Method and apparatus for suppressing narrow-band noise in passenger cabin of vehicle. RU patent 2504025.

FIELD: physics.

SUBSTANCE: apparatus for suppressing noise in the passenger cabin of a vehicle has at least one converter, a programmable computer and at least one acoustic sensor. The computer is configured to apply the electroacoustic model of the passenger cabin to the model of an adjustment system, having a fixed coefficient master controller which is connected to a variable coefficient unit, having a Yule parameter in form of a Yule Q unit. In the method, the first step involves determining and calculating the electroacoustic model and control law for at least one predetermined noise frequency. At the second step, the computer applies the control law to the electroacoustic model in real time in accordance with the current noise frequency to be suppressed.

EFFECT: improved method.

15 cl, 14 dwg

The present invention relates to a method and apparatus for noise suppression in the passenger cabin of the vehicle, in particular, vehicle, through active management. It finds its application in industry-related equipped with motor vehicles, and this term is understood in its broadest sense, including, in particular, light vehicles, heavy vehicles, road vehicles, railway vehicles, boats, barges, submarines, and in the field of electro-acoustic devices, such as car radios that can be added to this kind of function.

Some of the acoustic noise in the passenger compartment of a vehicle can have a wide variety, and others may be approximately . This, in particular, in the case of noise, generated by the rotation of the crankshaft that is known as the "humming noise", which is expressed in the noise spectrum which consists of lines whose frequencies are proportional to the frequency of crankshaft rotation, with one main frequency and harmonics.

These frequencies are changed in accordance with the speed of rotation of the crankshaft, but they, however, can be known with certainty thanks to the information, received from an tachometer, usually built into the vehicle.

Earlier already it was proposed to reduce, or even suppress these noises by means of active acoustic means. In connection with these, we can mention the description of the-art in the field of active control used in motor vehicles, the Elliot (Elliot) in December 2008, in an article entitled: "A review of the active noise and vibration control in road vehicle" (ISVR technical memorandum no. 981 - University of Southampton) (Review of active protection against noise and vibration in a road vehicle" (ISVR technical Memorandum №981 - University of Southampton).

There are two main systems of active acoustic control. First, the so-called system of "direct line" or with advance payment. Such a system needs loudspeaker, microphone determine the error, in which it is desirable to eliminate the noise and the regulator accepts a reference signal, correlated with the signal, subject to the elimination of nitric correction signal supplied to the loudspeaker. Such a system in schematic form is shown in figure 1, showing the prior art. Such a system, in particular, gave rise to a number of algorithms based on the method of Minimum mean squared error" (LMS-way): Fx-LMS, FR-LMS, the purpose of which is to minimize, in the sense of least squares, the output signal from the microphone to determine the error, and implement it through the processing of the reference signal.

In addition, in the case of the so-called system of "direct link" you can refer to the article Sano et al. (Sano and others), "NV counter-measure technology for a cylinder - On-Demand Engine-Development of active booming noise control applying adaptive notch filter" (SAE 2004), ("Technology countermeasures noise for the cylinder is performed at the request of debugging engine actively managed to suppress a buzzing noise, applying adaptive narrow-band rejection filter" (Society of automotive engineers (USA) 2004)). The authors present the algorithm based on adaptive ( filter, known frequency attenuation. The device is based on the algorithm, the structure of which is of the type "direct link", called FR-SAN, which is an adaptation algorithm FR-LMS, in the case where the noise subject to impairment refers to type. The implementation of this algorithm are not considered problems arising from a change in the transfer function of the passenger compartment, for example, in function of the number of passengers. In addition, in this manner, it is impossible to know, otherwise than experimentally, the description of the management system at frequencies other than the frequency at which it operates.

Secondly, the so-called system of "feedback" or with counter-reaction. Such a system is schematically shown in figure 2, showing the prior art. This system, unlike the so-called system of "direct link"does not need the reference signal. In this case, it is a traditional system with feedback, and can be used all the tools the traditional technique of automatic control (in particular, measure of robustness, stability analysis, performance). In particular can analyze the robustness of a closed system with respect to a change in the transfer function of the passenger compartment. Can also be investigated frequency response of the system, not only on the frequency of the suppression of the disturbances, but also at other frequencies.

The present invention relates to this second category of the so-called system of "feedback". To put it more specifically, it applies to the executable in the real-time active method for weakening through feedback narrowband noise, essentially of at least one specific frequency, in the passenger cabin of the vehicle by emitting the sound at least one Converter, usually loudspeaker, managed signal u(t) or U(t) based, respectively, on a case-SISO (single input - one output) or MIMO (multiple input-multiple outputs), generated programmable computing device, depending on the signal acoustic measurements y(t) or Y(t) in accordance with this case to be performed by at least one acoustic sensor, usually microphone, use one sensor corresponds to the case of SISO, a single input - single output - one variable, and using multiple sensors corresponds to the case MIMO range of inputs, a lot of exits variables, and on the first stage of designing the electroacoustic characteristics of the link, educated passenger cabin, the transducer and sensor is simulated by the electro-acoustic model as electro-acoustic transfer function, which is defined and calculated; after that define and calculate the control law based on the global model of the system in which this control law applies to electroacoustic transfer function, the output of which is optionally accepts signal to noise subject to impairment (R(t))to get the signal y(t) or Y(t) at the design stage, with the above mentioned law, management allows to produce the signal u(t) or U(t) as a function of acoustic measurements y(t) or Y(t), and the second stage is the use of the calculated control law is used in computing device to receive the signal u(t) or U(t), then sent to the transducer, depending on the signal y(t) or Y(t), adopted by the sensor, to weaken referred noise.

In accordance with the invention of the implemented control law, which contains the ula (Youla) to the Central controller and that is that in the aforementioned act only control parameter whirligig are coefficients that depend on noise frequency to be weakening, Central regulator has constant coefficients, the option of ula has the form of a filter infinite impulse response, and, after the definition and calculation of control law in memory of a computing device are kept at least above-mentioned variables, the coefficients, preferably in the table as a function of a certain frequency (certain frequency) noise p(t), which is used at the stage of design, and point-of-use, real-time:

- learn the current frequency noise, subject to impairment

make your computing device count control law contains a Central regulator with a parameter of July, using as a parameter whirligig stored in the memory of the coefficients of a certain frequency corresponding to the current frequency noise to be weakening.

In other words, the implemented control law, which contains part with constant coefficients, called a Central controller, and part with the odds as a function of noise frequency to be weakening, which is only a parameter of July, and the part of the regulator with variable coefficients is a filter with infinite impulse response, and after the determination and calculation of the control law, at least mentioned variables, the coefficients are stored in the memory of a computing device, preferably in the table as a function of a certain frequency (frequency) noise p(t), which is used at the stage of design, and point-of-use, real-time: find out the current frequency noise, subject to impairment and make computing device count control law contains a Central controller with constant coefficients to part with variable coefficients, using the as part with variable coefficients stored in the memory of the coefficients of a certain frequency corresponding to the current frequency noise to be weakening. Therefore, in the framework of the invention, to dampen the noise on at least one specific frequency, implemented the Central controller with constant coefficients, joined by a block with variable coefficients, which is a parameter of July in the form of a block (Q) of ula.

In the framework of the invention of the term "signal" refers to analog signals, such as electric signal, directly from the microphone and digital signals, as, for example, to output a signal block whirligig Q(q-1 ). It should also be understood that the terms "Converter" and "sensor" is used in General and functional significance, and that, in practice, these terms are associated interface circuitry, such as, in particular, an analog-to-digital or digital-to-analog converters filter(s) anti-aliasing, power (amps) (for speakers (speakers) and a microphone (s)). The term "signal" also covers cases: SISO, one input, one output, one variable (one sensor and, therefore, one entrance of acoustic measurements), and MIMO range of inputs, a lot of exits variables (several sensors and, therefore, multiple inputs acoustic measurements), whatever the number of speakers. Thus, the invention may be applied to the occasion SISO, one input, one output, one variable (the only microphone, that is, one only place in the passenger compartment will be weakened noise), and the occasion of MIMO range of inputs, a lot of exits variables (multiple MICS, that is, the same number of seats that will be weakened noise). One should also know that the invention is applied to a weakening of both the noise that takes place on that particular frequency, essentially constant over time (for example, noise refrigeration compressor in the truck)and noise, the frequency of which may change over time, and in this case, preferably at the design stage to define and calculate parameters of a whirligig, block Q(q -1 ), for several specific frequencies so that during the stage to take the result of calculating the parameter whirligig for that particular frequency, which corresponds to (equal to or near, that is, in fact, the best matches or otherwise against) current frequency noise to be weakening. You should understand that the smaller will be the channel step increments, the higher chance to get the result of calculating the parameter of July with a certain frequency, which corresponds to the frequency of the current noise to be weakening. In fact, it will be shown that in the law of management only option ula (in practice, it's coefficients) is a variable as a function of frequency noise, unlike the coefficients of the Central controller that remain constant and independent of the frequency noise.

