JP4857928B2 - Noise control device and noise control method - Google Patents

Description

  The present invention relates to a noise control device and a noise control method.

  In general, a noise control device and a noise control method have been proposed for measuring noise generated as a vehicle travels in a passenger compartment, for example, and generating a wave such as a sound wave or vibration that cancels the noise to reduce the noise. Yes.

In such a noise control device, a noise such as an acceleration sensor is affixed to a predetermined location without using a microphone for measuring noise, and noise is estimated from vibration detected by the sensor. Control devices have been proposed. For example, a noise control device has been proposed in which an acceleration sensor is attached to a point on the floor panel that has a high correlation with vehicle interior noise when the vehicle is running, noise in the vehicle interior is estimated from signals detected by the acceleration sensor, and noise control is performed. (See Patent Document 1).
Japanese Patent Laid-Open No. 8-2922771

  In such a noise control device, vehicle interior noise is estimated based on vibration detected by the sensor, and noise control is performed based on the estimated vehicle interior noise to reduce vehicle interior noise. If the vehicle breaks down, it becomes difficult to accurately estimate the vehicle interior noise, and the noise reduction effect is reduced.

  Therefore, in the case where a sensor breaks down, it is possible to cope with this problem by estimating the vehicle interior noise and performing noise control assuming that there is no broken sensor. However, as the number of failed sensors increases, it becomes difficult to accurately estimate vehicle interior noise. For this reason, there is a problem that even if noise control is performed, the noise reduction effect cannot be sufficiently obtained, and in some cases, the noise may be increased, and the noise reduction effect is impaired.

In order to solve the above-mentioned problem, the invention according to claim 1 is arranged on a vehicle body of a vehicle, and includes a plurality of sensors that detect vibrations of the vehicle body, a wave application unit that applies a controlled wave to the vehicle body, and the plurality of the plurality of sensors. A sensor failure detection unit that detects a failure of the sensor based on each output signal output from the sensor, and the vehicle based on the output signal of the normal sensor excluding the sensor where the failure is detected by the sensor failure detection unit A noise estimation unit that calculates an estimated value of vehicle interior noise that can be heard in a predetermined space of the vehicle interior, and when the sensor failure is detected by the sensor failure detection unit, the sensor failure is detected. A correlation value between the estimated value of the vehicle interior noise before detection and the estimated value of the vehicle interior noise after the failure of the sensor is detected, and the vehicle interior noise is calculated when the correlation value is equal to or greater than a predetermined value. Based on estimates of And a control unit for performing noise control for reducing the vehicle compartment noise and outputs the wave to the wave applying section Te, the output signal of a predetermined time period before the sensor failure detecting section detects a failure of the sensor An output signal storage unit for storing, and the noise estimation unit performs a predetermined filter process set based on a correlation between the output signal of the normal sensor and the vehicle interior noise on the output signal. And calculating an estimated value of the vehicle interior noise based on the sum of the output signals subjected to the filtering, and the control unit detects the output signal when the sensor failure detection unit detects a failure of the sensor. The sensor failure detection unit detects the sensor failure before the sensor failure detection unit detects the sensor failure based on the output signal stored in the storage unit and the filter processing before the sensor failure detection unit detects the sensor failure. An estimated value of indoor noise is calculated, and the output signal excluding the output signal of the sensor in which a failure is detected among the output signals stored in the output signal storage unit and the sensor failure detection unit Based on the filter processing after detecting a failure, the sensor failure detection unit calculates an estimated value of the vehicle interior noise after detecting the failure of the sensor, and from the two estimated values of the vehicle interior noise The noise control is performed based on the calculated correlation value .

  According to the present invention, in a noise control device that calculates an estimated value of noise based on vibration detected by a normal sensor and performs noise control based on the estimated value, the sensor fails when the sensor fails. Perform noise control without impairing the noise reduction effect by calculating the estimated noise value before and after and determining whether to implement or stop the noise control based on the correlation value of the estimated noise value before and after the sensor failure be able to.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  Typical causes of vehicle interior noise entering from outside the vehicle are engine noise caused by engine vibration, noise caused by road surface unevenness entering from tires during driving (hereinafter referred to as road noise). ), Wind noise generated by the airflow during driving. In this embodiment, reduction of road noise is mainly handled.

  FIG. 1 shows main propagation paths of vehicle body vibration and road noise caused by road surface unevenness.

  The vibration that becomes the main component of the road noise that has entered the vehicle body from the tire 200 first enters a highly rigid beam-like member called the member 140 from the mounting portion (not shown) of the axle 120 and the suspension 130. Thereafter, vibration propagates to a plate-like member called floor panel 110 surrounded by the member 140 and having relatively low rigidity, and the floor panel 110 vibrates. Furthermore, the vibration of the floor panel 110 causes air vibrations in the vehicle interior and causes a resonance phenomenon in the vehicle interior, so that road noise is heard in a predetermined space 100 (hereinafter referred to as a control space 100) in the vehicle interior. Although noise is also generated when the roof panel and window glass (both not shown) vibrate in addition to the floor panel 110, most of the road noise that enters mainly from the attachment portion of the suspension 130 is generated by the floor panel 110. It is known to be caused by vibration. For this reason, if noise control is performed so as to cancel road noise caused by vibration of the floor panel 110, road noise can be reduced.

  In the present invention, a sensor (described later) is arranged on the floor panel 110, vehicle interior noise is estimated based on the output signal of the sensor, a control command value is generated by the control unit, and based on this control command value A method is adopted in which control sound generated by an actuator (wave application unit) provided on the floor panel 110 is input into the passenger compartment.

  Here, the present invention uses a method of estimating the noise of the control space 100 from the signal of the acceleration sensor 10 without using a microphone as a sensor. Since the acceleration sensor 10 installed on the floor panel 110 is used, road noise caused by the floor panel 110 is handled as a control target. Here, the reason why the floor panel 110 is selected as the installation location of the acceleration sensor 10 is that coherence with the vehicle interior noise (definition will be described later) is high.

  In addition, since all the noises generated from the floor panel 110 are included in the control target, it is possible to handle a part of the engine noise and the wind noise generated by the air flowing through the bottom of the vehicle body.

  Further, the scope of the effect of the present invention is not limited to the category of noise reduction due to vibration of the floor panel 110. For example, vehicle interior noise generated by the same mechanism such as a dash panel, a front glass, and a roof panel (both not shown). The same effect can be obtained for the generation source if the present invention is used for the site.

  A schematic diagram of the noise control apparatus according to the present embodiment is shown in FIG.

  The noise control device according to the present embodiment includes an acceleration sensor 10 (10a, 10b, 10c, 10d) that is a sensor that measures vibration of the floor panel 110, and an actuator 20 (20a) that is a wave application unit that applies vibration to the floor panel 110. 20b) and a control unit 30 that calculates a control command value for reducing vehicle interior noise based on the signal obtained by the acceleration sensor 10 and controls the actuator 20.

  Here, the actuator 20 in the present embodiment is a so-called piezo-electric actuator.

  An input signal of the control unit 30 is an output of the acceleration sensor 10, and an output signal of the control unit 30 is a control command value to the actuator 20.

  The control unit 30 includes an amplification unit 31 (31a to 31f) for signal amplification and a control unit main body 32 that calculates and outputs a control command value for reducing vehicle interior noise.

  When the acceleration sensor 10 is a so-called charge charge type, the amplifying unit 31 also has a function of converting between charge and voltage.

  FIG. 3 is a block diagram showing the internal structure of the control unit main body 32.

A / D conversion section 33 (33a to 33e) is an acceleration signal alpha ₁ from the acceleration sensor _{10, α 2, α 3,} α 4 and the control command value _u 1, _{u 2} from the control portion main body 32 into a digital signal Convert.

The D / A converter 36 converts the control command values u ₁ and u ₂ calculated by the calculator 35 into analog signals and outputs the analog signals.

In the control unit main body 32, the noise in the control space 100 is reduced by using the acceleration signals α _{1 to} α ₄ output from the acceleration sensor 10 and the input signals (control command values u ₁ and u ₂ ) to the actuator 20. Calculate the control command value.

  A sufficient number of actuators 20 are attached to appropriate positions on the floor panel 110 of the vehicle body to reduce noise in the control space 100.

Here, the number of acceleration sensors 10 is generally required to be larger than the number of vibration sources. The specific number and installation positions of the acceleration sensors 10 are determined by coherency C _xy (ω) between each acceleration sensor 10 and the sound pressure of noise in the control space 100.

が十分高くなるように（例えば0.9以上）決定される。本実施形態では、加速度センサ１０とアクチュエータ２０の数は、それぞれ４個、２個とした。

Is determined to be sufficiently high (for example, 0.9 or more). In the present embodiment, the number of acceleration sensors 10 and actuators 20 is four and two, respectively.

Here, P _xy (ω) represents a cross power spectrum between acceleration and sound pressure, and P _xx (ω) and P _yy (ω) represent auto power spectra of acceleration and sound pressure, respectively. P ^H represents a Hermitian transpose matrix of P.

The noise estimation unit 34 calculates the estimated noise value SPL_est in the control space 100 using the accelerations α _{1 to} α ₄ output from the acceleration sensor 10 and the control command values u ₁ and u ₂ in the processing cycle one step before. To do.

The calculation unit 35 uses the estimated noise value SPL_est to calculate control command values u ₁ and u ₂ for the actuator 20 so as to reduce noise in the control space 100.

  In the present embodiment, the control unit main body 32 is mounted on a so-called CPU.

  A flowchart of the processing in the control unit 30 is shown in FIG.

In step S <b> 101, acceleration signals α _{1 to} α ₄ from the acceleration sensor 10 A / D converted by the A / D conversion unit 33 are input to the noise estimation unit 34. After this, the flow moves to step S102.

In step S < _b > 102, control command values u ₁ and u ₂ one step before A / D conversion by the A / D conversion unit 33 are input to the noise estimation unit 34. After this, the flow moves to step S103.

  In step S103, noise estimation processing is executed by the noise estimation unit 34, and an estimated value SPL_est of noise in the control space 100 is calculated from the signals obtained in S101 and S102. After this, the flow moves to step S104.

In step S104, the calculation unit 35 calculates control command values u ₁ and u ₂ for reducing noise in the control space 100 using the estimated noise value SPL_est calculated in S103. After this, the flow moves to step S105.

In step S105, the control command values u ₁ and u ₂ obtained in S104 are output to the D / A converter 36, and an output signal to the actuator 20 is output.

  The calculation unit 35 in FIG. 3 may be designed using any feedback control. For example, when designing as H∞ control, the following procedure may be followed.

The system model is a transfer function G _p (s) from the input voltage of the piezoelectric actuator 20 to noise. Here, s is a variable of Laplace transform.

For this transfer function G _p (s), the literature “D. McFarlane and K. Glover.“ A Loop Shaping Design Procedure Using H∞ Synthesis ”, IEEE Transactions on Automatic Control. Vol.37, no.6, June 1992 , Pp.759-769 ”, a control unit that reduces noise can be designed.

