Virtual position noise signal estimation for engine order cancellation

Abstract

A system and method for accurately estimating engine noise at a virtual microphone location, such as an ear location of an occupant, in an acoustic space in order to enhance the performance of an engine order cancellation (EOC) system is provided. A set of weights and transfer functions dependent on various vehicle parameters, such as frequency, load, and speed, can be employed to estimate the noise at a location where no physical microphone exists. Accurate estimation of engine noise at a virtual location, such as an ear location of an occupant, can be achieved using a frequency dependent weighted sum of filtered and unfiltered error signals measured at microphones mounted at different locations, such as a vehicle cabin, within the acoustic space that can not be located near the virtual location.

Description

Technical Field

[0001] The present invention relates to engine order elimination, and more specifically, to estimating noise at virtual microphone locations (such as locations near the ears of occupants in a vehicle). Background Technology

[0002] Active noise control (ANC) systems use feedforward and feedback structures to attenuate unwanted noise, adaptively removing unwanted noise within the listening environment, such as inside a vehicle cabin. ANC systems typically eliminate or reduce unwanted noise by generating canceling sound waves to disruptively interfere with the unwanted audible noise. When noise and "anti-noise" are combined to reduce the sound pressure level (SPL) at a location, destructive interference occurs; the "anti-noise" is substantially the same amplitude but out of phase with the noise. In a vehicle cabin listening environment, potentially unwanted noise sources include the engine, the interaction between the vehicle's tires and the road surface, and / or sound radiated by vibrations from other parts of the vehicle. Therefore, unwanted noise varies with vehicle speed, road conditions, and operating status.

[0003] Engine Order Cancellation (EOC) systems are specific ANC systems implemented in vehicles to reduce unwanted in-vehicle noise levels caused by narrowband acoustic and vibration emissions from the vehicle's engine and exhaust system or other rotating drivetrain components. EOC systems generate feedforward noise signals based on the engine or other rotating shaft angular velocities (such as revolutions per minute (RPM)) and use these signals and adaptive filters to reduce in-cabin SPL by radiating noise immunity via loudspeakers.

[0004] EOC systems are typically least mean square (LMS) adaptive feedforward systems, which continuously adjust adaptive filters based on RPM inputs from sensors mounted to the drive shaft and error signals from microphones located at various positions within the cabin.

[0005] An adaptive algorithm generates an anti-noise signal to eliminate noise at the location of the error microphone, rather than at the occupant's ear. The location of the error microphone affects the performance of the EOC system. Traditional noise cancellation algorithms rely on the assumption that cancellation at the error microphone location is closely related to the location of the nearest occupant. This relationship is frequency-dependent; the correlation between the physical microphone signal and the occupant's ear decreases as the noise frequency increases. This reduction in correlation is particularly impactful when the number of available error microphones and their placement are not optimal due to other vehicle requirements and / or limitations. Summary of the Invention

[0006] In one or more exemplary embodiments, a method is provided for estimating noise at a virtual microphone location for an engine order cancellation (EOC) system. The method may include receiving a plurality of estimated noise signals indicating noise at the location of each of a plurality of error microphones. Each estimated noise signal may be based at least in part on error signals from each of the plurality of error microphones. The method may further include filtering each estimated noise signal using a virtual path filter modeled according to a transfer function between each corresponding error microphone location and a virtual microphone location to generate a plurality of filtered estimated noise signals. The method may further include adaptively weighting each filtered estimated noise signal using weights that vary based on current vehicle conditions to generate a plurality of weighted filtered estimated noise signals. The method may further include generating an estimated virtual microphone noise signal indicating noise at the virtual microphone location based on the superposition of at least a plurality of weighted filtered estimated noise signals.

[0007] Each implementation may include one or more of the following features. Each virtual path filter may be a finite impulse response filter. Furthermore, each weight may be selected from multiple weights stored in a lookup table and derived based on the current vehicle conditions to be applied to each filtered estimated noise signal. Additionally, the virtual microphone position may correspond to the occupant's ear area.

[0008] The method may further include receiving the estimated virtual microphone noise signal from the adaptive filter controller and adjusting the adaptive transfer characteristics of the adaptive filter in part based on the estimated virtual microphone noise signal. The current vehicle condition may include the frequency of engine order noise, such that each weight varies at least based on that frequency. The current vehicle condition may also include at least one of engine load and vehicle speed, such that each weight is also selected based on at least one of engine load and vehicle speed.

[0009] Each of the multiple estimated noise signals can be divided into two signal paths, including a first signal path and a second signal path. The output of the second signal path may include multiple weighted and filtered estimated noise signals. To this end, the method may further include adaptively weighting each estimated noise signal in the first signal path using weights selected based on the current vehicle condition to generate multiple weighted estimated noise signals. Furthermore, generating a virtual microphone noise signal indicating an estimate of noise at a virtual microphone location based on the superposition of at least multiple weighted and filtered estimated noise signals may include generating the virtual microphone noise signal indicating an estimate of noise at a virtual microphone location based on the superposition of multiple weighted estimated noise signals from the first signal path and multiple weighted and filtered estimated noise signals from the second signal path.

[0010] One or more additional embodiments may relate to an EOC system comprising at least one adaptive filter, an adaptive filter controller, and a virtual position noise estimator. The adaptive filter may be configured to generate an anti-noise signal based on adaptive transfer characteristics and a reference signal received from a reference signal generator. The adaptive transfer characteristics of at least one adaptive filter may be characterized by a set of filter coefficients. The adaptive filter controller, including a processor and memory, may be programmed to adjust the set of filter coefficients based on the reference signal and an estimated virtual microphone noise signal indicating the noise at the virtual microphone location. The virtual position noise estimator may communicate with at least the adaptive filter controller.