It can be noted that parametrization of the ula (Youla) has been used to suppress the sinusoidal disturbance in a completely different field of technology: the management of vibrations active suspension. The relevant article is: "Adaptive narrow disturbance applied to an active suspension - an internal model approach" (Automatica 2005) "Adaptive narrowband indignation applied to the active suspension - internal model approach" (Automatic 2005), the authors of which are I.D.Landau et al. (.. and others). In this latter device option ula has a form of filter, finite impulse response (transfer function with a single polynomial without the denominator), whereas in the present invention will be shown that this option ula has the form of a filter infinite impulse response (transfer function with the numerator and the denominator). In addition, in this article, the calculation of the coefficients of the parameter of July is executed by means of adaptive devices, that is, information about the frequency of the perturbation is not known, in contrast to the present invention, in which this frequency is known on the basis of measurements, in particular, by the counter of the number of revolutions, and in which the coefficients of the parameters of ula is stored in tables to be used in the real time. Device and method corresponding to the invention, provides control law with a much more robust. In the specific case of the invention, this corresponds to a dead control law to changes in the parameters of the electro-acoustic model, i.e. to changes in passenger cabin configuration, that from the point of view of the industrial production is a fundamental element.

Also may be referred to the article "Adaptive control for interior noise control in rocket fairings" ("Adaptive management internal protection from noise in the missile gear"), Mark A.Mcever, 44-th conference AIAA/ASME/ASCE/AHS (American Institute of Aeronautics and Astronautics / ASME / American society of civil engineers) Structures, structural dynamics and Structures, dynamics of constructions and construction materials), April 7-10, 2003 Here again the option of ula is a filter with finite impulse response (FIR)filter, which creates problems with regard to the robustness of the system, the algorithm is adaptive and are not specifically intended for the suppression of a specific frequency.

- implemented the design stage on programmable computing device,

- determined and calculated parameter whirligig by discretization of continuous transfer function of the second order,

- at the second stage, at the design stage define and calculate polynomials Ro(q -1 ) and So(q -1 ) Central regulator so that the mentioned Central regulator himself provided the regulation reserve gain and phase, without the purpose of suppression of the disturbances,

- In case of SISO, one input, one output, one variable at the design stage:

a) on the first stage uses linear electro-acoustic model, and electro-acoustic model has the form of a discrete rational electroacoustic transfer function, and the said audio model is defined and calculated by the acoustic excitation of the passenger compartment through the Converter and acoustic measurements sensor, using then the process of identification of linear systems running with these measurements and model

b) to implement the second stage the Central controller, which is applied to a specific thus calculated and electroacoustic model, and the Central regulator has the form RS-regulator, consisting of two blocks

and Ro(q -1 ), and in the Central controller unit

generates a signal u(t) and takes as input an inverted output signal block Ro(q -1 )referred to block Ro(q -1 ) takes as input signal y(t)corresponding to the amount of noise p(t) and output signal electroacoustic transfer function of electroacoustic model, and the Central controller define and calculate,

c) the third stage to the Central controller attached parameter of July, which, therefore, represents a transfer block with variable coefficients, forming control law, while the parameter of ula has the form of a block Q(q -1 ), filter infinite impulse characteristics with

attached to the Central RS-regulator, and referred to block Q(q -1 )) whirligig, takes noise estimation obtained by calculation of signal u(t) and y(t) and as a function of electroacoustic of the transfer function and the output signal of the said unit (Q(q -1 )) ula is deducted from the inverted signal Ro(q -1 )sent by the entrance of the block

Central RS-regulator and the whirligig accordingly represents a transfer block with variable coefficients in the law of the control that contains the Central controller is bound to a parameter of July, determined and calculated for at least one frequency (p(t)) noise, including at least this specific frequency noise to be weakening, and point-of-use, real-time:

- learn the current frequency noise, subject to impairment

make your computing device count control law, containing RS-controller with a parameter of July, using as a parameter the whirligig coefficients calculated for noise frequency corresponding to the current frequency noise to be weakening, with the Ro(q -1 ) and So(q -1 ) are constant coefficients,

- at the design stage in the case of SISO, one input, one output, one variable, run the following transactions:

a) the first leg of the passenger cabin is subjected to acoustic stimulation, feeding on the transducer excitation signal, the spectral density, which is essentially a uniform on the effective frequency band

b) - the second stage define and calculate polynomials Ro(q -1 ) and So(q -1 the Central controller so that the mentioned Central regulator is equivalent to that of the regulator, which is calculated by means of placement of the poles of the closed-loop when using the Central regulator to electroacoustic transfer function, and n poles of the closed-loop placed on the n poles of the transfer function of the electroacoustic system,

c) the third stage for at least one frequency (p(t)) noise, including at least this specific frequency noise, subject to impairment define and calculate the numerator and denominator of the block (Q(q -1 )) ula in the law management as a function of the criterion of weakening, with the Q(q-1 ) is expressed in the form of relations

thus, to get the values of the coefficients of polynomials alpha(q -1 ) and? (q -1)/each frequency, the calculation of b(q-1 ) and C(q -1 ) is performed by obtaining the discrete transfer function

resulting from the discretization of the continuous transfer function of the second order polynomial b(q -1 ), is calculated by solving the equation zomb (Bezout),

and point-of-use, real-time, run the following transactions:

- computing device forced to rely control law, the Central controller with constant coefficients and parameter whirligig with variable coefficients, thus, to produce the signal u(t), sent by the Converter, as a function of acoustic measurements y(t), and using it to block (Q(q -1 )) whirligig the values of the coefficients of polynomials alpha(q -1 ) and? (q -1 )defined and calculated for a certain frequency corresponding to the current frequency,

- calculation of the noise assessment is obtained by applying the numerator of electroacoustic transfer function for u(t) and subtracting the result from the application of y(t) to the denominator of electroacoustic transfer function

- for electro-acoustic model used electro-acoustic transfer function, which has the form:

where d is the number of components of the delay in sampling periods in the system, and are polynomials q -1 , which have the form:

B(q -1 )=b 0 +B1 ·q -1 +...b nb ·q-nb

A(q -1 )=1+ a

1 ·q -1 +a...

na ·q-na

where b i and a

i represent a scalar value and q -1 represents the operator delay period sampling and estimates of the noise produced by applying the function q-d B(q -1 ) u(t) and subtracting the result from the application of y(t) to the function A(q -1 ),

for phase (b) polynomials Ro(q -1 ) and So(q -1 ) Central regulator is determined and calculated by way of placement of poles, n dominant poles closed circuit equipped with a Central controller is chosen equal to the n pole of electroacoustic transfer function, and m auxiliary poles are poles, located on a high frequency

- at the design stage:

a) on the first stage we use the linear electro-acoustic model, the electro-acoustic model has the form of the state representation, consisting of a matrix blocks: N, W, G and q -1 , I, and G is the transition matrix, and N represents the input matrix, W represents the output matrix and I is the identity matrix, said the view of the condition can be expressed by a recurrent equation:

X(t+Te)=G·X(t)+H·U(t)

Y(t)=W·X(t)

where X(t): state vector, U(t)is the vector of input signals Y(t): vector output signals

and speaking of electro-acoustic model is defined and calculated by the acoustic excitation of the passenger compartment through converters and acoustic measurements sensors, using then the process of identification of linear systems running with these measurements and model

b) to implement the second stage the Central controller, used to determined and calculated by the model, and the Central regulator has the shape of an "observer" status and feedback on the estimated state, which iteratively expresses

, state vector of the "observer", as a function of Kf, gain factor "observer", Kc, vector feedback estimated as, as previously defined and calculated electroacoustic model, i.e.:

where the control action

and referred to the Central controller define and calculate,

c) the third stage to the Central controller attached parameter (ula), which thus represents a transfer block with variable coefficients, forming control law, while the parameter of July, has the form of a block (Q) for MIMO range of inputs, a lot of exits - variables consisting of matrices (AQ), (BQ), (CQ) States, attached to a Central controller, also resulting in the form of presentation of state block (Q), whose output signal, folded with the output signal of the Central regulator, gives the signal, which forms the signal, the opposite U(t)is the input signal Y(t), which is deducted signal

the parameter of July, respectively, representing a transfer block with variable coefficients in the law of the control that contains the Central controller is bound to a parameter of July, determined and calculated at least for one frequency (p(t)) noise, including at least this specific frequency noise, subject to impairment calculation of coefficients matrixes: AQ, BQ, CQ, is performed by obtaining the discrete transfer functions

resulting from the discretization of the continuous transfer functions of the second order, and by placing the poles, as well as the asymptotic solution of the equation suppression,

and, point-of-use, real-time:

- learn the current frequency noise, subject to impairment

make your computing device count control law contains the Central controller with constant coefficients and parameter whirligig with variable coefficients, using as a parameter the whirligig coefficients that were calculated for the noise frequency corresponding to the current frequency noise, subject to impairment