  In this method, the evaluation formula

を満足するようなコントローラC_∞(s)を設計する。ここで、G_s(s)は重み関数W₁(s)とW₂(s)により重み付けされた伝達関数

Design a controller C _∞ (s) that satisfies Where G _s (s) is the transfer function weighted by weight functions W ₁ (s) and W ₂ (s)

によって求められる。最終的に（数式２）の評価式を満足するコントローラC_∞(s)を用いて、コントローラC(s)は

Sought by. Finally, using controller C _∞ (s) that satisfies the evaluation formula of (Formula 2), controller C (s) is

として算出される。

Is calculated as

  The constant ε is a parameter for determining the stability margin of the controller, and 0.2 to 0.3 is usually recommended. When mounted on the CPU, for example, C (s) may be discretized by performing bilinear transformation on the controller C (s) and mounted as an IIR filter.

FIG. 5 shows the structure of the noise estimation unit 34. Hereinafter, 1) Noise control performed when a failure of the acceleration sensor 10 occurs, and 2) Determination of whether or not noise control is performed when a failure of the acceleration sensor 10 occurs is described separately. Do.
1) Noise control performed when a failure occurs in the acceleration sensor In this embodiment, when a failure occurs in the acceleration sensor 10, a filter 50 that multiplies the acceleration signal is newly set, and an estimated noise value is set. Calculate and perform noise control.

  First, in order to describe the basic operation, the operation will be described with reference to FIG. 6 in which only portions related to the basic operation are extracted from the block diagram shown in FIG.

The acceleration signals α _{1 to} α ₄ and the control command values u ₁ and u ₂ as digital signals input to the noise estimation unit 34 are input to the filter 50 (50a to 50d) and the transfer function 60 (60a and 60b), respectively. The

Here, the filters 50 a to 50 d are blocks that shape the acceleration signals α _{1 to} α ₄ in order to express noise entering the control space 100 from the outside of the vehicle as a sum of certain signals. This filter 50 is designed so that the sum of the processed signals becomes an estimated value of noise entering from outside the vehicle.

  Transfer functions 60a and 60b indicate transfer functions from the input voltage to the piezoelectric actuators 20a and 20b to the sound pressure in the control space 100, respectively. The transfer functions 60a and 60b can be obtained by inputting white noise or an impulse signal to the piezo actuator 20 and performing system identification using the sound pressure signal and the input signal in the control space 100 obtained at that time. The method includes, for example, “Structural Dynamical Toolbox” which is a tool box of the control system design tool MATLAB, and subspace identification described in the literature “Adachi,“ System Identification for Control ”, Tokyo Denki University Press, 1996”. The method may be used.

  By multiplying the input voltages of the piezo actuators 20a and 20b by the transfer functions 60a and 60b, respectively, and adding them, an estimated value of noise generated in the control space 100 by the vibration (sound) generated by the piezo actuator 20 is calculated. The

The acceleration signals α _{1 to} α ₄ shaped by the filter 50 and the control command values u ₁ and u ₂ calculated by the transfer function 60 are added by the adding unit 70. By this processing, an estimated value SPL_est of the noise in the control space 100 created by the vibration entering from the outside of the vehicle and the vibration generated by the piezo actuator 20 is calculated.

  FIG. 7 shows a flowchart of processing performed by the noise estimation unit 34.

In step S <b> 201, the acceleration signals α _{1 to} α _{4 of} the acceleration sensors _{10 a to 10} d are A / D converted by the A / D conversion units 33 a to 33 d and input to the noise estimation unit 34. After this, the flow moves to step S202.

In step S202, the control command values u ₁ and u _{2 of the} previous step are A / D converted by the A / D conversion units 33e and 33f and input to the noise estimation unit 34. After this, the flow moves to step S203.

In step S203, W ₁ is multiplied by a filter 50 which has been previously stored in the acceleration signal alpha ₁ input in S201. Similarly, W ₂ is a filter 50 to the acceleration signal alpha ₂ is the W ₃ is a filter 50 to the acceleration signal alpha _3, W ₄ are multiplied is a filter 50 to the acceleration signal alpha _4. After this, the flow moves to step S204.

In step S204, the control command value u ₁ input in S202, G _p1 is multiplied is a filter 60a showing a transfer function from the input voltage to the actuator 20a which is stored in advance to the sound pressure in the control space 100 Is done. Similarly, the control command value u _2, G _p2 is a filter 60b showing a transfer function from the input voltage to the actuator 20b until the sound pressure at the control space 100 is multiplied.

Here, G _p1 and G _p2 are set in advance as IIR filters by performing inverse Z conversion after identifying the transfer function from the input voltage of the piezo actuator 20 to noise as a discrete time system. After this, the flow moves to step S205.

  In step S205, the sum of all the signals obtained in S203 and S204 is taken and the obtained signal is output.

  Next, a method for setting the filter 50 will be described.

FIG. 8 shows the relationship between input vibration (vibration source) to the vehicle body, acceleration, and vehicle interior noise. The input vibration f to the vehicle body is transmitted to each acceleration sensor 10a, 10b, 10c, 10d through the transfer function H (s). On the other hand, the input vibration f propagates the air in the passenger compartment and becomes noise in the control space 100. The transfer function of air propagation at this time is set as R (s). Further, the accelerations detected by the acceleration sensors 10a, 10b, 10c, and 10d are set as α ₁ , α ₂ , α ₃ , and α ₄ , respectively. Furthermore, the road noise measured in the control space 100 is set as SPL. At this time, the Laplace transform of the input vibration f is f _L (s), the Laplace transform of the signal SPL is SPL _L (s), and the Laplace transform of the acceleration signals α ₁ , α ₂ , α ₃ , α ₄ is α _L1 (s ), Α _L2 (s), α _L3 (s), and α _L4 (s), the relationship between the signals is expressed by the following equation.

ここで、Hは各要素が伝達関数である４行１列の行列である。この関係式を用いて、加速度センサ１０の信号から騒音の推定値を推定するためには、（数式７）を逆にｆについて解き、（数式６）に代入すればよい。したがって、

Here, H is a 4 × 1 matrix in which each element is a transfer function. In order to estimate the estimated value of noise from the signal of the acceleration sensor 10 using this relational expression, (Equation 7) may be solved for f on the contrary and substituted into (Equation 6). Therefore,

で表される。ここで、H⁺は伝達関数行列H(s)の逆関数を表す。ここで、Hは正方行列ではなく、長方行列であるので、逆行列を計算することはできない。そこで、擬似逆行列

It is represented by Here, H ⁺ represents an inverse function of the transfer function matrix H (s). Here, since H is not a square matrix but a square matrix, an inverse matrix cannot be calculated. Therefore, the pseudo inverse matrix

を用いて演算を行う。ただし、m_HをHの行の数、n_HをHの列の数としたときに、

Calculate using. However, when m _H is the number of rows of _H and n _H is the number of columns of H,

であることが、H⁺を計算できるための必要条件である。

It is a necessary condition to be able to calculate H ⁺ .

Since RH ⁺ is a 1-by-4 matrix, its elements are

とおくと、（数式８）は

(Formula 8) is

と変形することができる。ここで現れるW₁からW₄を図５に記載のフィルタ５０ａ〜５０ｄとして設定する。したがって、各加速度α₁からα₄に対するフィルタ５０（W）は

And can be transformed. W ₁ to W ₄ appearing here are set as the filters 50a to 50d shown in FIG. Therefore, the filter 50 (W) for each acceleration α ₁ to α ₄ is

の列ベクトルにより決定される。

Determined by the column vector.

  FIG. 9 shows a flowchart for setting the filter.

  In step S301, a transfer function H and a transfer function R are calculated. These may be calculated in advance and stored as data. After this, the flow moves to step S302.

In step S302, based on the transfer function H, a function H ⁺ = (H ^T · H) ⁻¹ · H ^T is calculated. After this, the flow moves to step S303.

In step S303, the filter 50 is calculated as W = RH ⁺ .

  An example of the frequency response of the filter 50 obtained by the above method is shown in FIG. 10A to 10D are graphs of the acceleration sensors 10a to 10d, respectively. The dotted line in each graph shows the transfer function from the acceleration signal of each acceleration sensor 10 to the noise in the control space 100, and the solid line shows the characteristic of the function obtained by multiplying the transfer function by the filter 50 calculated by (Equation 13). ing. Here, when the dotted line and the solid line are close to each other in a certain frequency band, it indicates that the filter 50 has performed a large weighting on the acceleration signal in the frequency band detected by the acceleration sensor. For example, when attention is paid to the vicinity of 300 Hz in 10A to 10D in FIG. 10, the solid line and the dotted line are close to each other in the graph 10C, and the solid line and the dotted line are separated in the other graphs 10A, 10B, and 10D. In this case, the acceleration signal detected by the acceleration sensor 10c is heavily weighted in the vicinity of 300 Hz, and the acceleration signals of the other acceleration sensors 10 are lightly weighted.

  FIG. 11 shows the measured noise value and the estimated noise value SPL_est calculated using the filter 50 described above. A dotted line indicates the measured noise value, and a solid line indicates the estimated noise value SPL_est calculated by the noise estimation unit 34. From FIG. 11, it is understood that the estimated noise value SPL_est is calculated with high accuracy by using this embodiment. Thus, in this embodiment, the noise in the vehicle interior can be estimated with high accuracy, and the vehicle interior noise can be effectively reduced.

  With the processing described above, it is possible to estimate the noise in the passenger compartment with high accuracy.

  However, when the acceleration sensor 10 breaks down, the number of acceleration signals that can be used to estimate the noise in the passenger compartment is different, so that the noise estimation accuracy decreases. For this reason, a sufficient noise reduction effect may not be obtained.

  As an example of this, when the noise is estimated using the filter 50 designed to estimate the noise using the four acceleration sensors 10a to 10d, the acceleration sensor 10c and the acceleration sensor 10d fail. FIG. 12 shows the estimation result when the noise is estimated only by the acceleration sensor 10a and the acceleration sensor 10b. In FIG. 12, the dotted line indicates the actual measured value of noise, and the solid line indicates the estimated noise value SPL_est. In this case, since the components estimated by the acceleration sensors 10c and 10d disappear, the estimated noise value SPL_est is small. Thus, when the number of acceleration sensors 10 is reduced from a certain number, the noise estimation accuracy is clearly lowered. Therefore, in order to maintain the noise estimation accuracy even when the acceleration sensor 10 fails, the noise estimation unit 34 needs to take some measures.

  In the present embodiment, as a countermeasure against a failure of the acceleration sensor 10, when a failure of the acceleration sensor 10 is detected, the filter 50 corresponding to all combinations of the acceleration sensors 10 excluding the failed acceleration sensor 10 is stored in the storage unit ( A method for estimating noise by changing the filter 50 according to the failed acceleration sensor 10 is used.

