processing, which includes an adaptive filter. The object of the signal processing is to isolate the speech signals representing the speech commands from the recorded audio signal in order to use these for controlling the device. The largely noise-free speech signal obtained with the signal processing is compared on the one hand with the noise signal and on the other hand with the audio signal, and the comparison results are used to control the adaptive filters.

This and comparable approaches for suppressing interference signal components contained in audio signals, also referred to as noise cancellation systems, are designed to continuously record ambient signals, ie audio signals, and then to filter out the interference signals. The ambient signal is continuously recorded, processed and filtered.

Advantages of the invention

The invention with the features of the independent claims has the advantage that an isolation of an audio useful signal from an audio signal, which contains in addition to the audio useful signal noise components, is improved and facilitated.

The invention pursues the approach to obtain information about an operating condition of devices or devices that cause noise during operation, preferably from existing knowledge about a control of these components or devices and to set a signal processing, even in the presence of this information to compensate for the generated noise , Findings about a control of such components or devices can be obtained from a switch position for such a component or a control signal for controlling such a component, which is present on a data bus. Thus, the signal processing can for example be preset to an expected noise of a component or device and already started as soon as the relevant control signal or the switch position or the like is detected. In contrast to known approaches, which turn off an adaptation to existing noise, thus a faster response and more efficient compensation of noise is possible.

In the context of the preferred application of the present invention, namely voice control of vehicle-operated devices or devices, the invention has the advantage that the recognition of voice commands spoken by a user, also referred to as voice instructions, improves the operation of these devices or devices, is made easier and faster. In this case, the invention has the further advantage that the identification of composite, overlapping, generated by components in the vehicle noise is automated and the filtering of noise and thus isolating the voice commands is optimized as a useful signal.

For this purpose, according to the invention for noise suppression in an audio signal, an operating state of at least one component which generates noises in at least one operating state is detected,
Generates control parameters in dependence on the detected operating state, processes the audio signal according to a processing mode to an audio user signal, wherein the processing mode is controlled by the control parameters, and finally outputs the audio audio signal thus obtained. The method can be advantageous in Form of a computer program or a module of an operating program of a control device to be executed, which is processed by a processor of the control unit.

For this purpose, a noise suppression device according to the invention comprises a detection module configured to detect an operating state of at least one component generating noise in at least one operating state, a control module configured to generate control parameters as a function of the detected operating state and
a signal processing module configured to process a supplied audio signal in accordance with a processing mode controllable via the control parameters and to output an audio user signal obtained by the processing. The said modules can be implemented, at least partially, preferably in the form of computer programs or computer program modules, which are processed by a processor of the control unit.

The noise reduction device can advantageously be brought into use in voice control, that is, with spoken voice commands operable devices or devices. By isolating the speech signal from the recorded with a sound transducer of the device or device to be used audio signal whose device-side detection is improved and accelerated. Error detections that trigger a misinterpretation and thus incorrect operation of the device or the device or that require a repeated input can be reduced or avoided altogether.

Particularly advantageously, the noise suppression according to the invention can thus be used in disturbed environments, such as voice-operated multimedia or driver information systems in motor vehicles, for example for selecting a device function and / or for further operation within a selected function, such as the voice input of an address, which subsequently as destination for a route calculation and route guidance from the current location to the destination.

Advantageously, the noise suppression device has a memory in which the operating states of the at least one component control parameters are assigned. In this case, the control module is designed to read out the control parameters assigned in the memory as a function of a detected operating state and feed them to the signal processing module.

Advantageously, the signal processing module is further configured to isolate a noise component from the audio signal and to output noise component parameters characterizing this noise component, and to configure the control module to adapt the control parameters stored in the memory as a function of the noise parameters.

In this way, control parameters that are used as a basis for modeling or simulating a noise of a specific noise source in the signal processing, on the actual circumstances, such as occupancy or loading of the vehicle interior, which has an effect on the characteristics of the noise, be adjusted. As a result, this leads to a more exact modeling of the noise and thus to a more precise isolation of the audio signal in the signal processing.