- at the design stage in the case of MIMO range of inputs, a lot of exits variables, execute the following operations:

a) the first leg of the passenger cabin is subjected to acoustic stimulation, feeding on the converters excitation signals, the spectral density, which is essentially a uniform on the effective frequency band moreover, the signals of excitation are in relation to each other,

b) - the second stage define and calculate the Central regulator so that it was equivalent to a controller with "observer" status and feedback calculated as by the poles when applying Central regulator to electroacoustic transfer function, and, with this purpose, selected zero gain "observer", that is Kf=0 (gain "observer" is equal to zero matrix), and the coefficient (Kc) feedback gain as selected in such a way to enter this the contour high frequency pole to ensure the robustness of the control law with parameter of July, and the calculation of the Kc is, for example, by the linear-quadratic optimization (LQ-optimization).

c) the third stage when considering the view zoomed "observer" status is determined and calculated poles block (Q) ula in the law of management, for at least one frequency (P(t)) noise, including at least this specific frequency noise to be weakening as a function of the criterion of weakening, so as to obtain the values of the coefficients of the parameter Ula for the/each frequency,

and point-of-use, real-time, run the following transactions:

make your computing device count control law, the Central controller with constant coefficients and parameter whirligig with variable coefficients, generating the signal U(t)is sent on a converters, as a function of acoustic measurements Y(t), and using the parameter whirligig the values of the coefficients defined and calculated for a certain frequency corresponding to the current frequency,

- at the second stage the calculation of the Kc is performed by means of linear-quadratic optimization (LQ-optimization)

- the method is fit for a set of specific frequency noise, subject to impairment and phase () is repeated for each of these specific frequencies, and point-of-use, in case when none of these specific frequencies does not match the current frequency noise, subject to impairment is made interpolation on the current frequency values of the coefficients of bloc (Q) whirligig, based on the values of the coefficients of the said unit (Q) whirligig, which are known for these specific frequencies,

- discretization of signals is made with a frequency Fe, and on stage (a) effective frequency band used for the excitation signal is essentially equal to [0, Fe/2],

- excitation signal has a uniform spectral density,

- before the stage of the application at the design stage, added the fourth stage (d)intended for checking the stability and robustness of the model of the electroacoustic system and control law, the Central regulator with a parameter of July, prior to the stages ((a) to (C)), by modeling the application of the control law, received at the stages of (b) and (C), the electro-acoustic model, obtained at the stage of (a), for the this particular frequency (these specific frequencies), and in case when some pre-defined criterion of stability and/or robustness is not satisfied, is the repetition of at least stage () when you change the criterion of weakening,

- at the fourth stage (d) at the design stage, in case, when the specified criterion of stability and/or robustness is not satisfied, the additional repetition phase (b) changes to supporting poles of the closed-loop,

- design stage is a preliminary stage and it is only done once, previously in relation to the stage of use, preserving in memory of results of definition and calculation to use point-of-use (for example, in the case of for SISO systems with a single input and a single output), are kept in the memory coefficients blocks R, S and Q of the calculated control law, as calculated electro modulation transfer function for the unit (Q) factor tables that can be implemented due to settlements for several specific frequencies),

- criterion weakening selected as a function of at least one of the two following elements: depth of attenuation (amplitude) and the width of the band attenuation

- current frequency noise, subject to impairment is recognized from the measurements produced by the counter of revolutions of the motor vehicle.

If we describe it in more General terms, the invention also relates to a device that is specifically tailored to the implementation of the method according to the invention, in order to weaken the narrow-band noise, essentially, on at least one specific frequency, the device contains at least one Converter, usually loudspeaker, managed signals, generated by the programmable computing device as a function of signal acoustic measurements performed at least one acoustic sensor, usually microphone, the law control is defined and calculated in the first stage-design referred to the calculated control law is used on the second stage of use, in a computing device, to generate a signal emitted by the transducer as a function of the signal received from the detector to weaken referred noise, and the device according to the invention, contains tools for implementation, in computing device, control law contains the whirligig to a Central regulator, while in the aforementioned act only control parameter ula has coefficients that depend on noise frequency to be weakening, Central regulator has constant coefficients, and the memory of a computing device stores at least mentioned variables coefficients, preferably, in a table as a function of a certain frequency (certain frequencies) R(t) noise used at the design stage.

The invention also refers to the medium, with the teams for the direct or indirect control of a computing device in such a way that it operated in accordance with the method according to the invention, and, in particular, real-time point-of-use.

The present invention will now be described in more detail, but without overlap, thus it is restricted by following description with reference to the attached drawings, in which:

figure 1, the relevant prior art is a schematic representation of the so-called system of "direct line" or proactive compensation for noise reduction systems;

figure 2, the relevant prior art is a schematic representation of the so-called system of "feedback" or with counter-reaction in the system attenuation;

figure 3, relevant prior art is a schematic representation of the principal scheme of electroacoustic closed system with control law for the passenger compartment of the vehicle;

figure 4 is a schematic representation of the time of excitation of a real acoustic system of the passenger compartment of a vehicle that is designed to define and calculate the electroacoustic model that will be used;

figure 5 is a representation of a closed system, corresponding electroacoustic model with the regulator RST-type, referred to as the Central the regulator, at T=0 and in the case of SISO, one input, one output, one variable;

figure 6 is an example of a direct function of the sensitivity and shows that, through the application of theorem Bode-district of Freudenberg-Luz (Bode-Freudenberg-Looze), the two areas, which are located above and below the axis of 0 dB, are equal to each other;

Fig.7 is a representation of a case-control law in the case of SISO, one input, one output, one variable applied to electroacoustic model and contains the Central controller's RS-type to which is attached the option of ula;

Fig.8 is a representation of the full schema management act with a Central controller RS-type to which is attached a parameter of July, and calculated in real-time point-of-use, to reduce noise in the passenger compartment;

figure 9 is a representation of a scheme to transfer the system consisting of 2 loudspeakers and two microphones, and, consequently, in the case of MIMO range of inputs, a lot of exits variables;

figure 10 is a representation of the structural scheme of the system, subject to regulation, that is, electro-acoustic model of the passenger compartment in case of MIMO range of inputs, a lot of exits - variables;

figure 11 is a representation of the structural scheme of the Central regulator, in the case of MIMO range of inputs, a lot of exits variables;

fig.12 is a representation of the structural scheme of the Central regulator applied to electroacoustic model of the passenger compartment in case of MIMO range of inputs, a lot of exits variables;

fig.13 is a representation of the structural scheme of the control law, the Central controller + parameter of July, applied to the electro-acoustic model of the passenger compartment in case of MIMO range of inputs, a lot of exits variables;

Fig.14 is a representation of the structural scheme of the control law, the Central controller + parameter of July, used in real time to dampen the noise, in the case of MIMO range of inputs, a lot of exits variables.

To synthesize control law, a model of a real system, composed of electro-acoustic and acoustic elements of passenger compartment, including loudspeaker (loudspeakers) (converters), microphone(s) (sensor), the associated electronic item (associated electronic components) (amplifiers, converters...). This model, referred to as "electro-acoustic model", must be in the form of a rational transfer function, that is, it should behave as a discrete filter infinite impulse response.

Note that because the computing device is digital, the implemented analog-to-digital and digital-to-analog converters, in particular, for the implementation of sampling analog signals. Thus, the computing device processes the sampled signals with The period (in seconds) and the frequency Fe=1/(in Hertz).

Taking into account the level of signals involved, the preferred option may be performed linear approximation of the real system, composed of electro-acoustic and acoustic elements of passenger compartment. In advanced alternative embodiments of the invention may also be designed to avoid nonlinear phenomena saturation or the like (for example, compression/expansion of the signals pass filter anti-aliasing...).

It should also take into account the fact that the equations governing the real response of the passenger compartment, are the equations in partial derivatives, that is, the transfer function of representing the accuracy of the real system, has finite dimension (model with distributed parameters). Thus, for realization of the invention necessary to find a compromise for the determination of electroacoustic model, and the order of the transfer function of this model is chosen with dimensionality, which reduced sufficiently to not cause too much computing, but is big enough to properly approximated by the model. This restriction results in the fact that you can avoid excess sample. In order example: maximum frequency of disturbance noise component 120 Hz can be selected sample rate, amounting to 500 Hz. One of the advantages of choosing to moderate the frequency of sampling is that it reduces the processing load on the inside of the car computing device. It should be noted that, since this amplifier speaker has a much higher sampling frequency (or even works with analog components), it is advisable to place between the output of a computing device and input loudspeaker low-pass filter that operates at a frequency loudspeaker amplifier, the frequency cut-off referred to filter is constant, in order to reduce the harmonic distortion, caused by transition between signals of different sampling periods.