  For example, when eight acceleration sensors 10 for noise estimation are used, combinations that cause the acceleration sensor 10 to malfunction are:

通りであるため、各組み合わせに対応したフィルタ５０を記憶部に記憶しておき、故障した加速度センサ１０が検知された場合には、故障した加速度センサ１０を除く加速度センサ１０の組合せに対応したフィルタ５０に変更して騒音の推定値を算出すればよい。なお、全ての加速度センサ１０が故障した場合には騒音の推定ができなくなるため、その場合のフィルタ５０は記憶しておく必要はない。

Therefore, the filter 50 corresponding to each combination is stored in the storage unit, and when a faulty acceleration sensor 10 is detected, the filter corresponding to the combination of the acceleration sensors 10 excluding the faulty acceleration sensor 10 is detected. The estimated value of noise may be calculated by changing to 50. Note that if all the acceleration sensors 10 fail, noise cannot be estimated, and the filter 50 in that case does not need to be stored.

  Hereinafter, the operation will be described with reference to FIG. 13 in which only the portion related to the filter change when the acceleration sensor 10 has failed is extracted from the block diagram shown in FIG.

In the noise estimation unit 34, the input acceleration signals α _{1 to} α ₄ of the acceleration sensor 10 are first input to an acceleration sensor failure detection unit 90 that is a sensor failure detection unit. Acceleration signals α _{1 to} α ₄ from the acceleration sensors 10 a to ₁₀ d are input to acceleration sensor failure detection units 90 a to 90 d of the acceleration sensor failure detection unit 90, respectively. The acceleration sensor failure detection unit 90 determines that the acceleration sensor 10 that outputs the acceleration signal has failed when, for example, the acceleration signal from the acceleration sensor 10 is smaller than a predetermined threshold value for a predetermined time. Thus, a failure of the acceleration sensor 10 is detected. When detecting a failure of the acceleration sensor 10, the acceleration sensor failure detection unit 90 outputs a failure detection signal to the filter setting unit 95.

Based on the failure detection signal from the acceleration sensor failure detection unit 90, the filter setting unit 95 newly adds the filter 50 to which the acceleration signal of the acceleration sensor 10 other than the acceleration sensor 10 in which the failure is detected is input among the filters 50a to 50d. Set to (change). That is, in the storage unit 96 provided in the filter setting unit 95, the filters 50 corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 are stored in advance, and the acceleration sensors 10 other than the failed acceleration sensor 10 are stored. The filter 50 corresponding to the combination is selected from the storage unit 96 and the filter 50 is newly set. The value obtained by multiplying the filter 50 set in this way by the acceleration signal and the signal obtained by multiplying the control command values u ₁ and u ₂ by the transfer function 60 are added in the same way as the noise estimation unit 34 shown in FIG. By adding at 70, an estimated value SPL_est of the noise in the control space 100 is calculated.

  FIG. 14 shows a flowchart in the noise estimation unit 34 in the present embodiment shown in FIG.

  In step S <b> 401, an acceleration signal is input to the noise estimation unit 34. After this, the flow moves to step S402.

  In step S402, the acceleration sensor failure detection unit 90 determines whether or not the acceleration sensor 10 has failed based on the time variation of the acceleration signal obtained in S401. If the acceleration sensor 10 has failed, a failure detection signal is output to the filter setting unit 95, and the flow proceeds to step S403. On the other hand, if the acceleration sensor 10 has not failed, the flow proceeds to step S404.

  In step S403, the filter setting unit 95 selects a new filter 50 and newly sets the filter 50. In the storage unit 96 provided in the filter setting unit 95, filters 50 corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 are stored in advance, and the acceleration at which a failure is detected according to the failure detection signal is stored. A filter 50 excluding the sensor 10 is selected from the storage unit 96 and set as a new filter 50. After this, the flow moves to step S404.

  In step S404, the control command value of one step before is input to the noise estimation unit 34. After this, the flow moves to step S405.

  In step S405, using the newly set filter 50, the filter 50 multiplies the acceleration signal from the acceleration sensor 10 excluding the acceleration sensor 10 in which a failure is detected. When no failure is detected in the acceleration sensor 10, the previous filter 50 is multiplied by the acceleration signal. After this, the flow moves to step S406.

  In step S406, the transfer function 60 stored in advance in the control command value is multiplied. After this, the flow moves to step S407.

  In step S407, all the signals obtained in S405 and S406 are added by the adder 70, and the obtained signal is output.

  FIG. 15 shows an example of a noise estimation result when the noise estimation unit 34 in the present embodiment is used.

  FIG. 15 shows the frequency characteristics of the filter 50 newly set to use the remaining acceleration sensors 10a and 10b when the acceleration sensors 10c and 10d fail. A graph 15a in FIG. 15 shows a graph of the acceleration sensor 10a, and a graph 15b in FIG. 15 shows a graph of the acceleration sensor 10b. Similar to FIG. 10, the dotted line represents the transfer function from the output signal of each acceleration sensor 10 to the noise in the control space 100, and the solid line represents the characteristic of the function obtained by multiplying the transfer function by the filter 50. In the frequency band where the dotted line and the solid line are close to each other in the graph, the acceleration signal in the frequency band detected by the acceleration sensor 10 has a large weight with respect to the estimated noise value SPL_est, and the filter 50 performs a large weighting. Yes.

  FIG. 16 shows the estimated noise value SPL_est when the filter 50 is newly set when the number of normal acceleration sensors 10 is changed from four to two. The dotted line is the measured noise value, and the solid line is the estimated noise value SPL_est. When the new setting of the filter 50 is not performed, the noise estimation accuracy decreases as shown in FIG. 12, but when the filter 50 is newly set, the noise estimation is performed with high accuracy as shown in FIG. Therefore, noise control with a high noise reduction effect can be performed.

2) Determination of whether or not to perform noise control when a failure occurs in the acceleration sensor 10 This operation will be described with reference to FIG.

  In the present embodiment, when a failure occurs in the acceleration sensor 10, as described above, the filter 50 multiplied by the acceleration signal is used to accurately estimate noise using the acceleration sensor 10 excluding the failed acceleration sensor 10. Is newly set. For example, if one of the four acceleration sensors 10 fails and the number of normal acceleration sensors 10 is three, noise estimation is performed if the filter 50 when the number of acceleration sensors 10 is four is used as it is. Therefore, the filter 50 in the case where the three acceleration sensors 10 are used in advance so that the remaining three acceleration sensors 10 excluding the failed acceleration sensor 10 can estimate the noise with high accuracy. Is stored in the storage unit 96, and the filter 50 is newly set (changed) using a failure detection signal corresponding to the failed acceleration sensor 10 as a trigger to estimate noise.

  This operation makes it possible to perform noise control without reducing the noise estimation accuracy as much as possible. However, when the number of failed acceleration sensors 10 increases, such a new setting of the filter 50 is performed. However, the noise estimation accuracy may be reduced.

  FIG. 17 shows the relationship between the number of normal acceleration sensors 10 and the correlation value R between the measured value and estimated value of noise. When there is no failure of the N acceleration sensors 10, the estimated noise value is close to the actual measurement value, and the correlation value is close to 1. On the other hand, when k acceleration sensors 10 have failed and the number of normal acceleration sensors 10 is (N−k), depending on the installation position of the failed acceleration sensor 10 (a combination of normal acceleration sensors). Since the noise estimation accuracy differs, the correlation value R between the measured noise value and the estimated value also varies. When the number of failed acceleration sensors 10 further increases and the number of normal acceleration sensors 10 is n, this variation further increases. As described above, when the number of failed acceleration sensors 10 increases, the noise estimation accuracy may be lowered even if the filter 50 described above is newly set.

  FIG. 18 shows the case where the number of normal acceleration sensors 10 is reduced with respect to the actual measured values of noise measured in the driver's seat and the passenger's seat by applying vibration to the left and right portions of the substantially central portion of the floor panel 110 of the vehicle body. It is a figure which shows the change of the estimation precision of noise. Here, the right figure of FIG. 18 shows the relationship between the number of normal acceleration sensors 10 and the correlation value R between the measured value and estimated value of noise in the driver's seat, and the left figure shows the normal acceleration sensor 10. The relationship between the number and the correlation value R between the measured value and estimated value of noise in the passenger seat is shown. As is clear from this figure, in this experimental condition, when the number of normal acceleration sensors 10 is three, noise can be estimated well in both the driver seat and the passenger seat, while the number of normal acceleration sensors 10 is When the number is two, it can be seen that the noise estimation accuracy varies greatly depending on the combination of the normal acceleration sensors 10, and the results differ between the driver seat and the passenger seat. For example, in combination 1, the correlation value R is about 0.2 or less for both the driver seat and the passenger seat, and the noise estimation accuracy is low. On the other hand, in the combinations 3 and 4, the noise estimation accuracy is relatively low in the passenger seat compared with the relatively low noise estimation accuracy in the vicinity of 0.65 at the driver seat, and the noise estimation accuracy is high.

  In the present embodiment, since there is no means for directly measuring the actual measured value of noise, the estimated noise value SPLN when all the acceleration sensors 10 are normal is used as the actual measured value of noise, and the faulty acceleration sensor 10 is replaced. The correlation value R between the corresponding estimated noise value SPLR and the estimated noise value SPLN is calculated, and a new setting of the filter 50 is performed only when the correlation value R is equal to or greater than the threshold value RT, that is, when the noise estimation accuracy is sufficient. Noise control is performed, and when the correlation value R is less than the threshold value RT, that is, when noise estimation accuracy is not sufficient, noise control is performed without damaging the noise suppression effect by stopping the noise control. is there.

  In the present embodiment, as shown in FIG. 5, an acceleration signal storage unit 80 that is an output signal storage unit is provided between the acceleration sensor failure detection unit 90 and the filter 50. Further, a noise control execution determination unit 81 is provided after the addition unit 70.

  Here, the acceleration signal when the acceleration sensor 10 is all normal is stored in the acceleration signal storage unit 80 at an arbitrary time, and when the acceleration sensor 10 fails, the acceleration signal stored in the acceleration signal storage unit 80 is stored. The estimated values SPLN and SPLR of the respective noises before and after the failure of the acceleration sensor 10 are calculated to determine whether noise control is to be performed or stopped based on the correlation value.

  Hereinafter, the operation of this embodiment will be described with reference to the flowchart of FIG.

  In step S <b> 501, the acceleration signal output from the acceleration sensor 10 is stored in the acceleration signal storage unit 80 when all the acceleration sensors 10 are normal. The acceleration signal to be stored is for a predetermined time. Here, the time for starting storage may be any time as long as all the acceleration sensors 10 are in a normal state. After this, the flow moves to step S502.

  In step S502, the acceleration sensor failure detection unit 90 determines whether or not the acceleration sensor 10 has failed. If a failure is detected in the acceleration sensor 10, the flow moves to step S503. On the other hand, if no failure is detected in the acceleration sensor 10, the flow proceeds to step S509.

  In step S503, the acceleration signal stored in the acceleration signal storage unit 80 is read. After this, the flow moves to step S504.

  In step S504, noise estimation when all the acceleration sensors 10 are normal is performed using the acceleration signal read from the acceleration signal storage unit 80 and the filter 50 when all the acceleration sensors 10 are normal. Calculate the value SPLN. After this, the flow moves to step S505.