Synthesizing a noise simulation or a noise model for the noise component of the audio signal from the control parameters read from the memory enables a particularly simple isolation of the audio signal from the audio signal, in the simplest case by subtracting the modeled noise from the recorded audio signal. As a result, this also leads to a particularly precise isolation of the audio user signal and thus the voice commands from the recorded audio signal.

Advantageously, the signal processing module has a filter module for processing the audio signal, and the control parameters comprise filter parameters for the filter module. In this case, the filter module is preferably embodied in the form of an adaptive filter and designed to start the processing of the audio signal with filter parameters which are read from the memory by the control module as a function of a detected operating state. In this way, the duration of the snapping of the adaptation algorithm to a noise component is considerably shortened. In addition, the adaptation of the filter can be directed to a specific one of several possible interference signal components in the audio signal. In the further course of the signal processing of the audio signal, the adaptation algorithm of the adaptive filter can automatically further adapt the start filter parameters used for its initialization.

list of figures

• Die einzige 1 zeigt ein Blockschaltbild des erfindungswesentlichen Teils einer erfindungsgemäßen Einrichtung zur Störunterdrückung

Embodiments of the invention are illustrated in the figures and are explained in more detail below.

• The only 1 shows a block diagram of the invention essential part of a device according to the invention for interference suppression

Description of the embodiments

In 1 is a block diagram of a device 5 for noise suppression in a recorded audio signal.

In a shielded noise environment, such as in a passenger compartment 1 a passenger car (car) cause certain vehicle-operated devices and vehicle components noises that a detection and implementation of a vehicle occupant, such as the driver 3 spoken speech commands 31 by a device to be operated by voice control 4 can complicate.

The device 5 includes at least one sound pickup 51 , Preferably a microphone, which is arranged in a vehicle interior and serves, by a user 3 For example, the driver, spoken voice instructions or voice commands 31 take. In the preferred embodiment described herein, there are at least two microphones 51 and 52 intended. These are arranged here in the headliner of the vehicle interior in front of the head of the driver, it being assumed that the driver is the main user for the voice control of the voice-controlled devices or devices. Another placement of the sound pickup 51 and 52 or the use of additional, additional sound pickup is possible and depending on the underlying application scenario makes sense. The first microphone 51 generates a first audio signal 510, the second microphone 52 a second audio signal 520 ,

In addition to the spoken voice commands 31 or voice instructions takes the sound pickup 51 , or present take the microphones 51 and 52 also more in the vehicle interior 1 existing noise 29 on. These existing in the vehicle interior 1 noise 29 are generated by other components in their operation.

• - ein Scheibenwischer 11, wobei sich die davon erzeugten Geräusche auch in Abhängigkeit einer Betriebsstufe, also beispielsweise Intervallbetrieb, langsamer Dauerbetrieb, schneller Dauerbetrieb unterscheiden
• - ein elektrischer Fensterheber 12, dessen elektromotorischer Antrieb je nach Bewegungsrichtung des Fensters unterschiedlichen Belastungen unterliegt und daher auch unterschiedliche Geräusche erzeugt,
• - eine Lüftung oder Klimaanlage 13, welche infolge der Luftströmung durch Luftführungen und -austritte Strömungsgeräusche erzeugt, wobei die Geräusche von einer Lüftungsstufe, daneben aber auch von einer Einstellung der Luftaustrittsöffnungen abhängen,
• - ein Fahrtrichtungsanzeiger 14, auch als Blinker bezeichnet, oder die zur Anzeige eines laufenden Blinkvorgangs synthetisch erzeugten und im Fahrzeug wiedergegebenen Klick-Geräusche und
• - eine elektromotorische und/oder elektropneumatische Sitzverstellung und - Anpassung.