In the framework of the present invention has been selected some specific form of electro-acoustic model, which will now be described. However, you should understand that in the framework of the present invention can be used and other forms of electro-acoustic model, and, in particular, in the case where the determination and calculation system of attenuation applied to the electroacoustic model does not give a satisfactory solution (see the description after the implementation of the additional time checking the stability and robustness of the model of electroacoustic systems and RS-regulator, with a parameter of July, during the design stage).

The transfer function of the electro-acoustic model that describes the response of the real electroacoustic system, can be expressed, between the points of u(t) and y(t) of the system, in the absence of any closed path. Let q -1 will be operator of the delay period, the sampling rate, the required transfer function, in the absence of any closed loop and noise (noise, subject to a weakening of the missing), has the following form:

where d is the number of components of the delay in sampling periods in the system,

In and are polynomials q-1 , q -1 represents the operator delay period of sampling. In particular:

B(q -1 )=b 0 +B1 ·q -1 +...b nb ·q-nb

A(q -1 )=1+ a

1 ·q -1 +a...

na ·q-na

where b i and a

i represent a scalar value.

The identification shall be performed by the excitation of a real system via the signal u(t), the spectral density, which is essentially a uniform in the frequency range [0, Fe/2], where Fe/2 represents the Nyquist frequency. It should be understood that the frequency (frequency) noise to be weakening, should also include in this same interval, and Fe, thus, is selected as a function of the high frequency noise to be weakening. This kind of later periods, while the excitation signal can be created, for example, pseudorandom binary sequence (PRBS-sequence). The excitement, schematically shown In figure 4, is the absence of disturbance of external noise. All test data u(t) and y(t)obtained for the time of testing of the real system (passenger cabin with his electric-acoustic components), written thus to be processed in privileged conditions batch processing.

Algorithms that can be used for identification of linear systems are numerous. For a brief overview of methodologies that can be used, you can refer, for example, to work I.D. Landau: "Commande des systemes" (I.D. Landau "systems Management") (2002). After receiving a rational transfer function should be checked identification to ensure that the resulting audio model is correct. There are various means of verification, the relevant put forward hypotheses about noise, influencing model (for example, the testimony of error of the forecast). To improve the reliability of the obtained model have the additional option to verify the obtained models via comparison between model simulations on the model and the real system, subjected (comparison of the amplitude and phase of the signal in the frequency range corresponding to the considered range for the suppression of the rebellion.

Preferably, such operation of identification with the excitation was performed for all configurations fill the passenger compartment of the real model. Such seeding can conform to the provisions occupied by passengers, devices (for example, additional seats), change of acoustic or electronic material, or any other conditions, is responsible for electro-acoustic characteristic of the passenger compartment. Therefore, it is desirable to identify all the configurations fill the passenger compartment, because multiple models received, in fact, have the divergences of the coefficient of amplification and phase for each frequency.

Now, after the receipt of the transfer function of electroacoustic model and after checking it with the appropriate tools will be synthesized control law to suppress disturbances variable frequency.

Describe the level of suppression of acoustic perturbations, which operates at the passenger compartment, is given by means of the direct function of the sensitivity of a closed system, which function is referred to as the Syp.

Suppose that the control law refers to the RST type, that is, the law, composed of three blocks, with T=0, a, R, S represent such polynomials that:

R(q -1 )=0 r +r 1 q -1 +...r nr ·q-nr

S(q -1 )=1+s-1 ·q -1 +...s ns ·q-ns

Control law is written in the following way:

RST-the regulator is a more General form of the introduction of a regulator for SISO, one input, one output, one variable. A closed system can in this case be as a structural scheme shown in figure 5, in which

represents the transfer function of the above-acoustical model. This structural pattern p(t) is the equivalent of acoustic perturbations, which was transferred to the output of the system, without loss of generality presentation.

Direct function (Syp) sensitivity can be defined as the transfer function between the signal (p(t)) perturbation and signal (y(t)) microphone. This transfer function describes the response of the closed path in relation to the suppression of acoustic disturbance.

In particular obtaining this function enables us to learn for any frequency quality suppression of the rebellion.

One can show that this function is written in the following way:

Since the purpose of the law was to make it possible for the suppression of the disturbances on the frequency fpert, the module Syp should be low on the frequency, in practice much lower than 0 dB.

Ideally, it would be desirable to Syp was lowest possible at all frequencies. However, this goal is unattainable due to theorem Bode-district of Freudenberg-Luz (Bode-Freudenberg-Looze), which shows that, if the system is asymptotically stable closed-loop is also resistant open loop:

This means that the sum of squares between the curve of the module sensitivity and the axis of 0 dB taken from them by a sign, equals zero. This implies that the weakening of perturbations in a certain frequency domain is inevitably leading to increased disturbance in other frequency regions.

Thereafter, as selected these poles, P is expressed and solved equation (2), an equation zomb (Bezout). Details of how the equation is solved zomb (Bezout), can be found, for example, in the above-mentioned work I.D.Landau (I.D. Landau) on pages 151 and 152. This is done by solving a system of Sylvester (Sylvester). In addition, this work associated computational procedures relevant programmes of the software Matlab® and Scilab®intended for the solutions of this equation. The choice of the poles can be made in accordance with different strategies. One of these strategies will be explained later.

Eliminating the influence of perturbations of the p(t) at the output is obtained on the frequencies:

Hence, for calculation of the regulator, the overwhelming indignation at the frequency fpert, part of S is determined a priori by the adoption of equation (2) what's decomposes on multipliers with the release of Hs-polynomial second order, for perturbations, i.e.:

h 1 =-2cos(2π.fpert/Fe)

if h 2 =1, we introduced a pair of complex zeros, without attenuation at the frequency of fpert.

If h 2 PD 1, S can be entered with a pair of complex zeros, if zero attenuation, the attenuation is determined as a function of the desired attenuation at a certain frequency.

In this case, the equation zomb (Bezout), the subject of the decision is:

In practice, the frequency noise, subject to repression, changes over time as a function, in particular, the speed of rotation of a cranked shaft of the vehicle, unit Hs must also change as a function of the mentioned frequency. In this case, this has the result that, for each frequency, subject to suppression, should be decided equation zomb (Bezout), which has the following form:

You notice that the solution of this equation, in particular, in real time, would have led to a large volume of calculations. Additionally, if you change the frequency of all coefficients S and R regulator should be changed. This results in a very heavy algorithm that requires significant computing power. Thus, even if the apply this simple solution with RS-regulator, preferably implement another solution, which is devoid of this problem and that minimizes the number of coefficients of the control law, changing the frequency of the perturbation subject to suppression.

Therefore, to address this problem, the following is a proposed solution based on the concept of parameterization of ula coachmen (Youla-Kucera) to controller RS-type.

Such a SISO system, one input, one output, one variable, adjustable regulator RS-type to which is attached a parameter of July, in schematically shown in Fig.7.

This kind of controller is based on the so-called "Central" RS-regulator, made of blocks Ro(q -1 ) and So(q -1 ), where Ro and So are polynomials q -1 .

Option ula is a block

where beta and α are polynomials q -1 .

As shown above, the units q-d B(q-1 ) and A(q -1 ) represent the numerator and denominator of the transfer function of the electroacoustic system, subject to regulation.

We can show that a link-the regulator made a way and shown in Fig.7, equivalent to controller RS-type, which blocks of R and S are equal:

Now, suppose that the Central controller formed and that it stabilizes the system.

Without parameterization of July, the characteristic polynomial (Po) system, as shown above, is written as follows:

When granting the Central controller parameter whirligig the characteristic polynomial of the system is written as follows:

P(q -1 )=A(q -1 ).(So(q -1 ).α(q -1 )-q-d B(q -1 ).β(q -1 )+q-d B(q -1 ).(Ro(q -1 ).α(q -1 )+A(q -1 ).β(q -1 ))

P(q -1 )=Po(q -1 ).α(q -1 )

you can see that the poles Q (zeros a) coexist with the poles of the closed-loop equipped with only the Central regulator, the characteristic polynomial of which is Po.

In addition, the equation:

can be used to determine the unit's through pre-defined block Hs, that is:

S'(q -1 ).Hs(q-1 )=So(q -1 ).α(q -1 )-q-d B(q -1 )b(q -1 )

that also is the equation of zomb (Bezout)that enables one, in particular, to find?, if defined α and Hs.

Let Sypo will be a direct function of the sensitivity of a closed system with a Central controller, without the whirligig.

Direct function of the sensitivity of a closed system with the regulator, provided with a parameter of ula is written as follows:

Consequently, based on a closed system containing a Central regulator, not having the purpose to suppress the sinusoidal disturbance, in particular at fpert, to the Central controller can be attached parameter of July, that will change the function (Syp) sensitivity while maintaining the poles of the closed-loop, equipped with a Central controller, to which are added the poles Q. In this case, in Syp frequency fpert can be created failure in the spectrum.

For this purpose the Hs, and C is calculated so that the transfer function of the

obtained in the result of the discretization of the continuous block of second-order method Tustin with "":

Hs, and C represent the polynomials q -1 the second degree, and u 1 , u 2 are coefficients of attenuation of the transfer function of the second order.