  In step S505, the filter 50 corresponding to the combination of the normal acceleration sensors 10 is newly set according to the failure of the acceleration sensor 10. The setting of the filter 50 is the same as that described in the flowchart of FIG. After this, the flow moves to step S506.

  In step S506, a failure is detected in the acceleration sensor 10 using the acceleration signal excluding the acceleration signal of the failed acceleration sensor 10 among the acceleration signals stored in the acceleration signal storage unit 80 and the newly set filter 50. The estimated value SPLR of the noise after being calculated is calculated. After this, the flow moves to step S507.

  In step S507, a correlation value R between the calculated estimated noise value SPLN and SPLR is calculated. After this, the flow moves to step S508.

  In step S508, the calculated correlation value R and the threshold value RT are compared. If correlation value R ≧ threshold RT, the flow moves to step S509. On the other hand, if correlation value R <threshold RT, the flow moves to step S510.

  In step S509, noise control is performed when the estimated noise value SPLR is calculated with high accuracy. The noise control is the same as described above.

  In step S510, the noise control is stopped when the estimated noise value SPLR is not calculated with high accuracy.

  The calculation of the correlation value R between the estimated noise value SPLN and the estimated noise value SPLR and the determination of whether or not to perform noise control are performed by the noise control execution determination unit 81.

  With the above operation, when the number of failed acceleration sensors 10 increases and the noise estimation accuracy is not sufficient, the noise control can be stopped. As a result, by performing noise control with an estimated value of low-accuracy noise, it is possible to eliminate the possibility of increasing the noise, and the noise reduction effect is not impaired.

  In the above embodiment, noise control may be performed when the correlation value R before and after failure of the acceleration sensor in the plurality of control spaces 100 is equal to or greater than the threshold value RT. Thereby, when the noise reduction effect cannot be obtained in all of the plurality of control spaces, the noise control is not performed, and the noise control capable of reliably obtaining the noise reduction effect in all of the plurality of control spaces can be performed.

  This example differs from the above-described embodiment in the operation of the acceleration signal storage unit 80. In this embodiment, the acceleration signal storage unit 80 continuously updates and stores the acceleration signal from the acceleration sensor 10 for a predetermined time, and the acceleration signal stored in the acceleration signal storage unit 80 when the acceleration sensor 10 fails. Noise estimation values before and after the failure of the acceleration sensor 10 are calculated using the acceleration signal before the failure occurs, and noise control is performed based on the correlation value of the noise estimation values before and after the failure of the acceleration sensor 10. Is performed or stopped.

  In order to perform this operation, in this embodiment, the acceleration signal storage unit 80 performs a so-called FIFO (First In First Out) ring buffer operation.

  Hereinafter, the operation of this embodiment will be described with reference to the flowchart of FIG.

In step S601, the acceleration signal storage unit 80 stores the latest acceleration signal output from the acceleration sensor 10 for a time _tmax . After this, the flow moves to step S602.

  In step S602, the acceleration sensor failure detection unit 90 determines whether or not the acceleration sensor 10 has failed. If a failure is detected in the acceleration sensor 10, the flow moves to step S603. On the other hand, if no failure is detected in the acceleration sensor 10, the flow proceeds to step S611.

  In step S603, the update of the memory in the acceleration signal storage unit 80 is stopped. After this, the flow moves to step S604.

In step S604, all acceleration signals stored after detecting a failure of the acceleration sensor 10 among the acceleration signals stored in the acceleration signal storage unit 80 are deleted. FIG. 21 shows an example of the acceleration signal stored in the acceleration signal storage unit 80. The failure of the acceleration sensor 10a is detected at time t _f in Figure 21. In this case, if the time at which updating of the acceleration signal storage unit 80 is stopped after the acceleration sensor failure detection unit 90 detects the failure of the acceleration sensor 10a is assumed to be _end , the acceleration signal output by the failed acceleration sensor 10a. Is stored in the acceleration signal storage unit 80 for a time (t _end −t _f ). For this reason, the acceleration signal at time (t _end −t _f ) is deleted from all acceleration signals stored in the acceleration signal storage unit 80. After this, the flow moves to step S605.

  Steps S605 to S612 are the same as steps S503 to S510 in the embodiment, and thus description thereof is omitted.

  With the above operation, when the number of failed acceleration sensors 10 increases and the noise estimation accuracy is not sufficient, the noise control can be stopped. As a result, by performing noise control with an estimated value of low-accuracy noise, it is possible to eliminate the possibility of increasing the noise, and the noise reduction effect is not impaired.

  A block diagram of this embodiment is shown in FIG.

  In the present embodiment, a correlation value R between the estimated noise value SPLN when all the acceleration sensors 10 are normal and the estimated noise value SPLR corresponding to the failed acceleration sensor 10 is calculated in advance. When the noise is higher than the threshold RT, that is, when the noise estimation accuracy is sufficient, the filter 50 is newly set to perform noise control, and the correlation value R is less than the threshold RT, that is, when the noise estimation accuracy is not sufficient. The noise control is performed without damaging the noise suppression effect by the operation of stopping the noise control.

  The failure detection signal of the acceleration sensor failure detection unit 90 is input to the noise control execution determination unit 82.

  The noise control execution determination unit 82 includes a CDT storage unit 83, receives a failure detection signal from the acceleration sensor failure detection unit 90, refers to a CDT value (described later) stored in advance in the CDT storage unit 83, and determines a CDT value. It is determined whether noise control is to be executed or stopped by using as an index. If it is determined that noise control is to be performed, a failure detection signal is output to the filter setting unit 95, and a new filter is set by the filter setting unit 95. On the other hand, when it is determined that the noise control is to be stopped, the failure detection signal is not output to the filter setting unit 95, and the operation of the noise estimation unit 34 is stopped.

  Here, the CDT value will be described.

  A table showing the case where “the correlation value R is equal to or greater than the threshold value RT” is a control determination table CDT, and the value described in this table is called a CDT value. The CDT value is set to “1” when the correlation value R is greater than or equal to the threshold value RT, and is set to “0” when the correlation value R is less than the threshold value RT. FIG. 23 shows an example of the control determination table CDT.

  The horizontal axis of the control determination table CDT represents a numerical value i (hereinafter referred to as a noise estimated position i) corresponding to the position where noise is estimated, and the vertical axis represents the number j of normal acceleration sensors 10. Here, the CDT values “1” and “0” set in the table are set according to the flowchart shown in FIG.

  In step S701, an estimated value SPLiN of noise when all N acceleration sensors 10 are normal at each noise estimated position i is calculated. After this, the flow moves to step S702.

In step S <b> 702, when k acceleration sensors 10 (1 ≦ k ≦ N−number of vibration sources) have failed at each noise estimation position i, all combinations _N C _j (j = N−) of the remaining acceleration sensors 10 are detected. The estimated noise value SPLij in k) is calculated. After this, the flow moves to step S702.

  In step S703, the correlation between SPLiN and SPLij is calculated, and a correlation value Rij is calculated. After this, the flow moves to step S704.

  In step S704, the correlation value Rij is compared with the threshold value RT. If correlation value Rij ≧ threshold RT, the flow moves to step S705. On the other hand, if correlation value Rij <threshold RT, the flow moves to step S706.

  In step S705, assuming that the noise estimation accuracy is high and the noise reduction effect by noise control can be expected, “1” meaning that noise control is performed is set to CDT (i, j). After this, the flow moves to step S707.

In step S706, assuming that the noise estimation accuracy is low and the noise reduction effect due to the noise control cannot be expected, “0” meaning to stop the noise control is set to CDT (i, j). After this, the flow moves to step S 7 07.

At step S707, the all i, whether the performance of step S 7. 01 to step S 7 06 against j is determined. All i, and performs step S 7. 01 to step S 7 06 against j, if all CDT (i, j) is set, and ends. On the other hand, all i, the step S 7. 01 to step S 7 06 has not been performed for j, if there is a CDT not set (i, j), the flow returns to step S 7 01.

  With the above operation, the CDT value in the control determination table CDT is set.

  In the above flowchart, when calculating the estimated noise value SPLiN, a standard acceleration signal may be used.

  Next, the noise control execution determination unit 82 using the control determination table CDT will be described with reference to the flowchart of FIG.

  In step S801, the acceleration sensor failure detection unit 90 determines whether or not the acceleration sensor 10 has failed. If the acceleration sensor failure detection unit 90 detects a failure of the acceleration sensor 10, the flow moves to step S802. On the other hand, if the acceleration sensor failure detection unit 90 does not detect a failure of the acceleration sensor 10, the flow moves to step S805.

  In step S <b> 802, a CDT value corresponding to a normal combination of acceleration sensors 10 is selected from the CDT storage unit 83. After this, the flow moves to step S803.

  In step S803, branching based on the CDT value selected in step S802 is performed. When the CDT value = 1, the flow proceeds to step S804 when the normal acceleration sensor 10 can estimate noise with high accuracy. On the other hand, if the CDT value = 0, it is difficult to accurately estimate noise with the normal acceleration sensor 10, and the flow moves to step S806.

  In step S804, a filter 50 corresponding to a combination of normal acceleration sensors 10 is newly set. The setting of the filter 50 is the same as that described in the flowchart of FIG. After this, the flow moves to step S805.

  In step S805, noise control is performed. The noise control is equivalent to the content described in the flowchart of FIG.

  In step S806, noise control is stopped.

  With the above operation, when the number of failed acceleration sensors 10 increases and the noise estimation accuracy is not sufficient, the noise control can be stopped. As a result, by performing noise control with an estimated value of low-accuracy noise, it is possible to eliminate the possibility of increasing the noise, and the noise reduction effect is not impaired.

  This example is an example in which the processing content of step S403 in the flowchart of FIG. 14 is changed in the above-described embodiment, and has the same configuration as the above-described embodiment. For this reason, only step S403 in which the processing content is different from that of the above-described embodiment will be described.

  The filter 50 is as described in (Formula 13).

により求められる。ここで、加速度センサ１０の故障時には伝達関数Rは変化せず伝達関数H⁺のみに新たな設定の必要が生じるため、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したフィルタ５０を予め記憶部９６に記憶しておく代わりに、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したH⁺を記憶しておき、それを用いてフィルタ５０を新たに設定することが可能ある。こうすることで、記憶部９６に記憶しておくデータ量が小さくなり、より少ない記憶容量でフィルタ５０の新たな設定を実現することができる。

Is required. Here, when the acceleration sensor 10 fails, the transfer function R does not change, and a new setting is required only for the transfer function H ^+. Therefore, the filter 50 corresponding to all combinations of the acceleration sensors 10 except for the failed acceleration sensor 10. Is stored in the storage unit 96 in advance, H ⁺ corresponding to all combinations of the acceleration sensors 10 except for the failed acceleration sensor 10 is stored, and the filter 50 is newly set using the stored H ^+. Is possible. By doing so, the amount of data stored in the storage unit 96 is reduced, and a new setting of the filter 50 can be realized with a smaller storage capacity.

  FIG. 26 shows a flowchart of new setting of the filter 50 in the present embodiment.