Such devices or vehicle components, also referred to as components, which in operation or in a specific operating state noises, here as noise 29 designate, generate, for example - and without limitation to these examples -

• - a windscreen wiper 11 , wherein the noise generated therefrom also differ depending on an operating stage, so for example, interval operation, slow continuous operation, fast continuous operation
• - an electric window 12 whose electromotive drive is subject to different loads depending on the direction of movement of the window and therefore also produces different noises,
• - a ventilation or air conditioning 13 which generates flow noise as a result of the air flow through air ducts and outlets, the noise being dependent on a ventilation stage, but also on an adjustment of the air outlet openings,
• - a direction indicator 14 , also referred to as a turn signal, or the click-sounds synthesized to indicate an ongoing flashing operation and reproduced in the vehicle
• - An electromotive and / or electropneumatic seat adjustment and - adaptation.

• - Warnsysteme, wie etwa eine Einparkhilfe 21, die bei Annäherung an ein Hindernis Warnsignale erzeugt oder ein Gurtwarner 22, der bei eingeschalteter Fahrzeugzündung oder bei festgestellter Fahrt des Fahrzeug und gleichzeitig nicht angeschnalltem Fahrzeugführer oder Fahrzeuginsassen ein Warnsignal erzeugt,
• - eine Audioanlage 23, welche bestimmungsgemäß in Rundfunksignalen oder auf einem Datenträger vorhandene Unterhaltungs- oder Informationsinhalte akustisch über im Fahrzeug angeordnete Lautsprecher 231, 232, 233, 234 wiedergibt, die für eine Sprachsteuerung jedoch Störsignale darstellen,
• - ein Fahrzeugnavigationssystem 24, welches bei bevorstehenden Fahrmanövern zum Befolgen einer Fahrroute akustische Zielführungshinweise über im Fahrzeug angeordnete Lautsprecher wiedergibt, und
• - eine Freisprecheinrichtung 25, welche das Sprachsignal eines Gesprächsteilnehmers über Fahrzeuglautsprecher ausgibt.

Other such devices or components that produce noise during operation, for example

• - Warning systems, such as a parking aid 21 which generates warning signals when approaching an obstacle or a belt warning device 22 which generates a warning signal when the vehicle ignition is switched on or when the vehicle is detected and at the same time unbelted driver or vehicle occupant,
• - an audio system 23 which intended in broadcasting or on a data carrier existing entertainment or information content acoustically via arranged in the vehicle speakers 231 . 232 . 233 . 234 which, however, represent interference signals for voice control,
• - a vehicle navigation system 24 which reproduces acoustic guidance instructions about loudspeakers arranged in the vehicle in the event of impending maneuvers to follow a driving route, and
• - a handsfree device 25 , which outputs the speech signal of a conversation participant via vehicle speakers.

These devices or components 11 . 12 . 13 . 14 . 21 . 22 . 23 . 24 . 25 represent potential noise sources in the context of the present invention.

The sound pickup 51 generated audio signal 510 represents or that of the sound recorders 51 and 52 recorded audio signals 510 and 520 thus represent a composite signal from the user 3 or vehicle occupants spoken voice instructions or voice commands 31 as Audionutzsignal and additional unwanted signal components 29 , ie interference signal components, by the components mentioned 11 . 12 . 13 . 14 . 21 . 22 . 23 . 24 . 25 are generated.

Unless otherwise stated, hereinafter for the sake of simplicity of illustration, an audio signal will be used 510 a sound pickup 51 went out.

This audio signal 510 becomes a signal processing module 53 fed. The task of the signal processing module 53 is to process the audio signal 510 supplied thereto according to a processing manner with the aim of extracting from the supplied audio signal 510 an audio signal 530 , that is, contained within, by a user 3 of the system spoken voice commands 31 or to isolate voice instructions and make them available for further processing.