[0011] A virtual position noise estimator may include a processor and a memory programmed to receive multiple estimated noise signals indicating the location of noise at each of a plurality of error microphones. Each estimated noise signal may be at least partially based on error signals from each of the plurality of error microphones. The virtual position noise estimator may also be programmed to filter each estimated noise signal using a virtual path filter modeled according to a transfer function between the location of each corresponding error microphone and a virtual microphone location to generate multiple filtered estimated noise signals. The virtual position noise estimator may also be programmed to adaptively weight each filtered estimated noise signal using weights selected and varying based on the current vehicle condition to generate multiple weighted filtered estimated noise signals. The virtual position noise estimator may also be programmed to generate an estimated virtual microphone noise signal indicating the noise at the virtual microphone location based on the superposition of at least a plurality of weighted filtered estimated noise signals.

[0012] Each implementation may include one or more of the following features. Each weight may be selected from multiple weights derived based on the current vehicle condition to be applied to each filtered estimated noise signal. The current vehicle condition may include the frequency of engine order noise, and each weight varies at least based on this frequency. Each virtual path filter may be a finite impulse response filter. Each of the multiple estimated noise signals may be divided into two signal paths, including a first signal path and a second signal path, wherein the output of the second signal path includes multiple weighted filtered estimated noise signals. In this respect, the virtual position noise estimator may also be programmed to adaptively weight each estimated noise signal in the first signal path using weights selected based on the current vehicle condition to generate multiple weighted estimated noise signals. Furthermore, the virtual position noise estimator may be programmed to generate a virtual microphone noise signal indicating an estimate of the noise at a virtual microphone position based on the superposition of multiple weighted estimated noise signals from the first signal path and multiple weighted filtered estimated noise signals from the second signal path. The virtual microphone position may be a fixed point in space corresponding to the occupant's ear and at a certain distance from multiple error microphones.

[0013] One or more additional embodiments may relate to a method for estimating noise at virtual microphone locations for an EOC system. The method may include receiving a plurality of estimated noise signals indicating noise at the location of each of a plurality of error microphones, wherein each estimated noise signal is at least partially based on error signals from each of the plurality of error microphones. The method may further include transmitting each estimated noise signal along a first signal path and a second signal path. The method may further include applying weights to each estimated noise signal in the first signal path, each weight being individually selected and varied based on current vehicle conditions to generate a plurality of weighted estimated noise signals. The method may further include filtering each estimated noise signal in the second signal path using a virtual path filter modeled according to a transfer function between each corresponding error microphone location and a virtual microphone location to generate a plurality of filtered estimated noise signals. The method may further include applying weights to each filtered estimated noise signal in the second signal path, each weight being individually selected and varied based on current vehicle conditions to generate a plurality of weighted filtered estimated noise signals. The method may further include generating a virtual microphone noise signal that indicates the noise at the virtual microphone location by superimposing a plurality of weighted estimated noise signals from a first signal path and a plurality of weighted filtered estimated noise signals from a second signal path.

[0014] Each implementation may include one or more of the following features. The proportion of each signal applied in the superposition may be adaptively controlled by its corresponding weight based on the current vehicle condition. The current vehicle condition may include the frequency of engine order noise, such that each weight is selected and varied individually based at least on that frequency. The virtual microphone position may correspond to the occupant's ear. Attached Figure Description

[0015] Figure 1 This is an environmental block diagram of a vehicle having an engine order elimination (EOC) system according to one or more embodiments of this disclosure;

[0016] Figure 2 It is based on one or more embodiments of this disclosure. Figure 1 Detailed view of the reference signal generator depicted in the figure;

[0017] Figure 3 This is a schematic block diagram illustrating an EOC system including a virtual location noise estimator according to one or more embodiments of the present disclosure;

[0018] Figure 4 It is based on one or more embodiments of this disclosure. Figure 3 An extended block diagram of the virtual position noise estimator depicted in the figure; and

[0019] Figure 5 This is an exemplary flowchart depicting a method for estimating noise at a location near an occupant's ear according to one or more embodiments of the present disclosure. Detailed Implementation

[0020] Detailed embodiments of the invention are disclosed herein as needed; however, it should be understood that the disclosed embodiments are merely illustrative of the invention, which may be embodied in different and alternative forms. The drawings are not necessarily drawn to scale; some features may be enlarged or minimized to show details of specific components. Therefore, the specific structural and functional details disclosed herein should not be construed as limiting, but rather serve as a representative basis for teaching those skilled in the art to apply the invention in different ways.

[0021] Figure 1This is an environmental diagram illustrating an Engine Order Cancellation (EOC) system 100 for a vehicle 102 with a reference signal generator 108. The reference signal generator 108 generates a reference signal x[n] corresponding to audible engine order noise for each engine order originating from the vehicle's engine and exhaust system 110. The EOC system 100 can be integrated with a feedforward and feedback active noise control (ANC) framework or system 104, which generates anti-noise by adaptively filtering the reference signal x[n] from the reference signal generator 108 using one or more microphones 112. The anti-noise signal y[n] can then be played through one or more speakers 124. S(z) represents the transfer function between a single speaker 124 and a single microphone 112. Although Figure 1 A single reference signal generator 108, microphone 112, and speaker 124 are shown for simplicity only; however, it should be noted that in addition to multiple speakers 124 (e.g., 4 to 8) and multiple microphones 112 (e.g., up to 8), a typical EOC system may also include multiple engine order noise reference signal generators 108.

[0022] refer to Figure 2 The reference signal generator 108 may include an RPM sensor 242, which can provide an RPM signal 244 (e.g., a square wave signal) indicating the rotation of the engine drive shaft or other rotating shaft indicating engine speed. In some embodiments, the RPM signal 244 may be obtained from a vehicle network bus (not shown), such as a controller area network (CAN) bus. Since the radiated engine order is proportional to the RPM of the drive shaft, the RPM signal 244 represents the frequency generated by the transmission system, which includes the engine and exhaust system. Therefore, the signal from the RPM sensor 242 can be used to generate a reference engine order signal corresponding to each of the engine orders of the vehicle. Thus, the RPM signal 244 can be used in conjunction with a lookup table 246 of engine order frequency versus RPM.