In addition, the operation discretization of continuous transfer function (in c) can be accomplished through a computational procedures that can be found, for example, in the programs of computational software specifically intended for the equipment of automatic regulation and control. In the case of Matlab® this is the function of "c2d".

It can be shown that the weakening M at frequency fpert is given as follows:

Besides, it is necessary to u 1 <1.

In addition, for equal treatment

, it is shown that the failure in the spectrum of functions (Syp) sensitivity is the wider, the more u 2 . But, the more wide is the gap in the spectrum, the more deformed |Syp| at frequencies other than fpert (a consequence of theorem Bode-district of Freudenberg-Luz (Bode-Freudenberg-Looze)). Therefore, when you select a u 1 , u 2 a compromise determined so as to create a sufficiently broad weakening around fpert without causing the ingot of a significant increase |Syp| at other frequencies. Typical values of attenuation coefficients are: u 1 =0,01 u 2 =0,1. These values can form the starting point for optimization.

After that, solving the equation zomb (Bezout) (10), one can calculate?.

It is shown that this choice Hs, and C produces a dip in the spectrum of functions (Syp) sensitivity, causing almost negligible impact in relation to Sypo on other frequencies, even if the applicable theorem Bode-district of Freudenberg-Luz (Bode-Freudenberg-Looze), which causes an increase in module Syp towards Sypo at frequencies other than fpert.

This increase Syp can reduce the robustness of a closed loop, which can be measured margin module (distance to the point of -1 from the provisions of the frequency open-loop system adjusted in the plane Nyquist) equal to the inverse value of the maximum [Syp| in the frequency range [0; Fe/2].

The main advantage of using parameterization of ula consists in the fact that a is of order 2:

Moreover,? is of the order of 1:

Therefore, in the case of the proposed system, consisting of the controller's RS-type to which is attached a parameter of July, the number of parameters, changing as a function of frequency disturbing noise, subject to repression, in the law of management is only 4. Calculation of these parameters as a function of frequency (f) of perturbation subject to suppression, can be made in advance, independently, by means of the solution of the equation zomb (Bezout) (10), during the design stage of the control law, and these settings can be stored in tables situated inside the car programmable computing device, and called, in real time, as a function of frequency, subject to suppression.

On Fig.8 shows the complete diagram of the control law (Central RS-controller + parameter (Q) ula).

To perform the synthesis of the regulator, it is preferable to use electro-acoustic model that can be qualified as a median, that is, the model, corresponding to the intermediate level of the passenger compartment from a number of electro-acoustic models corresponding to different configurations fill the passenger compartment.

It is preferable that the purpose of the synthesis of the Central regulator was to ensure that the stocks, having a special goal of suppression of the rebellion. This can be achieved, for example, by means of the method of placement of the poles, and, if necessary, you can refer to the above-mentioned work I.D.Landau (..), in particular, to the whole Chapter 3. More precisely, it may be performed so, as explained below.

Decide to run placing the poles of the closed-loop by placing n dominant poles closed contour on the n poles of the system, subject to regulation, that is, the roots of A(q -1 ), while n is the degree of the polynomial A. The preliminary determination of the unit So no, for the reason that there is no purpose to suppress the indignation by only one Central controller. During this operation, the Central regulator does not suppress disturbances p(t), but provides maximum robustness.

Can also be placed in a number of optional "high-frequency" of the poles, the value of which is between 0.05 and 0.5 on the complex plane (in the case in which there is no excess sample). Should take into account the fact that system is stable if all its poles are contained strictly within the unit circle in the complex plane. These utility poles play the role to improve the robustness of the control law when joining a parameter of ula.

After the poles of the closed-loop, that is, the roots of the Po(q -1 ) were selected, expressed Po(q -1 ), which is a polynomial q -1 degree n+m. After that, using the aforementioned procedures equation is solved zomb (Bezout):

where S o and S' o are unknown.

The Central controller, therefore, is defined and calculated.

Then the index is calculated parameter (Q) ula (that is, a and b), which are the only polynomials control law, varying as a function of frequency perturbations subject to suppression.

For each frequency (fpert) perturbation subject to repression, coefficients (u 1 , u 2 ) attenuation equations (12) are selected to adjust the depth of the weakening of Syp on the frequency, as well as the width of the failure in the spectrum (bandwidth) at fpert in Syp, leaving sufficient robustness that can be measured in the above margin module (maximum of Syp). Targets can be set, for example, stock modulo of 0.7, which corresponds to a high level of robustness of a closed loop, robustness, which will ensure the sustainability of active control variations in passenger cabin configuration.

It is known that the regulation of closed-loop is even more robust than nearer the poles of the closed path to the system, subject to regulation. Such a condition completely satisfied with this choice of accommodation pole during the synthesis of the Central controller.

Polynomials Hs(q-1 ) and C(q-1 ) are calculated as explained above, through sampling transfer function of the second order, and the equation zomb (Bezout) (10) is designed to determine b(q -1 ).

It is preferable that this calculation provides a definition alpha(q -1 ) and? (q -1 ) as functions fpert, was carried out throughout the frequency range in which you want to be suppressed indignation, a and beta can, for example, calculated for the frequency, changing with the increment of 2 Hz in the range, which is between 30 and 120 Hz.

In addition to our own electro-acoustic model (electro-acoustic models and model of the Central RS-controller, all the coefficients of the polynomials alpha(q -1 ) and? (q -1 ) as a function of fpert stored in memory table for these factors in computing device. Table allow to find the data to be used in real time, as a function of current conditions, in particular the current frequency noise, subject to weaken, and perhaps the current configuration of the filling in of the passenger compartment.

Therefore, the control law (RS-controller + option ula) in this case synthesized. At the secondary stage of the design stage, you can verify that it has the stability and an appropriate level of robustness (stock modulo >0.5), the modelling of a closed system and the suppression of the disturbances on the entire frequency range, for all configurations fill the passenger compartment, using electro-acoustic models identified in a variety of configurations. If this does not occur, designed control law change, affecting the coefficients u 1 , u 2 (depth and width of the frequency band suppression). If it's still not enough, you can try to take as electroacoustic model is a different model from among the models obtained for different configurations of the passenger compartment, or to influence the placement of additional poles of the closed circuit of high-frequency poles).

These preliminary stages of the design and synthesis requires significant computing, so it is preferable that they were performed in batch mode. As only a synthesis is made, these models could be applied in real-time in a computing device to achieve dampen the noise in the passenger compartment.

During the functioning of a computing device, in real time, as shown in Fig.8, the in-memory data, in particular the coefficients of the polynomials alpha(q -1 ) and? (q -1 ) for the parameter of July, is called as a function of the information on the current frequency noise, subject to a weakening of the coming, for example in an indirect way of measuring tachometer on the crankshaft. For current frequency values, which do not correspond directly to the frequencies of the input data of the table (the current frequency is located between two frequencies to the calculation of the table)may be evaluated with polynomial coefficients alpha(q -1 ) and? (q -1 ) through the implementation of interpolation between the calculated coefficients for two or more known values of the frequency. In the latter case, it is preferable that the channel step increments between the frequencies used for the calculation of the coefficients, not too large, usually appropriate is the grid of frequencies with a step of 2 Hz.

Summarizing the above example you can see that the invention relates to the executable in the real-time active way to weaken, through feedback, narrowband noise essentially , on at least one specific frequency in the passenger compartment of the vehicle by emitting the sound through at least one Converter, usually loudspeaker, managed signal u(t), generated programmable computing device as a function of signal (y(t)of acoustic measurements performed at least one acoustic sensor, usually microphone, while at the first stage of designing the electroacoustic characteristics of the link formed by the passenger cabin, the transducer and sensor is simulated by the electro-acoustic model as electro-acoustic transfer function, which is defined and calculated, then defined and calculated control law based on the global model of the system in which this control law applies to electroacoustic transfer function whose output optionally accepts a signal (p(t)) noise, giving the signal y(t) at the design stage, with the above mentioned law, management, makes it possible to produce the signal u(t) as a function of acoustic measurements y(t), and the second stage - using referred calculated control law is used in computing device in order to produce the signal u(t), then sent Converter, as a function of the signal y(t), adopted by the sensor, to weaken referred noise.