The further processing can in particular in a voice control of a vehicle-powered device 4 or system, for example, a driver information system or a vehicle multimedia system, which the isolated Audionutzsignal 530 or the speech commands contained therein 31 or convert voice prompts into control commands for the device or system. For example, the Audionutzsignal 530 be converted into control commands for selecting a device function, such as a navigation destination input. Next, the Audionutzsignal 530 Depending on the content, for example, for further operation within a selected function, such as the voice input of an address, which are subsequently implemented as a destination for a route calculation and route guidance from the current state to the destination.

The processing of the signal processing module 53 is from a control module 54 controlled. The control module leads to the control of the processing mode 54 the signal processing module control parameters 540 to. The signal processing module 53 is, for example, as in the case of the present embodiment, designed as a digital signal processor (DSP), which is adapted to, depending on the supplied control parameters 540 to perform certain filter functions.

To determine or determine the control parameters 540 uses the control module 54 Information about the operating states of the components 11 . 12 . 13 . 14 . 21 . 22 . 23 , 24, 25, which generate audible noises during operation in the vehicle interior. These operating states are from a detection module 55 captured and the corresponding operating state data to the control module 54 supplied or through the control module 54 from the acquisition module 55 accessed.

The acquisition module 55 can detect the operating conditions of the noise generating components 11 . 12 . 13 . 14 . 21 . 22 . 23 . 24 . 25 be directly connected to these components or their controls, such as switches. Next, the acquisition module 55 also operating status data via a data bus, such as the widespread in the automotive field CAN bus, over which the components connected to it are controlled retrieve. Furthermore, the detection module can also be connected to sensors which are set up to detect the operating state of components.

In the case of the exemplified windshield wiper 11 can its operating state, that is, whether switched off or, for example, in the interval mode, in continuous operation with low wiper speed or in continuous operation at high wiper speed, for example at a wiper switch, on the operating condition of the windshield wiper 11 is set to be tapped.

In the case of one over the vehicle bus 7 , here the CAN bus, controlled windscreen wiper 11 can provide information about the operating status of the windscreen wiper 11 also from the vehicle bus 7 be tapped.

For a windscreen wiper automatically controlled by a rain sensor 11 On the one hand, the signal of the rain sensor is evaluated by a windscreen wiper control in order to determine whether and, if appropriate, how intensively the windscreen is wetted by rain. In addition, the wiper control evaluates the switching signal of the wiper switch to determine whether the automatic control of the wiper 11 is activated. From this, the windscreen wiper control generates a drive signal for the electromotive drive of the windscreen wiper 11 which is via the vehicle bus 7 is transmitted. This signal contains information about the speed or, in intermittent operation, the frequency of the windscreen wiper deflection. This information may be from the state acquisition module 55 be removed from the vehicle bus.

In the case of the power window 12 can its operation and actuation direction, that is unconfirmed, close windows or open windows, for example, be detected via an associated control or be removed from the vehicle bus in control of the vehicle bus.

Analogously, this also applies to the other potential noise sources, such as interior ventilation and / or air conditioning 13 , Direction indicator 14 , electrical seat-adjustment 15 and other, not explicitly stated sources of noise.

In an analogous manner, the operating conditions of the other, already mentioned, in operation at least temporarily noise generating components such as warning signals generating warning systems, such as parking aid 21 , or Gurtwarner 22 , Audio or multimedia equipment 23 with acoustic reproduction of entertainment or informational content, vehicle navigation system 24 with acoustic output of route guidance notes or handsfree 25 with acoustic speech signal output from the acquisition module 55 be recorded.

The control module 54 is dependent on the detected operating conditions of the acquisition module 55 monitored components 11 . 12 . 13 . 14 . 21 . 22 . 23 . 24 . 25 the control parameters 540 ,

In a preferred embodiment, the control module engages 54 to a memory 56 to, in the possible operating conditions of the components 11 . 12 . 13 . 14 . 21 , 22, 23, 24, 25 each have control parameters 540 assigned. The control module 54 reads the assigned control parameters 540 from the memory as a function of the detected operating states 56 and forwards these to the signal processing module 53 to, its processing for the audio signal 510 according to the supplied control parameters 540 is controlled.