[0023] More specifically, lookup table 246 can be used to convert the RPM signal 244 into one or more engine order frequencies. The frequency of a given engine order at the sensed RPM, as retrieved from lookup table 246, can be supplied to frequency generator 248 to generate a sine wave at the given frequency. This sine wave represents a reference signal x[n] indicating engine order noise for a given engine order. Frequency generator 248 can be an oscillator, such as a quadrature oscillator, or any similar device for generating a sinusoidal reference signal indicating engine order noise. Since multiple engine orders may exist, EOC system 100 may include multiple reference signal generators 108 and / or frequency generators 248 for generating a reference signal x[n] for each engine order based on the RPM signal 244.

[0024] An engine rotating at 1800 RPM can be said to operate at 30 Hz (1800 / 60 = 30), which corresponds to the fundamental or primary engine order frequency. For a four-cylinder engine, two cylinders are ignited during each crankshaft rotation, producing a primary frequency of 60 Hz (30 x 2 = 60), which defines the sound of a four-cylinder engine at 1800 RPM. In a four-cylinder engine, this frequency is also referred to as the "second engine order" because it is twice the engine speed. At 1800 RPM, the other major engine orders of a four-cylinder engine are the fourth order at 120 Hz and the sixth order at 180 Hz. In a six-cylinder engine, the ignition frequency results in a dominant third engine order; in a V-10, the fifth engine order is dominant. As RPM increases, the ignition frequency increases proportionally. As previously described, the EOC system 100 may include multiple reference signal generators 108 and / or frequency generators 248 for generating a reference signal x[n] for each engine order based on the RPM signal 244. Furthermore, the ANC framework 104 within the EOC system 100 (e.g., adaptive filter 118, adaptive filter controller 120, secondary path filter 122) may be scaled to reduce or eliminate each of these multiple engine orders. For example, an EOC system reducing the 2nd, 4th, and 6th engine orders requires three ANC frameworks or subsystems 104, one for each engine order. Certain system components, such as the error microphone 112 and the noise-canceling speaker 124, may be common to all systems or subsystems.

[0025] Return to reference Figure 1 The characteristic frequencies of noise and vibration originating from the engine and exhaust system 110 can be sensed by one or more of the RPM sensors 242 optionally included in the reference signal generator 108. The reference signal generator 108 can output a reference signal x[n], which is a signal representing a specific engine order frequency. As previously mentioned, the reference signal x[n] may be at different engine orders of interest. Furthermore, these reference signals can be used individually or in various combinations known to those skilled in the art. The reference signal x[n] can be filtered S′(z) by the secondary path filter 122 using the modeled transfer characteristics of the estimated secondary path (i.e., the transfer function between the noise-canceling loudspeaker 124 and the error microphone 112).

[0026] Drivetrain noise (e.g., engine, drive shaft, or exhaust noise) is mechanically and / or acoustically transmitted into the passenger compartment and received by one or more microphones 112 within vehicle 102. One or more microphones 112 may be located, for example, in… Figure 1The headrest 114 of the seat 116 shown is located therein. Alternatively, one or more microphones 112 may be located in the headliner 115 of the vehicle 102, or in some other suitable location, to sense the acoustic noise field heard by the occupants within the vehicle 102. Engine, drive shaft, and / or exhaust noise is transmitted to the microphone 112 according to the transmission characteristics P(z) representing the main path (i.e., the transfer function between the actual noise source and the error microphone).

[0027] Microphone 112 can output an error signal e[n] representing the noise detected by microphone 112 present in the passenger compartment of vehicle 102. In EOC system 100, the adaptive transfer characteristic W(z) of adaptive filter 118 can be controlled by adaptive filter controller 120. Adaptive filter controller 120 can operate according to a known least mean square (LMS) algorithm based on the error signal e[n] and a reference signal x[n], which can optionally be filtered by filter 122 using a modeled transfer characteristic S′(z). Adaptive filter 118 is commonly referred to as a W-type filter. LMS adaptive filter controller 120 can update the filter coefficients of transfer characteristic W(z) based on the error signal e[n]. The adaptive or updated W(z) process that improves noise cancellation is called convergence. Convergence refers to the creation of an adaptive filter that minimizes the error signal e[n], which is controlled by a step size that controls the adaptive rate of a given input signal. The step size is a scaling factor that indicates the speed at which the algorithm converges to minimize e[n] by limiting the magnitude change of the adaptive filter coefficients based on each update of the adaptive filter 118.

[0028] The noise immunity signal y[n] can be generated by an adaptive filter 118 and an adaptive filter controller 120 based on an adaptive filter formed by a combination of the identified transfer characteristic W(z) and a reference signal or reference signal x[n]. The noise immunity signal y[n] ideally has a waveform such that when played through speaker 124, it generates noise immunity near the occupant's ears and microphone 112, which is substantially out of phase and has the same amplitude as the engine order noise audible to the occupants of the compartment. The noise immunity from speaker 124 can be combined with the engine order noise in the compartment near microphone 112, resulting in a reduction in the sound pressure level (SPL) caused by the engine order noise at that location. In some embodiments, the EOC system 100 can receive sensor signals from other acoustic sensors in the passenger compartment, such as acoustic energy sensors, acoustic intensity sensors, or acoustic particle velocity or acceleration sensors, to generate an error signal e[n].

[0029] Vehicles typically have other shafts that rotate at a different rate relative to the engine's RPM. For example, a drive shaft rotates at a rate associated with the engine, depending on the current gear ratio set by the transmission. A drive shaft may not have perfect rotational balance because it may have a degree of eccentricity. When rotating, this eccentricity causes rotational imbalance, thus exerting oscillating forces on the vehicle, and these vibrations may produce audible sounds in the passenger compartment. Other rotating shafts that rotate at a rate different from the engine include half-shafts or axes that rotate at a rate set by the gear ratio in their differential. In some embodiments, the reference signal generator 108 may have RPM sensors on different rotating shafts, such as drive shafts or half-shafts.