To put it more specifically, at the design stage:

a) - in a first phase as electro-acoustic model used discrete rational electro-acoustic transfer function, and the said audio model is defined and calculated by the acoustic excitation of the passenger compartment through converters and acoustic measurements sensor, using then the process of identification of linear systems using these measurements and transfer function model,

b) on the second step is to implement the control law, which contains the so-called "Central" RS-regulator, consisting of two blocks

and Ro(q -1 ), and in the Central controller unit

generates a signal u(t) and takes as input inverted output signal block Ro(q -1 )referred to block Ro(q -1 ) takes as its input signal y(t)corresponding to the amount of noise p(t) and output signal electroacoustic transfer function of electroacoustic model, and the Central controller is defined and calculated,

c) the third stage in the control law introduces the option of July in the form of block Q(q -1 )attached to the Central RS-regulator, and referred to block Q(q -1 )) whirligig, takes noise estimation obtained by calculation of signal u(t) and y(t) as a function of electroacoustic transfer functions, and the output of the said unit (Q(q -1 )) ula is deducted from the inverted signal Ro(q -1 ) sent to the entrance of the block

Central RS-regulator, and a parameter (Q(q -1 )) ula in the law of management, comprising a Central regulator is associated with a parameter of July, determined and calculated for at least one frequency (p(t)) noise, including at least this specific frequency noise, subject to impairment

and point-of-use, real-time:

- define the current frequency noise, subject to impairment

make your computing device count the law office containing RS-controller with a parameter of July, using as this option is selected, which was designed for a specific frequency appropriate to the current frequency noise to be weakening.

To date was presents a simple implementation, in which the passenger cabin, equipped with only a microphone and one speaker or group of speakers, all of which are excited by the same signal.

But the fact is that the noise reduction/a silence that can be obtained through the active management process, are very localized in space. In the aforementioned article, "A review of the active noise and vibration control in road vehicles" ("Overview of active protection against noise and vibration in road vehicles", Eliott (Elliott) indicates that the zone of silence around the microphone, defines the mismatch that does not exceed one-tenth of the wavelength noise, subject to repression, that is, approximately 110 cm for noise with frequency 30 Hz 55 cm for noise frequency of 60 Hz, 28 see for noise frequency of 120 Hz, at ambient temperature.

Thus, you can see that it is impossible to obtain a uniform noise reduction if there is only one microphone in the passenger cabin quite spacious vehicle and that multiply the number of microphones, defining mismatch, and distribute them in the passenger cabin thus, to increase the space in which there is noise reduction.

By way of example, figure 9 shows the scheme of electroacoustic transfer system 2*2 (2 loudspeaker, 2 microphones). In this example, microphone 1 sensitive to acoustic effects of the loudspeaker 1 (HP1) and the loudspeaker 2 (2). In addition, microphone 2 sensitive to acoustic effects of loudspeaker, 2 (2) and the loudspeaker 1 (HP1). Such a system, driven by way of example, can be modeled with the following transfer function matrix:

or in the case of (2*2):

View MIMO-system, multiple inputs, a lot of exits variables, through a transfer function, actually, not very comfortable, and is the preferred representation of the state, which is a generic representation of the linear systems (whether they are MIMO range of inputs, a lot of exits variables, or not).

nu is the number of inputs of the system (number of speakers or groups of speakers connected to each other);

ny is the number of outputs of the system (number of microphones);

n is the order of the system.

In the following description, to simplify the explanation, it is believed that nu=ny, but this is not a limitation, and the following description may also be applied to the case of nu>ny.

Representation of the state of electroacoustic system (passenger compartment) can be written as recurrent equation, called the "equation of state":

X: the vector of the system, has dimension (n*1)

U: the vector of input signals of the system that has dimension (nu*1)

Y: the vector of output signals, has dimension (ny*1)

G: the matrix referred to as "the matrix evolution", has the dimension (n*n)

N: the input matrix of the system, has the dimension (n*nu)

W: output matrix of the system, has the dimension (ny*n).

The coefficients of the matrix G, H, W define the linear system MIMO range of inputs, a lot of exits variables.

We assume that X(t) corresponds to the vector X in time (t) and X(t+Te) corresponds to the vector X in time (t+Te) time (that is, through sampling period after X(t)).

Control law based on this view, the state, and therefore, in the case of SISO, one input, one output, one variable must be defined model of the electroacoustic system, subject to regulation, (electro-acoustic model of the passenger compartment), that is, the coefficients of the matrix G, H, W.

Fig 10 shows a block diagram of the electro-acoustic model of the passenger compartment in case of MIMO range of inputs, a lot of exits variables, which I corresponds to the identity matrix, and which corresponds to formula (18). By analogy with the case of SISO, one input, one output, one variable, P(t) is a vector of disturbances on outputs, that is:

expressed in p 1 p ...ny , where p i is the outrage i signal.

In the case of SISO, one input, one output, one variable, the coefficients of the model of electroacoustic system, subject to regulation, are obtained through identification procedures during the design stage, that is, by instituting real electroacoustic system noises, having essentially uniform spectral density, nu speakers are initiated by means of signals that in relation to each other.

After this, the input data (measurement MICS) and outputs (signals for the speakers) are stored in memory in the computer and processed in it so as to obtain a representation of the state of the entire system, using this time authentication algorithms that are specifically designed for MIMO systems, multiple inputs, a lot of exits variables. These algorithms, for example, are presented in the Toolkit software, specialized in engineering, automatic control, such as, for example, Matlab®. It is also useful to appeal to working L.LJUNG, "System identification Theory for the user (. "Theory of identification systems for users"). Prentice Hall, Englewood Cliffs, N.S, 1987, the algorithms presented in this paper give rise tools, specialized on the identification, program software Matlab®. To the same algorithms for checking the correctness of the model produced for electro-acoustic system, subject to regulation.

Another possible implementation is an authentication nu*ny transfer functions, one after another, through identification tools for the case of SISO, one input, one output, one variable, and by instituting one speaker after another, and after this merger nu*ny models in a single model for MIMO range of inputs, a lot of exits variables. Such an Association could be done, for example, through the innovative method of minimum standard error of the algorithm described in the work of the Ph. de Larminat: "Automatic apphquée", Hermes, 2007

As in the case of SISO, one input, one output, one variable, it is desirable to identify for each of the configurations of the passenger compartment and to take as a model of electroacoustic system, which remain for the next part of the design stage of a model that could qualify as "median".

As soon as the I / o model electroacoustic system was received represent the state, and this model has been verified, then can be defined and calculated control law. Now, therefore, should be synthesized control law, which allows suppressing each of microphones, acoustic disturbance with frequency fpert, the mentioned frequency fpert may change over time.

With this purpose the concept of a Central regulator and the concept of a parameter of July, used in case of SISO, one input, one output, one variable are generalized for the case of MIMO range of inputs, a lot of exits variables.

It is believed that electroacoustic system described representation (18) state. One can show that the Central controller in case of MIMO range of inputs, a lot of exits variables has the form: "the observer status + feedback is evaluated as"in the form of:

is a vector of the status of "observer", has dimension(n*1)

Kf is a strengthening of the "observer" dimension (n*ny).

and control law is written as follows:

where Kc is a vector of feedback estimated the state of the system has dimension (nu*n).

It may be useful to refer to the work "Robustesse and commande optimale" (Alazard and others, publishing house CEPADUES, 1999, page 224 and 225).

In accordance with these formulas, figure 11 shows a block diagram of a Central regulator, the fig.12 shows a block diagram of the Central regulator applied to electroacoustic model of the passenger compartment, still in the case of a MIMO system, multiple inputs, a lot of exits variables. This latter structure is a traditional structure in the technique of automatic control. In accordance with the rule, known as the "separation principle", the poles of the closed-loop formed their own values, G-Kf.W eigenvalues of G-H.Kc, that is:

eig(G-Kf.W)∪eig(G-.).

eig(G-Kf.W) is the assigned pole filter, and

eig(G-H.Kc) are the poles of control.

the eig() refers to your own values.

Therefore, placing the poles of the closed-loop, equipped with a Central controller, can be done by selecting coefficients Kf and Kc, which are the settings of this management structure. Number of poles, to be placed is 2*n.

Thus, the observer estimated recruitment and feedback as selected as a Central regulator. In the case of SISO, one input, one output, one variable, it is shown that if n poles of the closed-loop placed on the n poles of the electroacoustic system (i.e. on the roots of the polynomials A(q -1 ), then the Central controller, which is not specifically suppresses the disturbances, but has a maximum robustness.

In the case of MIMO range of inputs, a lot of exits variables, the objective is also to Central regulator had maximum robustness, not having a special purpose suppression of the rebellion. Therefore, poles filtering selected equal to the poles of the system, subject to regulation. Thus, the need to Kf.W=0.

The most trivial solution is:

Therefore, the equation of the Central regulator takes the form:

Remains a place other n poles poles (eig(G-.)) regulation). Following what was done for the regulator in the case of SISO, one input, one output, one variable, these poles will be selected as a set of high-frequency poles designed to ensure the robustness of the control law. It should be noted that the situation MIMO range of inputs, a lot of exits variables, the number of coefficients Kc (nu*n) is greater than the number of poles (n), which can accommodate, and thus, these degrees of freedom can be successfully used to perform system of eigenvectors (choice not only of their own values, but also eigenvectors (G-H·Kc)).