In step S901, H ⁺ corresponding to the combination of the remaining acceleration sensors 10 excluding the failed acceleration sensor 10 is selected from the storage unit 96. Here, H ⁺ corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 is calculated in advance and stored in the storage unit 96. After this, the flow moves to step S902.

In step S902, the filter 50 (W) is calculated by (Equation 15) using H ⁺ calculated in S901.

  Hereinafter, even when the acceleration sensor 10 fails, the estimated noise value SPL_est is calculated with high accuracy by the same processing as in the above-described embodiment, and a high noise reduction effect can be obtained.

  This example is an example in which the processing content of step S403 in the flowchart of FIG. 14 is changed in the above-described embodiment, and has the same configuration as the above-described embodiment. For this reason, only step S403 in which the processing content is different from that of the above-described embodiment will be described.

  As described in (Formula 13), the filter 50 is

により算出される。ここで、加速度センサ１０の故障時には伝達関数Rは変化せずH⁺のみに新たな設定の必要が生じるため、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したフィルタ５０を予め記憶部９６に記憶しておく代わりに、実施例３に示したように、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したH⁺を予め記憶部９６に記憶しておくことで、フィルタ５０の新たな設定が可能となる。ここで、H⁺はHから算出が可能であるため、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したH⁺を予め記憶部９６に記憶しておく代わりに、故障した加速度センサ１０を除くすべての加速度センサ１０の組合せに対応したHを予め記憶部９６に記憶しておき、このHを用いてH⁺を算出することで、フィルタ５０の新たな設定が可能となる。このようにすることで、記憶部９６に記憶しておくデータ量がさらに小さくなり、より少ない記憶容量でフィルタ５０の新たな設定を実現することができる。

Is calculated by Here, when the acceleration sensor 10 fails, the transfer function R does not change and a new setting is required only for H ^+. Therefore, the filters 50 corresponding to all combinations of the acceleration sensors 10 except for the failed acceleration sensor 10 are preliminarily set. Instead of storing in the storage unit 96, as shown in the third embodiment, H ⁺ corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 is stored in the storage unit 96 in advance. Thus, a new setting of the filter 50 becomes possible. Here, since H ⁺ can be calculated from H, instead of storing H ⁺ corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 in the storage unit 96 in advance, the failed acceleration is detected. By storing H corresponding to all combinations of the acceleration sensors 10 except the sensor 10 in the storage unit 96 in advance, and calculating H ⁺ using this H, a new setting of the filter 50 becomes possible. By doing in this way, the data amount memorize | stored in the memory | storage part 96 becomes still smaller, and the new setting of the filter 50 is realizable with less memory capacity.

  FIG. 27 shows a flowchart of new setting of the filter 50 in the present embodiment.

  In step S911, H corresponding to the combination of the remaining acceleration sensors 10 excluding the failed acceleration sensor 10 is selected from the storage unit 96. Here, H corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 is calculated in advance and stored in the storage unit 96. After this, the flow moves to step S912.

  In step S912, the filter 50 (W) is calculated by (Equation 16) using H selected in S911.

  Hereinafter, even when the acceleration sensor 10 fails, the estimated noise value SPL_est is calculated with high accuracy by the same processing as in the above-described embodiment, and a high noise reduction effect can be obtained.

This example is an example in which the processing content of step S403 in the flowchart of FIG. 14 is changed in the above-described embodiment, and has the same configuration as the above-described embodiment. For this reason, only step S403 in which the processing content is different from the above-described embodiment will be described.
As shown in (Equation 13), the filter 50 is

で表される。ここでWは、W=[W₁、W₂、W₃、W₄]として表される伝達関数を要素とする１行４列行列である。

It is represented by Here, W is a 1 × 4 matrix having a transfer function expressed as W = [W ₁ , W ₂ , W ₃ , W ₄ ] as an element.

  When the acceleration sensor 10 fails, the transfer function R in (Equation 17) is a transfer function from the vibration source to the noise, and is not changed, and only the transfer function H changes. Here, since H is a transfer function from the vibration source to the acceleration signal of the acceleration sensor 10,

と、行列の形で表すことができる（以降、Hを伝達関数行列と呼ぶ）。（数式１８）において、Hの各要素H_a〜H_dは、それぞれ振動源から４つの加速度センサ１０ａ〜１０ｄまでの伝達関数である。

And can be expressed in the form of a matrix (hereinafter, H is referred to as a transfer function matrix). In (Formula 18), each element H _{a to} H _d of H is a transfer function from the vibration source to the four acceleration sensors 10 a to 10 _d .

Here, considering the transfer function matrix H when the acceleration sensor 10 fails, for example, when the acceleration sensors 10c and 10d fail, the rows of the transfer functions H _c and H _d corresponding to the failed acceleration sensor 10 are deleted. And transfer function matrix H again

とすることにより、加速度センサ１０ａと１０ｂのみで騒音の推定を行っていた場合と同じ状態とすることができる。したがって、加速度センサ１０の故障を検出した場合には、伝達関数行列Hの行の中で故障した加速度センサ１０に対応する行ベクトルを全て削除し、（数式１９）のように新たな伝達関数行列Hを設定する。次に、その伝達関数行列Hを用いて（数式９）から擬似逆行列H⁺を求め、最後にRH⁺を演算し、得られた行列の縦ベクトルを各フィルタ５０とすることにより、フィルタ５０を新たに設定することができる。

By doing so, it is possible to obtain the same state as when noise is estimated only by the acceleration sensors 10a and 10b. Therefore, when a failure of the acceleration sensor 10 is detected, all row vectors corresponding to the failed acceleration sensor 10 in the row of the transfer function matrix H are deleted, and a new transfer function matrix is obtained as in (Equation 19). Set H. Next, using the transfer function matrix H, a pseudo inverse matrix H ⁺ is obtained from (Equation 9), RH ⁺ is finally calculated, and the vertical vector of the obtained matrix is used as each filter 50, so that the filter 50 Can be newly set.

  By using this method, it is not necessary to store H corresponding to all combinations of the acceleration sensors 10 except the failed acceleration sensor 10 in the storage unit 96, and if only one H is stored, the filter 50 can be used. It becomes possible to newly set. By doing in this way, the data amount memorize | stored in the memory | storage part 96 becomes still smaller, and the new setting of the filter 50 is realizable with less memory capacity.

  FIG. 28 shows a flowchart of new setting of the filter 50 in the present embodiment.

  In step S921, the row of the transfer function matrix H corresponding to the failed acceleration sensor 10 is deleted, and a new transfer function matrix H is set. After this, the flow moves to step S922.

In step S922, the matrix RH ⁺ = R (H ^T · H) ⁻¹ · H ^T is calculated using the transfer function matrix H set in S921.

In step S923, the filter 50 is updated using each column vector of the calculated matrix RH ⁺ as the filter 50.

  Hereinafter, even when the acceleration sensor 10 fails, the estimated noise value SPL_est is calculated with high accuracy by the same processing as in the above-described embodiment, and a high noise reduction effect can be obtained.

  According to the noise control device of the embodiment of the present invention described above, the following operational effects can be obtained.

The invention of claim 1 is based on a plurality of sensors that are arranged in a vehicle body of the vehicle and detect vibrations of the vehicle body, a wave application unit that applies a controlled wave to the vehicle body, and output signals output from the plurality of sensors. A sensor failure detection unit that detects a failure of the sensor, and a sound that is heard in a predetermined space in the vehicle interior of the vehicle based on the output signal of the normal sensor excluding the sensor in which the failure is detected by the sensor failure detection unit A noise estimation unit that calculates an estimated value of vehicle interior noise, and when the sensor failure is detected by the sensor failure detection unit, the vehicle interior noise is estimated before the sensor failure is detected. A correlation value between the value and the estimated value of the vehicle interior noise after the failure of the sensor is detected. When the correlation value is equal to or greater than a predetermined value, the wave application is performed based on the estimated value of the vehicle interior noise. Wave in the part And a control unit for performing noise control by force reduces the vehicle interior noise, and an output signal storage unit for the sensor failure detecting section stores the output signal of a predetermined time period prior to detecting a failure of said sensor The noise estimation unit performs a predetermined filter process set based on a correlation between the output signal of the normal sensor and the vehicle interior noise on the output signal, and performs the filter process. The estimated value of the vehicle interior noise is calculated based on the sum of output signals, and the control unit stores the output stored in the output signal storage unit when the sensor failure detection unit detects a failure of the sensor. Based on the signal and the filter process before the sensor failure detection unit detects the sensor failure, the sensor failure detection unit calculates an estimated value of the vehicle interior noise before the sensor failure is detected. In addition, the output signal except the output signal of the sensor in which a failure is detected among the output signals stored in the output signal storage unit, and the filter after the sensor failure detection unit has detected the failure of the sensor The sensor failure detection unit calculates an estimated value of the vehicle interior noise after detecting the sensor failure based on the processing, and based on the correlation value calculated from the two estimated values of the vehicle interior noise, It is characterized by noise control .

According to this device, in a noise control device that calculates an estimated value of noise based on vibration detected by a normal sensor and performs noise control based on the estimated value, the sensor fails when the sensor fails. Estimated noise values before and after are calculated, and noise control is performed when the correlation value between estimated noise values before and after sensor failure exceeds a predetermined value, so noise reduction effect is not impaired be able to. Furthermore, according to this apparatus, the correlation value of the estimated noise values before and after the sensor failure can be easily calculated by storing the output signal before the sensor failure occurs .

According to a second aspect of the present invention, there are provided a plurality of sensors arranged on a vehicle body for detecting vibrations of the vehicle body, a wave applying unit for applying a controlled wave to the vehicle body, and output signals output from the plurality of sensors. A sensor failure detection unit that detects a failure of the sensor based on the sensor, and a predetermined space in the vehicle interior of the vehicle based on the output signal of the normal sensor excluding the sensor in which the failure is detected by the sensor failure detection unit The vehicle interior noise before the sensor failure is detected when the sensor failure detection unit detects the sensor failure. A correlation value between the estimated value of the vehicle interior noise and the estimated value of the vehicle interior noise after the failure of the sensor is detected, and when the correlation value is equal to or greater than a predetermined value, based on the estimated value of the vehicle interior noise Wave in the wave application part A control unit that performs noise control to output and reduce the vehicle interior noise, and an output signal storage unit that updates and stores the latest output signal for a predetermined time, and the noise estimation unit is normal A predetermined filtering process set based on the correlation between the output signal of the sensor and the vehicle interior noise is performed on the output signal, and the vehicle interior is performed based on the sum of the output signals subjected to the filter process. An estimated value of noise is calculated, and when the sensor failure detection unit detects a failure of the sensor, the control unit stops updating the storage of the output signal storage unit and is stored in the output signal storage unit. Of the output signals, the sensor failure detection unit detects the sensor failure based on the output signal before the sensor failure detection unit detects the sensor failure and the filter processing before the sensor failure detection unit detects the sensor failure. The failure detection unit calculates an estimated value of the vehicle interior noise before the failure of the sensor is detected, and the sensor failure detection unit detects a failure of the sensor among the output signals stored in the output signal storage unit. The sensor based on the output signal before detection and excluding the output signal of the sensor that has detected the failure, and the filtering process after the sensor failure detection unit has detected the failure of the sensor A failure detection unit calculates an estimated value of the vehicle interior noise after detecting a failure of the sensor, and performs the noise control based on a correlation value calculated from the two estimated values of the vehicle interior noise. It is said.