[0030] When vehicle 102 is in operation, processor 128 may collect and optionally process data from RPM sensor 242 and microphone 112 in reference signal generator 108 to construct a database or map containing data and / or parameters to be used by vehicle 102. The collected data may be stored locally in storage device 130 or in the cloud for future use by vehicle 102. Examples of data types related to EOC system 100 that may be used for local storage in storage device 130 include, but are not limited to, RPM history, microphone spectrum or time-related signals, microphone-based acoustic performance data, EOC tuning parameters, and primary engine order based on drive mode. Additionally, processor 128 may analyze RPM sensor and microphone data and extract key features to determine a set of parameters to be applied to EOC system 100. This set of parameters may be selected when triggered by an event. In one or more embodiments, processor 128 and storage device 130 may be integrated with one or more EOC system controllers, such as adaptive filter controller 120.

[0031] Figure 1 The simplified EOC system diagram depicted shows a secondary path, denoted by S(z), between each loudspeaker 124 and each microphone 112. As previously mentioned, an EOC system typically has multiple loudspeakers, microphones, and a reference signal generator. Therefore, a 6-loudspeaker, 6-microphone EOC system would have a total of 36 secondary paths (i.e., 6×6). Correspondingly, the 6-loudspeaker, 6-microphone EOC system could also have 36 S′(z) filters (i.e., secondary path filters 122) that estimate the transfer function of each secondary path. Figure 1 As shown, the EOC system also has a W(z) filter (i.e., adaptive filter 118) between each reference signal x[n] from the reference signal generator 108 and each loudspeaker 124. Therefore, a 5-reference-signal-generator, 6-loudspeaker EOC system can have 30 W(z) filters. Alternatively, a 6-frequency-generator, 6-loudspeaker EOC system can have 36 W(z) filters.

[0032] As previously discussed, narrowband engine noise cancellation systems can use multiple error microphones mounted in the vehicle's headliner or other locations to provide feedback to an adaptive algorithm. In conventional systems, the algorithm generates noise immunity to eliminate noise at the error microphone locations rather than at the occupant's ear area. Due to certain vehicle manufacturing and design requirements or limitations, the number of error microphones available and their placement may not be optimal. Vehicle manufacturers place microphones in the cabin for multiple functions, each with different requirements for optimal placement. For example, the optimal placement of noise cancellation microphones may not be consistent with the optimal placement of hands-free voice communication microphones (such as those used for telephone calls). Noise cancellation microphones tend to be located within 1 / 10 of the wavelength at the occupant's position and are therefore often placed directly above the occupant's head or in the headrest. Design constraints, such as the presence of a sunroof or glass sunroof, often hinder optimal EOC error microphone placement. On the other hand, hands-free telephone voice call microphones are placed to optimally detect the speaker's voice. These microphones are typically placed directly in front of the occupant's head, usually in fixed locations within the vehicle interior (such as rearview mirrors or dashboards).

[0033] The location of the noise cancellation (error) microphone affects EOC performance. Traditional noise cancellation algorithms rely on the assumption that the cancellation at the error microphone location is closely related to the location of the nearest occupant. However, this relationship is frequency-dependent. As the distance between the noise frequency and location increases, the correlation between the noise signal at the ear location and the error microphone decreases. Therefore, when the number of available error microphones and their placement may not be optimal, accurately estimating the noise at the occupant's ear location helps ensure that EOC performance is not compromised.

[0034] Existing techniques for estimating noise at the occupant's ear location are applicable to simple acoustic spaces with fixed-frequency noise (e.g., laboratory conditions), but may be unreliable when applied to automotive applications with dynamic frequency noise and acoustics. One or more embodiments of the present invention relate to systems and methods for more accurately predicting non-steady-state narrowband engine order noise at the ear location in complex acoustic environments, such as a vehicle cabin. The techniques described in this invention allow for a more reliable estimation of noise at the occupant's ear location (referred to as a virtual microphone location) based on signals from physically incorrect microphone locations. These techniques can employ a set of weights and transfer functions that depend on various vehicle parameters or conditions such as frequency, load, and speed to estimate noise at the occupant's ear location where a microphone is not present.

[0035] The systems and methods disclosed herein model the acoustics of highly reverberant environments such as vehicle cabins. As previously mentioned, engine noise typically originates from the vehicle's intake and exhaust. Sound waves propagate through the cabin along a direct path and may have numerous reflections before being summed at the occupant's ear location. To optimize cancellation performance, the signal at a physical error microphone can then be used with the acoustic model to estimate the signal at a virtual microphone location (such as an approximate location of the occupant's ear location) located away from the physical error microphone. This relationship can be modeled as the path from the error microphone (physical location (P)) to the occupant's ear location (virtual location (V)).

[0036] Figure 3 This is a schematic block diagram illustrating an EOC system 300 according to one or more embodiments of the present disclosure. As will be understood by those skilled in the art, the EOC system 300 may be a modified Filtered Least Mean Square (MFxLMS) EOC system. The EOC system 300 may correspondingly employ the MFxLMS adaptive algorithm for narrowband engine noise cancellation. However, other types of LMS-based EOC systems, such as Filtered Least Mean Square (FxLMS) systems, may be used to employ one or more aspects of the present invention.

[0037] EOC system 300 may include and combine Figure 1 The environmental diagram illustrates and describes components similar to those of the EOC system 100. For example, the EOC system 300 may include a reference signal generator 308, which includes at least an RPM sensor 342 and a frequency generator 348 (depicted as an oscillator) for generating a sinusoidal engine order noise reference signal x[n] having frequency characteristics derived from the noise and vibration of the engine and exhaust system 310. Figure 1 Similarly, for ease of illustration, the EOC system 300 is shown as having a reference signal generator 308, an error microphone 312, and a speaker 324. In applications, the EOC system 300 can be a scalable multiple-input multiple-output (MIMO) system operating for multiple engine orders, multiple speaker outputs, and multiple error microphones. The EOC system 300 can also be scaled to estimate noise signals at multiple virtual microphone locations (e.g., occupant ear locations), as described in more detail below.