The processing of the signal processing module 53 will be described below using the example of the windscreen wiper 11 described as noise source.

As a control parameter 540 be the signal processing module 53 In this case, information is transmitted, which is a synthesizing of the operating noise of the windscreen wiper 11 enable in its current operating state. This information includes an interval duration at interval operation and an angular velocity or displacement speed of the wiper blade. In addition, this information includes a pattern, for example a frequency distribution of the noise generated by the wiper blade on the windshield, as far as this in the passenger compartment 1 is perceptible.

The signal processing module 53 is adapted from this information a simulation of that of the windshield wiper 11 generate wiping noise in its current interval mode. This simulated wiping noise will be in the signal processing module 53 from the recorded audio signal 510 subtracted. The result of this subtraction is the audio signal 530 ,

A further improvement is achieved by using the synthetically generated, simulated wiping noise with the actual noise of the windshield wiper 11 compared and adjusted accordingly. The purpose of this is an additional adaptive filter, via which the synthetic wiping noise simulation before the subtraction of the recorded audio signal 510 is directed. The filter coefficients of this adaptive filter can with a common algorithm, such as an LMS (= Least Mean Square) algorithm from the simulation of the wiper noise and the result of the difference formation, so the Audionutzsignal 530 , adapted and optimized. As a result, this leads to improved modeling of the synthesized wiper noise, which more accurately reflects the actual operating noise of the windshield wiper 11 corresponds, and thus also to a Audionutzsignal 530 , largely of noise components, here the operating noise of the windscreen wiper 11 , is free.

The adaptive filter is thus able to detect deviations of the modeled wiper noise from that with the microphone or microphones 51 and 52 absorbed actual wiper noise, which may result from a occupation or loading of the passenger compartment or a wear of the wiper blades in operation.

The adaptive filter transfer function thus obtained may also apply the original information for synthesizing the simulation of the windshield wiper operating noise to the associated control parameters 540 adapt. These customized control parameters 540 can in the store 56 be written so that at a next start-up of the wiper 11 improved or more accurate control parameters 540 ready to stand.

The processing of the signal processing module will be described below 53 the example of the audio system 23 described as noise source.

The audio system 23 gives entertainment or information content, hereinafter simplistic multimedia signals, in acoustic form over speakers again. These are, for example, on the left and right A-pillar, the driver and the front passenger door and the left and right C-pillar and the left and right rear door arranged, the main Schallabstrahlrichtung the speaker on the vehicle interior, so the cabin 1 is aligned.

The direction of incidence of the multimedia signals on the microphones 51 and 52 , which are in the area of the driver's head 3 in the headliner of the passenger compartment 1 are different from those of the driver 3 spoken speech commands 31 to operate the voice-operated device 4 ,

Now shows the acquisition module 55 on that the audio system 23 is in operation reads the control module 54 the assigned control parameter 540 from the store 56 and forwards these to the signal processing module 53 to control its processing.

In response to the supplied control parameters 540 the signal processing module processes those from the microphones 51 and 52 supplied audio signals 510 and 511 in the sense of a directional characteristic such that signal components, which from the direction of the head of the driver 3 that is, those that come from a different direction, some of the vehicle doors or the A and C pillars, are more weighted. A distinction of these signals is the stereoscopic recording with two microphones 51 and 52 and thus by comparing signal propagation times possible.

Furthermore, the signal processing module 530 by the control parameters preferably also controlled to that of the audio signal system 23 to obtain generated multimedia signal directly from the audio signal system and to subtract from the microphone signals or an audio signal generated from the two microphone signals. Preferably, before the subtraction, the multimedia signal may also pass through an adaptive filter with which the transmission paths from the loudspeakers to the microphones 51 and 52 , such as frequency-selective attenuation due to loading or reinforcements by reflections, be modeled. This measure also improves the isolation of the Audionutzsignals 530 from the audio signal recorded with the microphones.