According to this device, in a noise control device that calculates an estimated value of noise based on vibration detected by a normal sensor and performs noise control based on the estimated value, the sensor fails when the sensor fails. Estimated noise values before and after are calculated, and noise control is performed when the correlation value between estimated noise values before and after sensor failure exceeds a predetermined value, so noise reduction effect is not impaired be able to. Further, according to this apparatus, according to this apparatus, the correlation value of the estimated noise values before and after the sensor failure can be easily calculated by storing the output signal before the sensor failure occurs. Can do.

According to a third aspect of the present invention, in the noise control device according to the first or second aspect, the filter processing is a weighting set based on a correlation between the output signal of the normal sensor and the vehicle interior noise. In the weighting process, the weighting process coefficient is set so that the output signal having a higher correlation with the vehicle interior noise has a higher contribution to the estimated value of the vehicle interior noise. is doing.

  According to this apparatus, the weighting process is performed so that the output signal having a higher correlation with the vehicle interior noise has a higher contribution to the estimated value of the vehicle interior noise, and the vehicle interior noise can be estimated with high accuracy.

  According to a fourth aspect of the present invention, in the noise control apparatus according to the third aspect, the weighting process is performed for each predetermined frequency band.

  According to this apparatus, the weighting process is performed for each predetermined frequency band, and the vehicle interior noise can be estimated with high accuracy.

  According to a fifth aspect of the present invention, in the noise control device according to the third or fourth aspect, the filter processing includes: the output signal of the normal sensor; the output signal; and an estimated value of the vehicle interior noise. It is characterized in that the transfer function G is multiplied by the weighting coefficient.

  According to this apparatus, the vehicle interior noise can be estimated with high accuracy from the output signal of the sensor and the transfer function G.

  Further, in the noise control device according to any one of claims 3 to 5, the noise estimation unit may calculate the weighting coefficient corresponding to all combinations of the normal sensors. It is characterized in that a storage unit is provided which stores the weighting coefficient calculated in advance and calculated.

  According to this apparatus, the processing time can be shortened by storing the weighting coefficient in advance.

According to a seventh aspect of the present invention, in the noise control device according to the first or second aspect, the filter processing includes the output signal of the normal sensor and a vehicle body that is a source of vibration of the output signal and the vehicle body. Multiplying the function (H ^T · H) ⁻¹ · H ^T based on the transfer function H between the vibration source and the transfer function R between the vehicle body vibration source and the estimated value of the vehicle interior noise It is a feature.

  According to this apparatus, the filter process is calculated based on the transfer function H, and the vehicle interior noise can be estimated with high accuracy.

The invention according to claim 8 is the noise control device according to claim 7, wherein the noise estimation unit is configured to use the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal sensors. Is calculated in advance, and a storage unit for storing the calculated function (H ^T · H) ⁻¹ · H ^T is provided.

  According to this apparatus, the storage unit stores less capacity, and the storage capacity required by the storage unit can be reduced.

The invention according to claim 9 is the noise control apparatus according to claim 7, wherein the noise estimation unit is configured to use the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal sensors. The transfer function H used for calculating the transfer function H is calculated in advance, and a storage unit for storing the calculated transfer function H is provided.

  According to this apparatus, the storage unit stores less capacity, and the storage capacity required by the storage unit can be reduced.

  In the noise control device according to claim 7, the noise estimation unit may calculate the transfer function H corresponding to a case where the sensor failure detection unit has not detected a failure of the sensor. When the sensor failure detection unit detects a failure of the sensor, the row component of the transfer function H corresponding to the sensor in which the failure is detected is deleted and a new transfer function H is obtained. It is characterized by calculating.

  According to this apparatus, the transfer function H can be easily calculated.

  The invention according to claim 11 is the noise control device according to any one of claims 7 to 10, wherein the number of the vehicle body vibration sources is the number of wheels provided in the vehicle.

  According to this apparatus, the number of wheels is the number of vehicle body vibration sources, so that the noise in the passenger compartment can be estimated with high accuracy.

According to a twelfth aspect of the present invention, in the noise control device according to any one of the first to eleventh aspects, the correlation value is calculated in advance according to the position of the predetermined space and the number of normal sensors. And when the sensor failure detection unit detects a failure of the sensor, the control unit controls the noise based on the correlation value stored in the correlation value storage unit. It is characterized by performing.

  According to this apparatus, since the correlation value of the estimated noise value before and after the occurrence of the sensor failure is calculated and stored in advance, when the sensor failure occurs, by referring to the corresponding correlation value, the noise It is possible to easily determine whether to execute the control.

The invention according to claim 13 is the noise control device according to any one of claims 1 to 12 , wherein the noise control is performed based on the correlation values for a plurality of the predetermined spaces. .

  According to this apparatus, when a sensor failure occurs, noise control can be performed on a plurality of spaces without impairing the noise reduction effect.

The invention according to claim 14 is the noise control device according to any one of claims 1 to 13 , wherein the sensor failure detection unit is configured to detect the sensor based on an output value and an output time of the output signal. It is characterized by detecting a failure.

  According to this apparatus, it is possible to easily detect a sensor failure based on the output value and output time of the output signal of the sensor.

Furthermore, the invention of claim 15 is the noise control device according to any one of claims 1 to 14 , wherein the wave is a vehicle body vibration applied to the vehicle body or a sound generated in the vehicle interior of the vehicle body. It is characterized by being.

  According to this device, vehicle interior noise can be reduced by applying vibration or sound to the vehicle body.

  Further, according to the noise control method of the embodiment of the present invention described above, the following operational effects can be obtained.

According to a sixteenth aspect of the present invention, there is provided a first step of detecting vibrations of a vehicle body of a vehicle at a plurality of detection points, and a second step of detecting the detection points at which the vibration is not normally detected among the plurality of detection points . steps and, interior noise heard at a predetermined space in the cabin of the vehicle based on the vibration from the normal the detection region other than the detection portion which is not normally detected the vibration of the plurality of detecting locations A third step of calculating an estimated value of the vehicle interior noise of the vehicle interior before detecting the detected location when the detected location where the vibration is not normally detected among the detected locations is detected. A correlation value between the estimated value and the estimated value of the vehicle interior noise after detecting the detected portion is calculated, and the vehicle body is controlled based on the estimated value of the vehicle interior noise when the correlation value is a predetermined value or more. Made wave A fourth step of performing a noise control for reducing the pressure teeth the vehicle interior noise, the vibration of a predetermined time period prior to detection of the detection portion which is not able to properly detect the vibration of the plurality of detection points A third step of storing a predetermined filtering process set based on a correlation between the vibration from the normal detection location and the vehicle interior noise with respect to the vibration. The estimated value of the vehicle interior noise is calculated based on the sum of the vibrations that has been subjected to the filtering process, and the fourth step cannot detect the vibration normally among the plurality of detection points. When the detection location is detected, before the vibration is not normally detected based on the stored vibration and the filter processing before detecting the detection location where the vibration is not normally detected The estimated value of the vehicle interior noise before detecting the detection location is calculated, and the vibration and the vibration except for the vibration at the detection location where the vibration is not normally detected among the stored vibrations are normalized. Based on the filter processing after detecting the detection location that has not been detected, the estimated value of the vehicle interior noise after detecting the detection location where the vibration is not normally detected is calculated, and the two The noise control is performed based on a correlation value calculated from the estimated value of the vehicle interior noise .

According to this method, in a noise control method for calculating an estimated value of noise based on vibration detected at a normal detection location and performing noise control based on the estimated value, when the detection location of vibration fails, Noise is reduced because noise estimates are calculated before and after the vibration detection location fails and the correlation value of the noise estimation values before and after the vibration detection location fails is greater than or equal to a predetermined value. Noise control without impairing the effect can be performed. Further, according to this method, the noise before and after detecting the detection location where the vibration is not normally detected is stored, so that noise estimation before and after detecting the detection location where the vibration is not normally detected can be stored. The correlation value of values can be easily calculated .

According to a seventeenth aspect of the present invention, a first step of detecting vibrations of a vehicle body of a vehicle at a plurality of detection points, and detecting the detection points at which the vibrations are not normally detected among the plurality of detection points. A vehicle that can be heard in a predetermined space in a vehicle interior of the vehicle based on the second step and the vibrations from the normal detection points other than the detection points where the vibrations are not normally detected among the plurality of detection points. A third step of calculating an estimated value of indoor noise, and the vehicle interior before detecting the detection location when the detection location where the vibration is not normally detected among the plurality of detection locations is detected. A correlation value between the estimated value of noise and the estimated value of the vehicle interior noise after detecting the detection location is calculated, and the vehicle body is based on the estimated value of the vehicle interior noise when the correlation value is equal to or greater than a predetermined value. Controlled A fourth step of performing noise control for applying motion to reduce the vehicle interior noise, and a fifth step of updating and storing the latest vibration for a predetermined time, wherein the third step includes: A predetermined filter process set based on the correlation between the vibration from the normal detection point and the vehicle interior noise is performed on the vibration, and the filter process is performed based on the sum of the vibrations. An estimated value of vehicle interior noise is calculated, and the fourth step stops updating the memory of the vibration when detecting the detected portion where the vibration is not normally detected among the plurality of detected portions. Then, the vibration before detecting the detection point where the vibration is not normally detected among the stored vibrations and the filtering process before detecting the detection point where the vibration is not normally detected Base And calculating an estimated value of the vehicle interior noise before detecting the detected location where the vibration is not normally detected, and detecting the detected location where the vibration is not normally detected among the stored vibrations The vibration before the vibration and the vibration at the detection location where the vibration is not normally detected except the vibration and the filter processing after the detection location at which the vibration is not normally detected are detected. Based on the correlation value calculated from the two estimated values of the vehicle interior noise after calculating the estimated value of the vehicle interior noise after detecting the detection location where the vibration is not normally detected based on the noise It is characterized by performing control .

According to this method, in a noise control method for calculating an estimated value of noise based on vibration detected at a normal detection location and performing noise control based on the estimated value, when the detection location of vibration fails, Noise is reduced because noise estimates are calculated before and after the vibration detection location fails and the correlation value of the noise estimation values before and after the vibration detection location fails is greater than or equal to a predetermined value. Noise control without impairing the effect can be performed. Further, according to this method, the noise before and after detecting the detection location where the vibration is not normally detected is stored, so that noise estimation before and after detecting the detection location where the vibration is not normally detected can be stored. The correlation value of values can be easily calculated.

The invention according to claim 18 is the noise control method according to claim 16 or 17 , wherein the filter processing is set based on a correlation between the vibration detected at the normal detection location and the vehicle interior noise. The weighting process is performed such that the weighting process coefficient is set so that the vibration having a higher correlation with the vehicle interior noise has a higher contribution to the estimated value of the vehicle interior noise. It is a feature.