[0038] exist Figure 3 In the schematic block diagram, the error microphone 312 is depicted as an adder (or summing operator / element). Furthermore, the actual primary path P(z) and the actual secondary path S are represented by boxes using elements 350 and 352, respectively. p The transfer function of (z). For illustrative purposes, Figure 3The component division between acoustic domain 354 and electrical domain 356 is also described.

[0039] The reference signal x[n] can be filtered by the first-stage path filter 358. The first-stage path filter 358 can use the modeled transfer characteristics of x[n]. The reference signal is filtered to generate a filtered reference signal, and this transfer characteristic estimates the secondary path (i.e., transfer function) x′[n) between the noise-canceling loudspeaker 324 and the virtual microphone position representing the occupant's ear. Similar to EOC system 100, EOC system 300 may include a first adaptive filter 318 and an adaptive filter controller 320. The adaptive transfer characteristic W(z) of the first adaptive filter 318 may be controlled by the adaptive filter controller 320, partially based on the filtered reference signal, according to an LMS-based adaptive algorithm, using x′[n]. The adaptive filter controller 320 may actively update the filter coefficients of the first adaptive filter 318 to improve noise cancellation. The filter coefficients of the first adaptive filter 318 may be referred to as active filter coefficients. EOC system 300 may include a second adaptive filter 360, which also has an adaptive transfer characteristic W(z) characterized by a set of filter coefficients. The second adaptive filter 360 may be a copy of the first adaptive filter 318. Therefore, the filter coefficients of the second adaptive filter 360 may be referred to as passive filter coefficients.

[0040] As shown in the figure, the reference signal x[n] can also be filtered by the second adaptive filter 360 to generate the actual noise-resistant signal y[n]. The second adaptive filter 360, controlled by the adaptive filter controller 320 through a set of filter coefficients, can generate the noise-resistant signal y[n] according to an adaptive algorithm. In the acoustic domain, the noise-resistant signal y[n] can be converted into sound by the loudspeaker 324. As mentioned earlier, the actual secondary path S represented by box 352 p (z) represents the transfer function between the loudspeaker 324 and the error microphone 312. The signal y′[n] represents the actual secondary path S at the physical error microphone 312. p (z) The audible noise immunity filtered is called the actual error microphone noise immunity y′[n]. The actual error microphone noise immunity y′[n] can be combined with the main noise d[n] from the engine and exhaust system 310, as filtered by the actual main path P(z) (represented by box 350) at the error microphone 312. The error microphone 312 can output an error signal e[n] indicating the residual engine noise present in the compartment (i.e., the noise not eliminated by the noise immunity).

[0041] In the electrical domain, the noise-resistant signal y[n] can be filtered by the second-stage path filter 362 to generate an estimated noise-resistant signal. The second-stage path filter 362 can use the modeled transfer characteristics of y[n]. By filtering the noise-resistant signal, the modeled transfer characteristics estimate the secondary path (i.e., transfer function) between the noise-resistant speaker 324 and the error microphone 312. This estimate of the noise-resistant signal... This indicates the estimated noise immunity at the physical location of the error microphone 312. As shown, the estimated noise immunity signal can be subtracted from the error signal e[n] at adder 364. To generate an estimated noise signal at error microphone 312, or simply to generate an estimated noise signal. Estimating noise signals It can provide an estimate of engine noise at the physical error microphone location.

[0042] To estimate engine noise at a virtual microphone location (e.g., the area around an occupant's ear), the EOC system 300 may also include a virtual location noise estimator 366. The virtual location noise estimator 366 may include an acoustic model to estimate the noise signal based on signals from the physical microphone location (i.e.,...). The virtual position noise estimator 366 provides an improved estimate of noise at a virtual microphone location, such as the location representing an occupant's ear. Therefore, the virtual position noise estimator 366 models the path from one or more physical locations of the error microphone 312 to the virtual location representing the occupant's ear. Thus, the virtual position noise estimator 366 can output an estimated virtual microphone noise signal. The estimated virtual microphone noise signal provides an estimate of the engine noise at a virtual microphone location (e.g., the occupant's ear location). As previously discussed, in complex dynamic acoustic environments (such as a vehicle cabin), it is difficult to accurately predict non-steady-state narrowband noise at virtual locations (such as the occupant's ear location). According to one or more embodiments of this disclosure, the virtual location noise estimator 366 may employ a set of weights and transfer functions that depend on various vehicle parameters or conditions such as frequency, load, and speed to estimate the engine noise at the occupant's ear location (i.e., the virtual microphone location). Therefore, as shown, the RPM signal or a frequency derived from the RPM signal at the reference signal generator 308 may be provided to the virtual location noise estimator 366.

[0043] Figure 4 This is an extended block diagram of a virtual position noise estimator 366 according to one or more embodiments of the present disclosure. A relatively large frequency range can be divided into multiple frequency compartments, each of which constitutes a relatively narrow frequency range. Figure 4The example implementation of the virtual position noise estimator 366 shown describes the estimation of noise at a single virtual microphone location (e.g., the occupant's ear location) for one engine order and one frequency bin, using noise signals from four physical error microphones. However, this concept can support any number of physical and virtual microphones (e.g., ear locations) and can be further scaled to take into account multiple engine orders and frequency bins.