The filter coefficients of the adaptive filter determined in this way can be used as part of the control parameters 540 In the storage room 56 be filed. When the audio system is restarted, the adaptive filter or, more generally, the signal processing of the signal processing module 530 be initialized with these filter coefficients, so that an adaptation can be accelerated.

Also, a combination of noise from a plurality of the aforementioned potential noise sources can be processed by the inventive approach. This is facilitated in particular by the state detection or the control of the processing mode of the audio signal processing as a function of the detected operating states of the possible sources of interference, or even made possible in the first place. Thus, as non-operational sources of interference in the signal processing can be disregarded from the outset, which compared to purely adaptive methods, which are based solely on the recorded audio signals, significantly better results both in terms of precision and the speed of interference suppression.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 0411360 A1 [0004]

Claims (10)

Noise suppression device comprising a detection module (55) configured to detect an operating state of at least one component (11, 12, 13, 14, 21, 22, 23, 24, 25) generating noise in at least one operating state, a control module (54) configured to generate control parameters (540) as a function of the detected operating state, a signal processing module (53) adapted to process a supplied audio signal (510) in accordance with a processing mode controllable via the control parameters (540) and to output an audio user signal (530) obtained by the processing. Noise suppression device according to Claim 1 and further comprising a memory (56) in which the operating states of the at least one component (11, 12, 13, 14, 21, 22, 23, 24, 25) are assigned control parameters (540) and wherein the control module (54) continues designed to read out in dependence on a detected operating state, the control parameter (540) assigned in the memory (56) and to supply it to the signal processing module (53). Störunterdrückungseinrichtung after Claim 2 wherein the signal processing module (53) is further adapted to isolate a noise component from the audio signal (510) and to output noise parameters characterizing this noise component, and wherein the control module (54) is adapted to control the control parameters (540) stored in the memory (56). depending on the noise parameters. Noise suppression device according to one of the preceding claims, wherein the signal processing module (53) comprises a filter module for processing the audio signal (510) and wherein the control parameters (540) comprise filter parameters for the filter module. Noise suppression device according to Claim 4 wherein the filter module is implemented in the form of an adaptive filter, wherein the control module (54) is adapted to initialize the filter with the filter parameters and wherein the filter module is adapted to begin processing the audio signal (510) with the pre-assigned filter parameters and automatically adapt the filter parameters in the further course of the processing of the audio signal (510). A noise suppression device according to any one of the preceding claims, wherein the signal processing module (53) is adapted to synthesize from the control parameters (540) a noise simulation for the noise portion of the audio signal (510) and the noise simulation to isolate the audio useful signal (530) from the audio signal (510 ) to extract. Voice control device, set up for controlling a device (4) by means of spoken voice commands (31) at least one sound pickup (51, 52) for recording at least one spoken voice command (31), a noise reduction device according to any one of the preceding claims, to which the audio signal (510) containing the recorded voice command (31) is supplied, and a command conversion module for converting the audio user signal (530) into control signals for controlling the device (4). Method for noise suppression in an audio signal, wherein an operating state of at least one component (11, 12, 13, 14, 21, 22, 23, 24, 25) which generates noises in at least one operating state is detected, wherein control parameters (540) are generated as a function of the detected operating state, wherein the audio signal (510) is processed into an audio payload signal (530) according to a manner of processing, the processing being controlled by the control parameters (540), and wherein the audio payload signal (530) is output. A computer program comprising instructions for causing the noise canceling means to follow one of Claims 1 to 6 or the voice control device Claim 7 the process steps after Claim 8 performs. Computer-readable medium on which the computer program is based Claim 9 is stored.

Images