  According to this method, the weighting process is performed so that the vibration having a higher correlation with the vehicle interior noise has a higher contribution to the estimated value of the vehicle interior noise, so that the vehicle interior noise can be estimated with high accuracy.

The invention according to claim 19 is the noise control method according to claim 18 , characterized in that the weighting process is performed for each predetermined frequency band.

  According to this method, the weighting process is performed for each predetermined frequency band, and the vehicle interior noise can be estimated with high accuracy.

The invention according to claim 20 is the noise control method according to claim 18 or 19 , wherein the filter processing includes the vibration detected at the normal detection position, the vibration and an estimated value of the vehicle interior noise. The transfer function G between the two is multiplied by the weighting coefficient.

  According to this method, the vehicle interior noise can be accurately estimated from the vibration of the vehicle body and the transfer function G.

The invention according to claim 21 is the noise control method according to any one of claims 18 to 20 , wherein the weighting processing coefficients corresponding to all combinations of the normal detection locations are calculated in advance. The coefficient of the weighting process is stored.

  According to this method, the processing time can be shortened by storing the weighting processing coefficients in advance.

The invention according to claim 22 is the noise control method according to claim 16 or 17 , wherein the filter processing is performed by detecting the vibration detected at the normal detection location and the body of the vibration and the vibration of the vehicle body. Multiplying the function (H ^T · H) ⁻¹ · H ^T based on the transfer function H between the vibration source and the transfer function R between the vehicle body vibration source and the estimated value of the vehicle interior noise It is a feature.

  According to this method, the filter processing is calculated based on the transfer function H, and the vehicle interior noise can be estimated with high accuracy.

The invention according to claim 23 is the noise control method according to claim 22 , wherein the functions (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal detection points are calculated in advance. The calculated function (H ^T · H) ⁻¹ · H ^T is stored.

  According to this method, the storage capacity can be reduced and the storage capacity can be reduced.

The invention according to claim 24 is the noise control method according to claim 22 , wherein the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal detection points is calculated. The transfer function H to be used is calculated in advance, and the calculated transfer function H is stored.

  According to this method, the storage capacity can be reduced and the storage capacity can be reduced.

The invention according to claim 25 is the noise control method according to claim 22 , wherein the transfer function H corresponding to the case where all the detection points are normal is stored, and the vibration is normal among the plurality of detection points. When a detection location that has not been detected is detected, a row component of the transfer function H corresponding to the detection location is deleted, and a new transfer function H is calculated.

  According to this method, the transfer function H can be easily calculated.

According to a twenty-sixth aspect of the present invention, in the noise control method according to any one of the twenty-second to twenty-fifth aspects, the number of vehicle body vibration sources is the number of wheels provided in the vehicle.

  According to this method, the number of wheels is the number of vehicle body vibration sources, so that the noise in the passenger compartment can be estimated with high accuracy.

According to a twenty-seventh aspect of the present invention, in the noise control method according to any one of the sixteenth to twenty-sixth aspects, the correlation value is calculated in advance according to the position of the predetermined space and the number of normal detection points. The noise control is performed based on the correlation value when the detection location where the vibration is not normally detected among the detection locations is detected.

  According to this method, since the correlation value of the estimated value of noise before and after detection of the detected portion where the vibration is not normally detected is calculated and stored in advance, the detected portion where the vibration is not normally detected is detected. At this time, by referring to the corresponding correlation value, it is possible to easily determine whether to perform noise control.

The invention according to claim 28 is the noise control method according to any one of claims 16 to 27 , wherein the noise control is performed based on the correlation values for a plurality of the predetermined spaces. .

  According to this method, it is possible to perform noise control without impairing the noise reduction effect on a plurality of spaces when a detection location where vibration is not normally detected is detected.

The invention according to claim 29 is the noise control method according to any one of claims 16 to 28 , wherein the detection location where the vibration is not normally detected based on the magnitude and detection time of the vibration is detected. It is characterized by detecting.

  According to this method, it is possible to easily detect a detection location where vibration is not normally detected based on the magnitude of vibration and the detection time.

Further, the invention of claim 30 is the noise control method according to any one of claims 16 to 29 , wherein the wave is a vehicle body vibration applied to the vehicle body or a sound generated in the vehicle interior of the vehicle body. It is characterized by being.

  According to this method, vehicle interior noise can be reduced by applying vibration or sound to the vehicle body.

  As mentioned above, although the Example of this invention was explained in full detail with drawing, an Example is only an illustration of this invention and this invention is not limited only to the structure of an Example. Accordingly, it is a matter of course that the present invention includes any design change within a range not departing from the gist of the present invention.

  For example, the number of acceleration sensors 10 is not limited to four, and a necessary number can be set.

  The number of piezoelectric actuators 20 is not limited to two, and an arbitrary number smaller than the number of acceleration sensors 10 can be set.

  Further, the feedback control performed by the calculation unit 35 is not limited to the H∞ control, and any feedback control may be used.

  In the above description, the application example to the case of road noise reduction has been described. However, the present invention can also be applied to other noises. For example, when used for engine noise, the same estimation is performed by pasting the acceleration sensor 10 on a floor panel and a dash panel (not shown). If the acceleration sensor fails, the noise estimation unit 34 similarly filters 50. Make a new setting. For wind noise, the same estimation is performed by attaching the acceleration sensor 10 to the front pillar (not shown) and the roof (not shown). If the acceleration sensor fails, the noise estimation unit 34 similarly sets the filter 50. Make new settings.

  In the present embodiment and examples, a signal line that is fed back to the control unit main body 32 and is returned to the control command value that is an output of the control unit main body 32 is output from the control unit main body 32. The control command value is subjected to D / A conversion and then A / D converted again and input to the control unit main body 32. Therefore, the control command value input to the control unit main body 32 is delayed by one processing cycle. Therefore, another form of deleting the feedback of the control command value in the control unit 30 and forming the feedback inside the control unit main body 32 may be used.

  Furthermore, in the present embodiment, the noise estimation unit 34, the calculation unit 35, and the acceleration sensor failure detection unit 90 are configured separately, but the noise estimation unit 34 and the acceleration sensor failure detection unit 90 are configured inside the calculation unit 35. However, noise estimation and control command value calculation can be performed simultaneously.

It is a figure which shows the main propagation path of the vibration of a vehicle body by the influence of the unevenness | corrugation of a road surface, and road noise. 1 is a schematic diagram of a noise control device in an embodiment of the present invention. It is a block diagram which shows the structure inside a control part main body. It is a flowchart of the process in a control part. It is a block diagram of the noise estimation part in embodiment of this invention. It is the block diagram simplified in order to demonstrate operation | movement of the noise estimation part shown in FIG. It is a flowchart of the process in the noise estimation part when all the acceleration sensors are normal. It is a figure which shows the relationship between the vibration source to a vehicle body, an acceleration signal, and vehicle interior noise. It is a flowchart of a filter calculation process. It is a figure which shows the frequency response example of the filter by embodiment of this invention. It is a figure which shows the measured value of noise, and the estimated value calculated in the noise estimation part. It is a figure which shows the measured value of the noise when an acceleration sensor fails, and the estimated value calculated by the noise estimation part. It is the block diagram simplified in order to demonstrate operation | movement of the noise estimation part shown in FIG. It is a flowchart of the process in a noise estimation part. It is a figure which shows the frequency response example of the filter when an acceleration sensor fails. It is a figure which shows the measured value of noise, and the estimated value calculated in the noise estimation part. It is a graph showing the relationship between the number of normal acceleration sensors and the estimation accuracy of noise. It is another graph showing the relationship between the number of normal acceleration sensors and noise estimation accuracy. It is a flowchart of an embodiment of the present invention. 3 is a flowchart of the first embodiment. In Example 1, it is explanatory drawing of the acceleration signal which an acceleration signal memory | storage part memorize | stores. 6 is a block diagram of Embodiment 2. FIG. It is a figure explaining control judgment table CDT. It is a flowchart which shows the setting method of the CDT value in the control determination table | surface CDT. 10 is a flowchart of Example 2. 10 is a flowchart of new setting of a filter according to the third embodiment. 10 is a flowchart of new setting of a filter in the fourth embodiment. 10 is a flowchart of new setting of a filter according to the fifth embodiment.

Explanation of symbols

10, 10a, 10b, 10c, 10d Acceleration sensor (sensor)
20, 20a, 20b Piezo actuator (wave application unit)
30 control unit 31 amplifying unit 32 control unit main body 33 A / D conversion unit 34 noise estimation unit 35 calculation unit 36 D / A conversion unit 50 filter 60 transfer function 70 addition unit 80 acceleration signal storage unit (output signal storage unit)
81, 82 Noise control execution determination unit 83 CDT storage units 90, 90a, 90b, 90c, 90d Acceleration sensor failure detection unit (sensor failure detection unit)
95 Filter setting unit 96 Storage unit 100 Control space 110 Floor panel 120 Axle 130 Suspension 140 Member 200 Tire

Claims (30)