[0044] As shown in the figure Figure 4 The virtual position noise estimator 366 can be based on estimated noise signals measured at four physical error microphones P0, P1, P2, and P3. To provide an estimated virtual microphone noise signal at the first virtual microphone position V0 The virtual position noise estimator 366 may include a W-value applied to the signal between each physical microphone and each virtual microphone. pv The first set of weights is 410. This represents the estimated noise signal measured at four physical error microphones (P0, P1, P2, P3). It can be divided into two signal paths. The estimated noise signal in the first signal path 412 It can be used for the first set of weights 410 (e.g., W). 00 W 10 W 20 W 30 The direct input of the signal is used to generate a weighted estimated noise signal.

[0045] The virtual position noise estimator 366 may also include H pv [p][v] represents a set of virtual path filters 414, which typically represent the transfer function from each physical error microphone (P) to each virtual microphone (V) representing the occupant's ear location. The virtual path filters 414 can be finite impulse response (FIR) filters and can be designed using a deconvolution process. For example, during the design and calibration process for a specific vehicle environment, signals from the actual microphone at the error microphone location and at a location approximating the occupant's ear location can be measured first, and then a deconvolution process can be used to determine the transfer function H between each physical error microphone location and each ear location (virtual microphone location). pv [p][v]. Transfer function H pv [p][v] can ultimately be converted into a digital FIR filter, and then used as a virtual path filter 414 in the virtual position noise estimator 366 in the EOC system 300 and the corresponding adaptive algorithm. This is the estimated noise signal in the second signal path 416. The virtual path filter 414 can be used to filter and generate the filtered estimated noise signal for each of the four physical error microphones (P0, P1, P2, P3).

[0046] Each virtual microphone position can be a predetermined fixed point in space relative to a physical microphone or sensor (such as error microphone 312). The virtual microphone position may correspond to the location of an occupant's ear. In some embodiments, the virtual microphone position may represent the ear location of a typical occupant. In some other embodiments, each virtual microphone position can be adjusted to accommodate different types of occupants and / or different vehicle configurations. For example, occupant settings can be used to customize the virtual microphone positions to better estimate the occupant's ear location based on occupant characteristics (such as torso height) and seat position settings. As another example, in vehicles with more flexible cabin configurations, such as those configured for autonomous driving, one or more virtual microphone positions can be adjusted based on whether the vehicle is in autonomous driving mode.

[0047] According to one or more embodiments, the virtual microphone position can be determined from a set of virtual microphone positions selected based on occupant head tracking. For example, the position of the occupant's head can be sensed, and the virtual microphone position can be adjusted based on the current head position. In embodiments with adjustable virtual microphone positions, the corresponding virtual path filter 414 can be similarly adapted or modified to represent an appropriate transfer function from each error microphone position to each adjustable virtual microphone position. As an example, the virtual path filter 414 can be designed, calibrated, and stored for each error microphone 312 and multiple virtual microphone positions associated with each occupant. The appropriate virtual path filter 414 can then be selected based on the current virtual microphone position, whether selected by the occupant through system settings or sensed via head position tracking.

[0048] The virtual position noise estimator 366 may also include a signal estimator derived from R that can be applied to the signal between each filtered physical microphone and each virtual microphone. pv The second set of weights is represented by 418. This is the filtered estimated noise signal. It can be used as a weight for the second set of 418 (e.g., R). 00 ,R 10 ,R 20 ,R 30 The direct input is used to generate a weighted and filtered estimated noise signal.

[0049] According to one or more embodiments, the final estimated virtual microphone noise signal at the first virtual microphone position V0 can be generated at adder 420 from the superposition of all signals (i.e., filtered and unfiltered noise signals) in the two signal paths. The proportion of each signal applied in the superposition can be determined by the weight W. pv and R pv Control. The EOC system 300 can adaptively adjust these weights based on vehicle parameters or conditions, such as frequency. For example, the acoustic response of the passenger compartment to engine noise excitation varies with frequency. Additionally, the EOC system and adaptive algorithms handle tonal noise, which, unlike broadband noise, has signal statistics that change rapidly with frequency. Therefore, the weights 410 and 418 applied by the virtual position noise estimator 366 can also be varied relative to frequency to account for changes in the amplitude and phase of the added noise signals.

[0050] Selectable frequency-dependent weights can be derived during system design and calibration. For each combination of error microphone and virtual microphone locations, the signals measured at the error microphone and virtual microphone locations (e.g., the occupant's ear area) can be divided into entries in a frequency bin table with a predetermined range. For example, each frequency bin could be 3 Hz wide. An adaptive algorithm such as LMS can then be used to calculate the weight W for each frequency bin. pv and R pv Both. Using this method, a set of frequency-related weights can be derived, which are optimized until the residuals are minimized to below a defined threshold. Therefore, the virtual position noise estimator 366 may also include weights W for each combination of error microphone and virtual microphone positions. pv and R pv The lookup table 422 contains entries for each weight, divided by frequency bins. Lookup table 422 receives the frequency values ​​424 of the engine order noise and outputs the appropriate weights W. pv and R pv This is applied adaptively based on frequency to each estimated noise signal, both unfiltered and filtered. Figure 3 As shown, the frequency of the engine order noise can be derived from the reference signal generator 308. During EOC operation, when the engine noise frequency changes, weights can be selected from the appropriate frequency bin according to lookup table 422 and applied by the virtual position noise estimator 366. In one or more alternative implementations, the adaptive weights can also be functions of engine load, vehicle speed, and other vehicle parameters or conditions, rather than functions of frequency or functions other than frequency.

[0051] By mixing filtered and unfiltered error microphone signals and adaptively weighting these signals based on current vehicle parameters (e.g., frequency, load, and / or speed), an accurate and reliable estimate of non-steady-state narrowband noise can be provided at a virtual microphone location (e.g., the occupant's ear area) in complex and dynamic acoustic environments such as a vehicle cabin. Alternatively, the final estimated virtual microphone noise signal at the virtual microphone location... It is possible to estimate the noise signal from only the weighted signal. The superposition (i.e., the superposition of noise signals along the first signal path 412) is generated. According to another embodiment, the final estimated virtual microphone noise signal at the virtual microphone location... Estimated noise signal from only weighted filtering The superposition (i.e., the superposition of noise signals along the second signal path 416) is generated.