Another way to start the calculation of the COP is to perform a linear-quadratic optimization (LQ-optimization)in respect of which there is an extensive literature. You can, for example, mention the work of "Robustesse and commande optimale", publishing house CEPADUES, 1999, page 69-79. Also, there is the possibility to calculate the coefficients of the matrix (COP) to perform what Ph. de Larminat calls LQ-optimization-type, i.e. optimization based on the horizon Tc. The details of this LQ-optimization-type can be found in the work of the Ph. de Larminat: "Automatic appliquee", Hermes, 2007, In particular, with the work related calculation procedure for the software Matlab®intended for calculation of coefficients Kc, relevant LQ-optimization-type.

In the law of management, as it is symbolically shown in fig.13 parameter (Q) ula is itself a block MIMO range of inputs, a lot of exits variables, whose representation of the state can be written in the following way:

where X Q is a vector parameter status of ula.

In this case the law Department of the Central controller, fitted with a parameter of ula is written in the following form:

moreover, this control law corresponds feedback as an observer, associated with feedback as a parameter of ula.

Now you will learn how to determine the block parameters (Q) so as to ensure the suppression of disturbances known frequency.

In the case of SISO, one input, one output, one variable, transfer function

has been calculated by the discretization of the continuous transfer function of the second order, and C was then denominator parameter of July, and the Hs is used in equation(zomb (Bezout), allowing to find?, the numerator of this ratio whirligig.

In the case of MIMO range of inputs, a lot of exits variables, for each output-i use the model of unregulated ones:

For each output i this model unregulated perturbation is written as follows:

is a state vector of the model perturbations i (dimension 2*1)

Z 2i is an additive perturbation output i (dimension 1*1), whereas:

It should be noted that the choice of the form G 2i ; W 2i is not unique. It was accepted that the canonical representation of the monitoring capabilities.

hS 1i and hS 2i subtracted from the numerator of the transfer function

resulting from the discretization of the continuous transfer function of the second order, identical to that used in the case of SISO, one input, one output, one variable:

when this:

Hs i (q-1 )=h 0 +h 1i ·q -1 +h 2i ·q -2

Discretization of continuous transfer function can be performed, for example, by means of computational procedures "c2d" program software Matlab®.

In this case, the equation of state observer, increased by models perturbations on outputs, can be written in the following way:

when this:

Kf 1 has the dimension (2*ny, ny)

Kc 1 has the dimension (nu, 2*ny)

This vector, which is a state vector unregulated model.

Equation (29) the observer can also be written in the following way:

Now you should choose the coefficients Kf 2 to place the poles of the portion of the enlarged observer.

Choice for poles 2ny roots denominators α i (q-1 )that is made in case of SISO, one input, one output, one variable is generalized to the case of MIMO range of inputs, a lot of exits variables.

To put it more accurately, eig(G 2i-Kf 2i. 2i ) is equal to the roots of the above polynomials α i (q -1 ), which the polynomials are, as described above, the result of a discretization of the continuous transfer function of the second order.

Calculation Kf 2i as a function G 2i , W 2i, and C i (q -1 ) is a traditional operation accommodation poles. To perform this operation you can, for example, use the procedure in Matlab®specifically designed for this operation, the name of which is "the PLACE" ("PLACE").

The latter condition matrix Kf 2i is the diagonal blocks, that is:

Remains choose Kc 2 dimension (nu*2ny). This choice is not free if you want to get the asymptotic suppression output disturbances.

Kc 2 , must satisfy the so-called equations "asymptotic of suppression, which have the following form:

T a ·G 2 G·T a-H·G a =0

Explanation of equations (36) and (37) can be found in the work of the work of the Ph. de Larminat: "Automatic appliquée", Hermes, 2007, page 202 - 205. Solution of equations (37) as a result leads to solving a system of Sylvester. It should be noted that the above-mentioned work is available computational procedure to program the software Matlab® for asymptotic solutions of the equations of suppression.

Comparing equations (24) and (25) with equation (34) one can notice that this structure with increased state observer is not that other, as a Central regulator, in the form in which it was defined, equipped with a parameter of July, but where signs of equations (24) and (25):

A Q =G 2-Kf 2 ·2 W

B Q =Kf 2

It should be noted that these equations are fair, because it was made Kf=0.

Therefore, for each frequency perturbations of the coefficients A Q B Q C Q can be calculated during configuration management act and stored in tables, so called point-of-use, as a function of fpert, in a computing device, running in real time. On figure 14 shows the scheme of application of the law management point-of-use, real-time, programmable computing device.

Block (Q) ula can be realized as the transition matrix in such a way as to minimize the number of variables coefficients in this block. Such operation can be performed, for example, by means of the procedure "ss2tf" in Matlab®.

As shown above, the setting of parameters of the control law lies in the choice of poles management (carried out by means of parameters Kc), which impact on the robustness of the control law. For each frequency, you can choose attenuation coefficients u 1i , u 2i continuous transfer functions of the second order, that influence the frequency band width and depth of the suppression of the disturbances on the frequency fpert.

It should be noted that the robustness of the control of a closed system can be estimated by calculating infinite norms of the transition matrix between P(t) and Y(t) (compilation of case SISO, one input, one output, one variable). Since the calculation of infinite norms of the transition matrix is performed by calculation of singular values of the mentioned transition matrix, here again it is possible to use program software Matlab® and, in particular, the function SIGMA (SIGMA) of the "management instrumentation".

These opportunities settings are a generalization of the adjustment possibilities case SISO, one input, one output, one variable.

In summary, we note that the control law (the Central controller + option ula), intended for application to the electro-acoustic model of the passenger compartment of the vehicle, in the case of MIMO range of inputs, a lot of exits variables, is obtained through the following stages:

- obtaining of electroacoustic model of a passenger vehicle, which is a linear MIMO system, multiple inputs, a lot of exits variables, in the form of the state representation, calculated through identification,

- synthesis of a Central regulator in the form of observer status and feedback estimated as, Kf is selected as Kf=0,

- selection of coefficients Kc, which correspond to the high-frequency poles, to ensure the robustness of the control law (LQ-optimization) and, in particular LQ-optimization-type),

- choice of attenuation coefficient u 1i , u 2i for the grid frequency perturbations, subject to repression, such a grid is, in particular, in the case in which over time can meet several current frequency noise to be weakening, or in the case when the frequency noise varies over time (as in the case of SISO, one input, one output, one variable, point-of-use can be made interpolation of variables as a function of frequency),

- calculation of coefficients of a parameter of July, which is stored in tables computing device to be used in real-time point-of-use.

Please note that the reduction in the number of factors that are placed in the table, can be done by selecting all ?C; equal to the frequency of the perturbation.

Therefore, the invention implemented the Central controller with a parameter of July, which has the form of a filter infinite impulse response, with at least one entrance and at least one output, the number of which is a function of selected forms of implementation (SISO, one input, one output, one variable, MIMO range of inputs, a lot of exits variables, the number of sensors and transducers,...).

In the above cited as an example embodiments of the invention in order to simplify the explanation was the case of the suppression of one single frequency. However, the invention can be applied to suppression of several frequencies at the same time, and this case will be described now.

And really, regardless of whether there is a case SISO, one input, one output, one variable or case MIMO range of inputs, a lot of exits variables, you can suppress simultaneously more than one frequency. This leads to the introduction of a second or even a third failure in the spectrum of functions (Syp) sensitivity. However, taking into account theorem theorem Bode-district of Freudenberg-Luz (Bode-Freudenberg-Looze), it follows that the creation of one or more of the gaps in the spectrum of the sensitivity functions necessarily causes an increase |Syp| at other frequencies, and hence reduction of robustness.

In the following description, it assumes that you are suppressing the two bands, but in any case this is not a restriction and is given only as an example. These two frequencies are:

- current frequency, which is designated here as fpert, if you use the designations used so far in this document and

the second frequency associated with fpert, and that is designated as η.fpert, and η not necessarily represent integer, t] can represent a constant, not necessarily being an integer, but it can also be a fpert, the only condition is that the function η.(fpert) should be continuous.

In the case of SISO, one input, one output, one variable, is still available equation (10) zomb (Bezout):

S'(q -1 ).Hs(q -1 )+q-d B(q -1 )b(q-1 )=So(q -1 ).α(q -1 )

Here Hs, and C represent the polynomials q -1 4-th degree, and u 11 , u 12 , u 21 , u 22 are coefficients of attenuation, allowing, as in the case of suppression, specify the width and depth of the failure of the weakening of the curve representing the module Syp, a(q -1 ) is a polynomial of 4th order, and b(q -1 ) is a polynomial 3rd order. Thus, the number of variable coefficients in the law of management is higher: there are 4 additional coefficient varied as a function of fpert.