A plurality of sensors arranged on the vehicle body for detecting vibrations of the vehicle body;
A wave applying unit for applying a controlled wave to the vehicle body;
A sensor failure detection unit that detects a failure of the sensor based on each output signal output by the plurality of sensors, and the output signal of the normal sensor excluding the sensor in which the failure is detected by the sensor failure detection unit A noise estimation unit that calculates an estimated value of vehicle interior noise that can be heard in a predetermined space in the vehicle interior of the vehicle, and the sensor failure detection unit detects a sensor failure when the sensor failure is detected by the sensor failure detection unit. A correlation value between the estimated value of the vehicle interior noise before detection of the vehicle interior noise and the estimated value of the vehicle interior noise after the failure of the sensor is detected, and when the correlation value is greater than or equal to a predetermined value, A control unit that performs noise control to reduce the vehicle interior noise by outputting a wave to the wave application unit based on an estimated value of the vehicle interior noise, and a predetermined value before the sensor failure detection unit detects the sensor failure Minutes And an output signal storage unit for storing said output signal,
The noise estimation unit performs a predetermined filter process set based on the correlation between the output signal of the normal sensor and the vehicle interior noise on the output signal, and the output signal subjected to the filter process An estimated value of the vehicle interior noise based on the sum of
When the sensor failure detection unit detects a failure of the sensor, the control unit is configured to output the output signal stored in the output signal storage unit and the sensor failure detection unit before the sensor failure is detected. Based on the filter processing, the sensor failure detection unit calculates an estimated value of the vehicle interior noise before detecting the sensor failure, and a failure is detected from the output signals stored in the output signal storage unit. After the sensor failure detection unit detects the failure of the sensor based on the output signal excluding the output signal of the sensor that has been detected and the filter processing after the sensor failure detection unit detects the failure of the sensor A noise control apparatus that calculates an estimated value of the vehicle interior noise and performs the noise control based on a correlation value calculated from the two estimated values of the vehicle interior noise . A plurality of sensors arranged on the vehicle body for detecting vibrations of the vehicle body;
A wave applying unit for applying a controlled wave to the vehicle body;
A sensor failure detection unit that detects a failure of the sensor based on each output signal output by the plurality of sensors, and the output signal of the normal sensor excluding the sensor in which the failure is detected by the sensor failure detection unit A noise estimation unit that calculates an estimated value of vehicle interior noise that can be heard in a predetermined space in the vehicle interior of the vehicle, and the sensor failure detection unit detects a sensor failure when the sensor failure is detected by the sensor failure detection unit. A correlation value between the estimated value of the vehicle interior noise before detection of the vehicle interior noise and the estimated value of the vehicle interior noise after the failure of the sensor is detected, and when the correlation value is greater than or equal to a predetermined value, A control unit that performs noise control to reduce the vehicle interior noise by outputting a wave to the wave application unit based on the estimated value of the vehicle interior noise, and updates and stores the latest output signal for a predetermined time. Output signal storage Equipped with a,
The noise estimation unit performs a predetermined filter process set based on the correlation between the output signal of the normal sensor and the vehicle interior noise on the output signal, and the output signal subjected to the filter process An estimated value of the vehicle interior noise based on the sum of
When the sensor failure detection unit detects a failure of the sensor, the control unit stops updating the storage of the output signal storage unit, and the sensor out of the output signals stored in the output signal storage unit The sensor failure detection unit detects the sensor failure based on the output signal before the failure detection unit detects the sensor failure and the filter processing before the sensor failure detection unit detects the sensor failure. The estimated value of the vehicle interior noise before detection is calculated, and the output signal stored in the output signal storage unit is the output signal before the sensor failure detection unit detects the sensor failure. The sensor signal based on the output signal excluding the output signal of the sensor that has detected the failure and the filter processing after the sensor failure detection unit has detected the sensor failure. And wherein the detection unit calculates the estimated value of the vehicle interior noise after the detection of a failure of the sensor, performs the noise control based on the correlation value calculated from the estimated value of the two said compartment noise Noise control device. The filtering process performs a weighting process set based on a correlation between the output signal of the normal sensor and the vehicle interior noise, and the weighting process is performed for the output signal having a higher correlation with the vehicle interior noise. The noise control apparatus according to claim 1 or 2, wherein the coefficient of the weighting process is set so that the degree of contribution to the estimated value of the vehicle interior noise is high.   The noise control apparatus according to claim 3, wherein the weighting process is performed for each predetermined frequency band.   The filter process is characterized by multiplying the output signal of the normal sensor, a transfer function G between the output signal and the estimated value of the vehicle interior noise, and a coefficient of the weighting process. Item 5. The noise control device according to Item 3 or 4.   The said noise estimation part is provided with the memory | storage part which calculates beforehand the coefficient of the said weighting process corresponding to all the combinations of the said normal sensor, and memorize | stores the calculated coefficient of the said weighting process. The noise control device according to any one of 5. The filter processing is a function (H ^T · H) ⁻¹ · H based on the normal output signal of the sensor and a transfer function H between the output signal and a vehicle body vibration source that is a source of vibration of the vehicle body. The noise control device according to claim 1 or 2, wherein ^T is multiplied by a transfer function R between the vehicle body vibration source and the estimated value of the vehicle interior noise. The noise estimation unit calculates the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal sensors in advance, and calculates the calculated function (H ^T · H) ⁻¹ · H ^T The noise control apparatus according to claim 7, further comprising a storage unit that stores The noise estimation unit calculates in advance the transfer function H used for calculating the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal sensors, and the calculated transfer function The noise control apparatus according to claim 7, further comprising a storage unit that stores H.   The noise estimation unit includes a storage unit that stores the transfer function H corresponding to the case where the sensor failure detection unit has not detected the sensor failure, and the sensor failure detection unit has detected the sensor failure. The noise control apparatus according to claim 7, wherein a new transfer function H is calculated by deleting a row component of the transfer function H corresponding to the sensor in which a failure is detected.   The noise control device according to any one of claims 7 to 10, wherein the number of vehicle body vibration sources is the number of wheels provided in the vehicle.   A correlation value storage unit in which the correlation value is calculated and stored in advance according to the position of the predetermined space and the number of normal sensors is provided, and the control unit is configured so that the sensor failure detection unit detects failure of the sensor. The noise control apparatus according to any one of claims 1 to 11, wherein, when detected, the noise control is performed based on the correlation value stored in the correlation value storage unit. The noise control apparatus according to any one of claims 1 to 12 , wherein the noise control is performed based on the correlation values for a plurality of the predetermined spaces. It said sensor failure detecting section, the noise control device according to any one of claims 1 to 13, characterized in that for detecting a failure of the sensor based on the output value and the output time of the output signal. The noise control device according to any one of claims 1 to 14 , wherein the wave is a vehicle body vibration applied to the vehicle body or a sound generated in the vehicle interior of the vehicle body. A first step of detecting vehicle body vibrations at a plurality of detection points;
A second step of detecting the detection point where the vibration is not normally detected among the plurality of detection points;
An estimated value of vehicle interior noise that can be heard in a predetermined space in the vehicle interior of the vehicle based on the vibration from the normal detection locations excluding the detection locations where the vibration is not normally detected among the plurality of detection locations. A third step of calculating;
When the detection location where the vibration is not normally detected among the plurality of detection locations is detected, the estimated value of the vehicle interior noise before detecting the detection location and the detection location after detecting the detection location A correlation value with the estimated value of the vehicle interior noise is calculated, and when the correlation value is equal to or greater than a predetermined value, a controlled wave is applied to the vehicle body based on the estimated value of the vehicle interior noise to reduce the vehicle interior noise. A fourth step for performing noise control,
A fifth step of storing the vibration for a predetermined time before detecting the detection point where the vibration is not normally detected among the plurality of detection points, and
In the third step, the predetermined filter processing set based on the correlation between the vibration from the normal detection location and the vehicle interior noise is performed on the vibration, and the vibration subjected to the filter processing is performed. An estimated value of the vehicle interior noise based on the sum of
In the fourth step, when the detection location where the vibration is not normally detected among the plurality of detection locations is detected, the stored vibration and the detection location where the vibration is not normally detected are detected. Based on the filter processing before detection, an estimated value of the vehicle interior noise before detecting the detection location where the vibration is not normally detected is calculated, and the vibration among the stored vibrations is normal The vibration is not normally detected based on the vibration excluding the vibration at the detection point that is not detected and the filter process after the detection point is not detected normally. calculating an estimated value of the vehicle interior noise after the detection of the detection portion, to perform the noise control based on the correlation value calculated from the estimated value of the two said compartment noise Noise control method according to claim. A first step of detecting vehicle body vibrations at a plurality of detection points;
A second step of detecting the detection point where the vibration is not normally detected among the plurality of detection points;
An estimated value of vehicle interior noise that can be heard in a predetermined space in the vehicle interior of the vehicle based on the vibration from the normal detection locations excluding the detection locations where the vibration is not normally detected among the plurality of detection locations. A third step of calculating;
When the detection location where the vibration is not normally detected among the plurality of detection locations is detected, the estimated value of the vehicle interior noise before detecting the detection location and the detection location after detecting the detection location A correlation value with the estimated value of the vehicle interior noise is calculated, and when the correlation value is equal to or greater than a predetermined value, a controlled wave is applied to the vehicle body based on the estimated value of the vehicle interior noise to reduce the vehicle interior noise. A fourth step for performing noise control,
And updating and storing the latest vibration for a predetermined time,
In the third step, the predetermined filter processing set based on the correlation between the vibration from the normal detection location and the vehicle interior noise is performed on the vibration, and the vibration subjected to the filter processing is performed. An estimated value of the vehicle interior noise based on the sum of
In the fourth step, when the detection location where the vibration is not normally detected among the plurality of detection locations is detected, update of the storage of the vibration is stopped, and the vibration among the stored vibrations is detected. The vibration can be normally detected based on the vibration before detecting the detection point where the detection is not normally detected and the filtering process before detecting the detection point where the vibration is not normally detected. And calculating an estimated value of the vehicle interior noise before detecting the detected location, and the vibration before detecting the detected location where the vibration is not normally detected among the stored vibrations, Based on the vibration excluding the vibration at the detection location where vibration is not normally detected and the filtering process after detecting the detection location where the vibration is not normally detected. And calculating an estimated value of the vehicle interior noise after detecting the detected portion where the vibration is not normally detected, and performing the noise control based on a correlation value calculated from the two estimated values of the vehicle interior noise The noise control method characterized by performing . The filtering process performs a weighting process that is set based on a correlation between the vibration detected at the normal detection location and the vehicle interior noise, and the weighting process has a high correlation with the vehicle interior noise. The noise control method according to claim 16 or 17 , wherein the coefficient of the weighting process is set so that the degree of contribution of the vehicle interior noise to the estimated value increases. The noise control method according to claim 18 , wherein the weighting process is performed for each predetermined frequency band. The filtering process is characterized by multiplying the vibration detected at the normal detection location, a transfer function G between the vibration and the estimated value of the vehicle interior noise, and a coefficient of the weighting process. The noise control method according to claim 18 or 19 . Previously calculated coefficients of the weighting processing corresponding to all combinations of normal the detection point, it stores the calculated coefficient of the weighting process to any one of claims 18 to 20, wherein The noise control method described. The filter processing is a function (H ^T · H) ⁻¹ · H based on a transfer function H between the vibration detected at the normal detection location and a vibration source of the vehicle body that is a source of vibration of the vehicle body. The noise control method according to claim 16 or 17 , wherein ^T is multiplied by a transfer function R between the vehicle body vibration source and the estimated value of the vehicle interior noise. The function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal detection points is calculated in advance, and the calculated function (H ^T · H) ⁻¹ · H ^T is stored. The noise control method according to claim 22 . The transfer function H used to calculate the function (H ^T · H) ⁻¹ · H ^T corresponding to all combinations of the normal detection points is calculated in advance, and the calculated transfer function H is stored. The noise control method according to claim 22 , wherein: The transfer function H corresponding to the case where all the detection points are normal is stored, and the detection point corresponding to the detection point is detected when the detection point where the vibration is not normally detected is detected among the plurality of detection points. 23. The noise control method according to claim 22 , wherein a new transfer function H is calculated by deleting a row component of the transfer function H. The noise control method according to any one of claims 22 to 25 , wherein the number of the vehicle body vibration sources is the number of wheels provided in the vehicle. The correlation value is calculated and stored in advance according to the position of the predetermined space and the number of normal detection points, and the detection points where the vibration is not normally detected among the plurality of detection points are detected. The noise control method according to any one of claims 16 to 26 , wherein the noise control is performed based on the correlation value. Noise control method according to any one of claims 16 to 27, characterized in that the noise control on the basis of the correlation values for a plurality of said predetermined space. The noise control method according to any one of claims 16 to 28 , wherein a detection location where the vibration is not normally detected is detected based on the magnitude and detection time of the vibration. The wave is noise control method according to any one of claims 16-29, characterized in that a sound generator to the passenger compartment of the vehicle body vibration or the vehicle body is applied to the vehicle body.

Images