[0052] Refer again Figure 3 Estimated virtual microphone noise signal This can be provided as feedback to the adaptive filter controller 320, which is then used to adaptively update the filter coefficients of adaptive filters 318 and 360. According to one or more embodiments, the EOC system 300 may include an internal error loop 368 for the LMS-based adaptive filter controller 320. In this case, the estimated virtual microphone noise signal can be added at adder 370. With internal noise immunity signal The components are combined to generate an internal error signal g[n]. This internal error signal g[n] can then be provided as feedback to the adaptive filter controller 320. An internal error loop 368 can be employed in the MFxLMS system to adapt the adaptive filter 318 using a relatively simple LMS algorithm, which can accelerate convergence by avoiding the delays introduced by other LMS-based systems. For example, compared to other LMS algorithms, the internal error loop 368 allows for larger step sizes, resulting in faster convergence.

[0053] Figure 5 This is a flowchart depicting a method 500 for estimating noise at a virtual microphone location in an EOC system (such as EOC system 300). One or more steps for estimating the noise at the virtual location may be performed by a virtual location noise estimator 366. For example, the virtual location noise estimator 366 may receive multiple estimated noise signals. As provided in step 510. Each estimated noise signal It may be based at least in part on the error signal e[n] from each of the plurality of error microphones 312. According to one or more embodiments of this disclosure, the plurality of estimated noise signals... It can be split into two signal paths, such that each estimated noise signal It can be transmitted along both the first signal path 412 and the second signal path 416, as provided at step 520.

[0054] At step 530, weights can be applied to each of the multiple estimated noise signals in the first signal path. For example, the first group of weights is 410 (W) pv It can be applied to multiple estimated noise signals. To generate multiple weighted estimated noise signals As mentioned earlier, each weight in the first set of weights 410 can be individually selected based on the current vehicle parameters or conditions (such as engine order frequency) and adaptively applied to the corresponding estimated noise signal.

[0055] At step 540, the virtual path filter 414 can be used to filter each estimated noise signal in the second signal path. The filtering is performed based on the transfer function H between each corresponding error microphone position (P0, P1, P2, P3) and the virtual microphone position (V0). pv [p][v] Modeling to generate multiple filtered estimated noise signals At step 550, weights can be applied to multiple filtered estimated noise signals in the second signal path. Each of them. For example, the second group has a weight of 418 (R). pv It can be applied to multiple filtered estimated noise signals. To generate multiple weighted and filtered estimated noise signals As mentioned earlier, each weight in the second set of weights 418 can be individually selected based on the current vehicle parameters or conditions (such as engine order frequency) and adaptively applied to the corresponding filtered estimated noise signal.

[0056] At step 560, each signal from the first signal path 412 (i.e., unfiltered) and each signal from the second signal path 416 (i.e., filtered) can be combined to generate an estimated virtual microphone noise signal indicating the noise at the first virtual microphone position V0. For example, the virtual position noise estimator 366 can be based on multiple weighted estimated noise signals from the first signal path 412. and multiple weighted filtered estimated noise signals from the second signal path 416 The superposition of these signals is used to generate the estimated virtual microphone noise signal. The weights applied to each filtered and unfiltered signal can be adaptively adjusted based on the current vehicle conditions to increase the proportion of each corresponding signal in the final superposition, thereby obtaining an estimated virtual microphone noise signal.

[0057] In addition to providing an accurate estimate of noise at virtual locations where no physical microphone is present, the EOC system 300 can reliably estimate noise at virtual locations across different frequency ranges in complex and dynamic acoustic environments. Furthermore, the EOC system 300 allows for a reduction in the final physical microphone count within the vehicle compartment. Many conventional EOC systems rely on dedicated microphones to approximate the location of an occupant's ears. Moreover, these dedicated EOC microphones are additional microphones used in some vehicles for voice recognition and hands-free telecommunications, resulting in a total of seven or more microphones. Adaptively weighting filtered and / or unfiltered microphone signals and mixing at least some of the signals to provide an accurate estimate of noise at virtual microphone locations remote from physical microphones allows the EOC system to utilize other vehicle microphones for EOC purposes. Therefore, the systems and methods of this disclosure can also reduce the overall vehicle microphone count by reusing at least some existing voice microphones for EOC purposes. This can further reduce costs and is useful when design requirements dictate fewer microphones placed in locations that are not traditionally optimal for EOC.

[0058] Any one or more of the controllers or devices described herein include computer-executable instructions that can be compiled or interpreted from computer programs created using various programming languages ​​and / or techniques. Typically, a processor (such as a microprocessor) receives and executes instructions, for example, from memory, a computer-readable medium, etc. Processing units include non-transitory computer-readable storage media capable of executing the instructions of a software program. Computer-readable storage media can be, but is not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof.

[0059] Those skilled in the art will understand that functionally equivalent processing steps can be performed in the time domain or the frequency domain. Therefore, although each signal processing block is not explicitly illustrated in the figures, signal processing can be performed in the time domain, the frequency domain, or a combination thereof. Furthermore, while the various processing steps are explained in typical digital signal processing terminology, equivalent steps can be performed using analog signal processing without departing from the scope of this disclosure.

[0060] While exemplary embodiments have been described above, these embodiments are not intended to describe all possible forms of the invention. Rather, the terminology used herein is descriptive rather than restrictive, and it should be understood that various changes may be made without departing from the spirit and scope of the invention. Furthermore, features of various implementations may be combined to form other embodiments of the invention.