In the case of MIMO range of inputs, a lot of exits variables, the matrix G 2i of the equation (27) is now a dimension of 4*4, that is:

and, also

It should be noted that the choice of the form G 2i W 2i is not unique. It was accepted that the canonical representation of observability.

the hS 1i hS 2i , hS 3i hS 4i are coefficients of the numerator of the transfer function

the resulting sample rate of the product of two continuous transfer functions of the second order, identical to those used in the case of SISO, one input, one output, one variable, that is:

H si (q-1 )=h 0 +h 1i ·q -1 +h 2i (q -1 )+h 3i (q -1 )+h 4i (q -1 )

From this it follows that now:

Kf 2 has the dimension (4*ny, ny)

Kc 2 has the dimension (nu, 4*ny),

the G2 is given by equation (31), but has the dimension (4ny*4ny).

this time is a dimension (4ny*1), and matrix W2 this time has the dimension (ny*4ny). Equations (36) and (37) asymptotic suppression remain unchanged. The solution of such system MIMO range of inputs, a lot of exits variables, similar to the above occasion of the suppression of a single frequency.

What was described for the number of simultaneously jammed frequency, equal to two, and can be extended to a high number of frequencies. However, as explained above, the increase in the number of the repressed frequencies leads to the loss of robustness, quickly becoming unacceptable.

It should be understood that the principle of the invention: the Central controller is attached to a parameter of July, can be applied in practice to dampen the noise are different than what was described above. In particular, the other can be a type of electroacoustic models, methods of determining and/or synthesis of the Central regulator and parameter ula may also be different, and for the practical implementation of 43 of these other methods may be helpful to refer to the specified literature.

1. The method of active, real-time weakening narrowband noise, essentially , of at least one specific frequency, in the passenger cabin of the vehicle by means of feedback, including the emission of a sound at least one Converter, usually loudspeaker, managed signal u(t) or U(t) in the case of SISO or MIMO, respectively, and the signal is generated programmable computing device as a function of the signal acoustic measurements y(t) or Y(t)corresponding to the specified occasion, acoustic measurements performed by means of at least one acoustic sensor, usually microphone, use one sensor corresponds to the case of SISO, a single input - single output - one variable, and using multiple sensors corresponds to the case MIMO range of inputs, a lot of exits - a set of variables, with the first stage - the stage of design - model electroacoustic characteristic link formed by the passenger cabin, the transducer and sensor by means of electro-acoustic model in the form of electro-acoustic transfer function which is defined and calculated, and then determine and calculate the control law based on the global model of the system in which the specified control law apply to electroacoustic transfer function, the output of which is additionally supplied subject to a weakening of the signal noise p(t), so that at this stage of design to get the signal y(t) or Y(t), while the control law allows forming of the signal u(t) or U(t) as a function of acoustic measurements y(t) or Y(t), and the second stage - point-of-use - for easing referred noise use the calculated control law in computing device for signal u(t) or U(t)sent then the Converter, as a function of the signal y(t) or Y(t)obtained from the sensor is characterized in that implement control law, which includes the whirligig to the Central controller and that is that this law only control parameter ula has coefficients that depend on the frequency subject to a weakening of the noise, with a Central regulator has constant coefficients, the option of ula has a view of the filter infinite impulse the characteristic, after the definition and calculation of control law in memory of computing devices preserve at least the specified variables coefficients, preferably in the table, in the form of functions of a certain frequency or certain frequency noise p(t)used at the design stage, the point-of-use, real-time: get the current frequency noise, subject to impairment provide with the help of a computing device calculation of the control law, including the Central controller with a parameter of July, when this is used as a parameter whirligig stored in the memory of the coefficients of a certain frequency corresponding to the current frequency noise to be weakening.

5. The method of claim 2, or 3, wherein the stage of (b) polynomials Ro(q -1 ) and So(q -1 ) Central proportionate and expect by way of placement of the poles of the closed loop, n dominant poles closed circuit equipped with a Central controller is chosen equal to the n pole of electroacoustic transfer function and m auxiliary poles poles are located at high frequency.

6. The method according to claim 1, characterized in that at the design stage: (a) at the first stage uses linear electro-acoustic model, the electro-acoustic model of representation of matrix blocks N, W, G, q -1 and I, where G is the transition matrix, N - input matrix, W - output matrix and I is the identity matrix, the representation of the state can be expressed by a recurrent equation X(t+Te)=G·X(t)+H·U(t)Y(t)=W·X(t), where X(t) is the state vector, U(t) is the vector of input signals; Y(t) is a vector of output signals, and electro-acoustic model determined and calculated by the acoustic excitation of the passenger compartment through converters and acoustic measurements sensors, using then the process of identification of linear systems running with the specified dimensions and specified model, b) - the second stage implement the Central controller applied to the specified defined and calculated by the model, and the Central regulator has the appearance of an "observer" status and feedback on the estimated state, which iteratively expresses

, a state vector of the "observer", as a function Kf, gain factor "observer", Kc, vector feedback estimated as, as previously defined and calculated electroacoustic model, there are:

where the control action

and define and calculate the specified Central regulator, c) - the third stage to the Central regulator added option for the formation of ula control law, while the parameter of ula has a view of the block Q for the case of MIMO range of inputs, a lot of exits - a lot of variables, consisting of matrices AQ, BQ, CQ States, attached to a Central controller, also resulting in the form of the state representation; block Q, the output of which is built with the output signal of the Central regulator, forms the signal, which generates a signal, the opposite U(t), and at the entrance of which signal Y(t), which is deducted signal

; and the option of July in the law of management, comprising a Central regulator is associated with a parameter of July, determined and calculated by at least one noise frequency p(t), including at least the specified certain frequency noise, subject to a weakening; calculation of coefficients matrixes AQ, BQ, CQ is performed by obtaining the discrete transfer functions

resulting from the discretization of the continuous transfer functions of the second order, and by placing the poles, and also solutions of the equations asymptotic suppression, point-of-use, real-time: - get the current frequency noise, subject to impairment - provide with the help of a computing device calculation of the control law, including the Central controller with constant coefficients and parameter whirligig with variable coefficients, using as a parameter the whirligig ratios for noise frequency corresponding to the current frequency noise to be weakening.

7. The method of claim 6, wherein said at the design stage to perform the following operations: a) the first leg of the passenger cabin is subjected to acoustic stimulation, feeding on the converters excitation signals, spectral density, which is essentially uniform at the effective frequency band, and excitation signals are in relation to each other, b) - the second stage define and calculate the Central regulator so that it was equivalent to a controller with "observer" status and feedback calculated the state, through the placement of poles in applying Central regulator to electroacoustic transfer function, with the purpose of choosing a zero gain "observer", that is Kf=0, and the coefficient of the COP strengthening state feedback choose so, to ensure the robustness of the control law with parameter whirligig, through LQ-optimization, c) - the third stage when considering the view zoomed "observer" status is determined and calculated coefficients block Q whirligig in the law of management of at least one noise frequency P(t), including at least the specified certain frequency noise, subject to impairment in the form of test functions weakening thus, to get the parameter values of the coefficients of ula for the specified frequency or each frequency, point-of-use, real-time, perform the following operations: - provide with the help of a computing device calculation of the control law, the Central controller with constant coefficients and parameter whirligig with variable coefficients to obtain the signal U(t)sent to inverters, as a function of acoustic measurements Y(t), the use for parameter whirligig the values of the coefficients defined and calculated for a certain frequency corresponding to the current frequency.

8. A method according to claim 2, 3, 6 and 7, wherein the method is fit for a set of specific frequency noise, subject to impairment and stage (C) repeat for each of these specific frequencies, point-of-use in the case when none of the specified frequencies does not match the current frequency noise, subject to impairment produce interpolation on the specified current frequency values of the coefficients block Q whirligig on the basis of the values of the coefficients of the specified block Q whirligig, which are known to these specific frequencies.

9. A method according to claim 2, 3, 6 and 7, wherein the discretization of signals is made with the frequency Fe, and at the stage of (a) effective frequency band used for the excitation signal is essentially equal to [0, Fe/2].

10. A method according to claim 2, 3, 6 and 7, wherein the stage before use, a stage of designing add a fourth step (d), intended for checking the stability and robustness of the model of the electroacoustic system and control law, the Central regulator with a parameter of July, prior to the stages a)-c), by modeling the application of the law management obtained at stages b) and (c)applied to the electroacoustic model, obtained at the stage of (a), for a specified a certain frequency or certain frequency, and in case when the specified criterion of stability and/or robustness is not satisfied, produce repetition at least stage c) by changing the criterion of abating.

11. The method according to any one of claims 1 to 3, 6 and 7, wherein the design stage is a preliminary stage and it is carried out once previously in relation to the stage of use, preserving in memory of results of definition and calculation to use point-of-use.

12. The method according to any one of claims 1 to 3, 6 and 7, wherein the current frequency noise, subject to a weakening of the measure produced by the counter of revolutions of the motor vehicle.

13. The method according to any one of claims 1 to 3, 6 and 7, wherein the noise occurs on one particular frequency fpert.

14. The method according to any one of claims 1 to 3, 6 and 7, wherein the noise produced by the two specific frequencies: from the first frequency fpert and the second frequency η.fpert where η is constant or is continuously changing with fpert.