Claims

1. A method for estimating noise at a virtual microphone location for an engine order cancellation (EOC) system, the method comprising: Receive multiple estimated noise signals indicating the location of noise at each of a plurality of error microphones, wherein each estimated noise signal is at least partially based on the error signal from each of the plurality of error microphones; Each estimated noise signal is filtered using a virtual path filter modeled based on the transfer function between each corresponding error microphone position and the virtual microphone position to generate multiple filtered estimated noise signals. Each filtered estimated noise signal is adaptively weighted using weights that vary based on the current vehicle condition to generate multiple weighted filtered estimated noise signals; as well as A virtual microphone noise signal, indicating the noise at the virtual microphone location, is generated by superimposing at least the plurality of weighted and filtered estimated noise signals.

2. The method of claim 1, wherein each virtual path filter is a finite impulse response filter.

3. The method of claim 1, wherein each weight is selected from a plurality of weights, the plurality of weights being stored in a lookup table and derived based on the current vehicle condition to be applied to each filtered estimated noise signal.

4. The method of claim 1, wherein the virtual microphone position corresponds to the occupant's ear.

5. The method of claim 1, further comprising: The estimated virtual microphone noise signal is received at the adaptive filter controller; as well as The adaptive transfer characteristics of the adaptive filter are adjusted in part based on the estimated virtual microphone noise signal.

6. The method of claim 1, wherein the current vehicle condition includes the frequency of engine order noise, such that each weight varies at least based on the frequency.

7. The method of claim 6, wherein the current vehicle condition further includes at least one of engine load and vehicle speed, such that each weight is also selected based on at least one of the engine load and the vehicle speed.

8. The method of claim 1, wherein each of the plurality of estimated noise signals is divided into two signal paths comprising a first signal path and a second signal path, wherein the output of the second signal path comprises the plurality of weighted filtered estimated noise signals, the method further comprising: Each estimated noise signal in the first signal path is adaptively weighted using weights selected based on the current vehicle condition to generate multiple weighted estimated noise signals.

9. The method of claim 8, wherein generating the estimated virtual microphone noise signal indicating noise at the virtual microphone location based on the superposition of at least the plurality of weighted filtered estimated noise signals comprises: An estimated virtual microphone noise signal indicating the noise at the virtual microphone location is generated by superimposing the plurality of weighted estimated noise signals from the first signal path and the plurality of weighted filtered estimated noise signals from the second signal path.

10. An engine order elimination (EOC) system, the engine order elimination system comprising: At least one adaptive filter is configured to generate an anti-noise signal based on an adaptive transfer characteristic and a reference signal received from a reference signal generator, wherein the adaptive transfer characteristic of the at least one adaptive filter is characterized by a set of filter coefficients. An adaptive filter controller, including a processor and memory, is programmed to adjust the set of filter coefficients based on the reference signal and an estimated virtual microphone noise signal indicating noise at the virtual microphone location. as well as A virtual position noise estimator, communicating with at least the adaptive filter controller, comprising a processor and a memory, is programmed to: Receive multiple estimated noise signals indicating the location of noise at each of a plurality of error microphones, wherein each estimated noise signal is at least partially based on the error signal from each of the plurality of error microphones; Each estimated noise signal is filtered using a virtual path filter modeled based on the transfer function between each corresponding error microphone position and the virtual microphone position to generate multiple filtered estimated noise signals. Each filtered estimated noise signal is adaptively weighted using weights selected and changed based on the current vehicle condition to generate multiple weighted filtered estimated noise signals. as well as The estimated virtual microphone noise signal, indicating the noise at the virtual microphone location, is generated by superimposing at least the plurality of weighted filtered estimated noise signals.

11. The system of claim 10, wherein each weight is selected from a plurality of weights derived based on the current vehicle condition and applied to each filtered estimated noise signal.

12. The system of claim 10, wherein the current vehicle condition includes the frequency of engine order noise, and each weight varies at least based on the frequency.

13. The system of claim 10, wherein each virtual path filter is a finite impulse response filter.

14. The system of claim 10, wherein each of the plurality of estimated noise signals is divided into two signal paths comprising a first signal path and a second signal path, wherein the output of the second signal path comprises the plurality of weighted filtered estimated noise signals, and the virtual position noise estimator is further programmed to: Each estimated noise signal in the first signal path is adaptively weighted using weights selected based on the current vehicle condition to generate multiple weighted estimated noise signals.

15. The system of claim 14, wherein the virtual location noise estimator is programmed to generate the estimated virtual microphone noise signal indicating noise at the virtual microphone location based on the superposition of the plurality of weighted estimated noise signals from the first signal path and the plurality of weighted filtered estimated noise signals from the second signal path.

16. The system of claim 15, wherein the virtual microphone position is a fixed point in space corresponding to the occupant's ear and at a certain distance from the plurality of error microphones.

17. A method for estimating noise at a virtual microphone location for an engine order cancellation (EOC) system, the method comprising: Receive multiple estimated noise signals indicating the location of noise at each of a plurality of error microphones, wherein each estimated noise signal is at least partially based on the error signal from each of the plurality of error microphones; Each estimated noise signal is transmitted along the first signal path and the second signal path; Weights are applied to each estimated noise signal in the first signal path, and each weight is individually selected and varied based on the current vehicle condition to generate multiple weighted estimated noise signals. Each estimated noise signal in the second signal path is filtered using a virtual path filter modeled based on the transfer function between each corresponding error microphone position and the virtual microphone position to generate multiple filtered estimated noise signals. Weights are applied to each filtered estimated noise signal in the second signal path, with each weight being individually selected and varied based on the current vehicle condition to generate multiple weighted filtered estimated noise signals. as well as A virtual microphone noise signal, indicating an estimate of the noise at the virtual microphone location, is generated by superimposing the plurality of weighted estimated noise signals from the first signal path and the plurality of weighted filtered estimated noise signals from the second signal path.

18. The method of claim 17, wherein the proportion of each signal applied in the superposition is adaptively controlled by its corresponding weight based on the current vehicle condition.

19. The method of claim 18, wherein the current vehicle condition includes the frequency of engine order noise, such that each weight is selected and varied individually based at least on the frequency.

20. The method of claim 17, wherein the virtual microphone location corresponds to the occupant's ear.

Images