JPH07191760A - Active type vibration controller and active type noise controller - Google Patents

Abstract

PURPOSE:To evade growth to the real divergence of control. CONSTITUTION:In the case that the period of vibration is judged to be constant in a step 201, and in addition, the possibility of the divergence of an adaptive digital filter is judged to be high in the steps 202, 203, the step is shifted to the step 204, and the zero cross point ZP of the intersection point (size 0) of waveform drawn by each filter coefficient of the adapative digital filter and a straight line is determined. Then, in the case that the number of pieces of the zero cross points ZP is judged to be '2' in the step 205, the step is shifted to the step 206, and a positive-to-negative zero cross point ZP. at which a filter coefficient shifts from positive to negative and a negative-to-positive zero cross point ZP2 at which the filter coefficient shifts from negative to positive are set. Then, in the case that the number of pieces of the zero cross points becomes over '3' afterward, the filter coefficient of negative polarity among the filter coefficients in a positive area and the filter coefficient of positive polarity among the filter coefficients in a negative area are set forcedly to '0'.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an active vibration control device for reducing vibrations by interfering control vibrations with periodic vibrations emitted from a vibration source such as an engine and propagating in a vehicle body, and noises of the engine and the like. The present invention relates to an active noise control device for reducing noise by interfering a control sound with periodic noise transmitted from a power source to a vehicle compartment, etc., and particularly, by filtering a reference signal representing a vibration or noise generation state. Active vibration control device and active noise control device provided with an adaptive digital filter for generating a drive signal for driving a control sound source or a controlled vibration source, and means for sequentially updating each filter coefficient of the adaptive digital filter according to an adaptive algorithm. In the above, the control divergence is suppressed and stable vibration reduction control or noise reduction control is executed.

[0002]

2. Description of the Related Art As a conventional technique for reducing vibration, there is, for example, a device described in Japanese Utility Model Publication No. 1-41952, and such a conventional vibration reducing device is a device for reducing vibration of a vehicle body. In brief, by appropriately controlling the phase of the control signal for the vibration exciter fixed to the vehicle body, it is possible to input an exciting force having a phase opposite to the vibration to the vehicle body from the vibration exciter. , The vibration was canceled by the excitation force to reduce the vibration of the vehicle body.

However, this conventional vibration reducing device has a structure in which a sinusoidal control signal is applied to the vibration exciter so that a vibration force having a phase opposite to that of the vibration is simply generated. If the characteristics or the like change due to deterioration of each component, for example, there is a high possibility that a sufficient vibration reducing effect cannot be obtained, and in some cases, the vibration may be worsened.

As a conventional technique capable of coping with such a defect, there is, for example, an active noise control device described on pages 515-516 of "Acoustic Society of Japan 1992 Spring Research Presentation Lecture Collection". The conventional technology is synchronous Fi
It is an active noise control device to which the tered-X LMS algorithm is applied. That is, the LMS algorithm is one of adaptive algorithms. For example, when the LMS algorithm is applied to a noise reduction device, a reference signal indicating the noise generation state is taken in and the reference signal is filtered by an adaptive digital filter. While processing and generating a signal for driving a control sound source, a residual noise signal representing a noise reduction state is taken in, and each filter coefficient of the adaptive digital filter is sequentially updated according to the LMS algorithm based on the residual noise signal and the reference signal. Will be configured. By using such an adaptive algorithm, even if the characteristics of the control system are unknown or the characteristics of the control system fluctuate from the initial state, the control sound that can reduce noise can be obtained. Can be generated from the control sound source.

It should be noted that the LMS algorithm disclosed in the above-mentioned collection of papers is particularly the synchronous Filtered-X LMS.
It is called an algorithm, which is advantageous in reducing periodic vibrations and noise, and is characterized in that an impulse train synchronized with the fundamental frequency of noise is applied. That is, since the reference signal is an impulse train, multiplication is not necessary and convolution calculation can be performed only by addition. In some cases, addition is also unnecessary. Therefore, there is an advantage that the amount of calculation is significantly reduced and the processing can be performed at high speed. is there.

[0006]

However, even in the conventional device using the LMS algorithm as described above, the filter coefficient of the adaptive digital filter becomes unstable when some disturbance input occurs from the stable control state. If you don't have any measures to avoid
There is a problem that the control diverges and the vibration and noise increase on the contrary. Therefore, conventionally, for example, when the level of the residual vibration signal or the residual noise signal is higher than the level before the control is started, it is determined that the control has diverged and the control itself is interrupted. But with this,
This is not an effective solution because the vibration reduction control and noise reduction control itself stop.

The present invention has been made by paying attention to the unsolved problems of the above-mentioned conventional techniques, and in particular, in order to reduce periodic vibrations and noises, it is expected that the control will diverge. It is an object of the present invention to provide an active vibration control device and an active noise control device that can be prevented.

[0008]

In order to achieve the above object, the invention according to claim 1 is a control vibration source capable of generating a control vibration that interferes with a periodic vibration generated from a vibration source, and Reference signal generation means for detecting the vibration occurrence state of the vibration source and outputting it as a reference signal, residual vibration detection means for detecting the vibration after the interference and outputting it as a residual vibration signal, and the control by filtering the reference signal. An adaptive digital filter for generating a driving signal for driving a vibration source, and an adaptive digital filter for updating a filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the vibration after the interference based on the reference signal and the residual vibration signal. An active vibration control device including a processing means, and determines whether or not the adaptive digital filter is in a stable state with no tendency to diverge. The constant state determination means, a predetermined threshold value that is a value in the middle of or in the vicinity of the range that the filter coefficient of the adaptive digital filter can take in the stable state, and the magnitude of each filter coefficient of the adaptive digital filter. A magnitude relationship determining means for determining a relationship; a storage means for storing the determination result of the magnitude relationship determining means when the stable state determining means determines that the stable state is in correspondence with each filter coefficient; And a filter coefficient compulsory setting means for forcibly setting the corresponding filter coefficient to the threshold value when the judgment result of the relation judging means and the stored contents of the storage means are compared and the two do not match. It is a thing.

In order to achieve the above object, the invention according to claim 2 is a control vibration source capable of generating a control vibration that interferes with a periodic vibration generated from the vibration source, and a vibration of the vibration source. Reference signal generating means for detecting a generation state and outputting as a reference signal, residual vibration detecting means for detecting the vibration after the interference and outputting as a residual vibration signal, and driving the control vibration source by filtering the reference signal. An adaptive digital filter for generating a driving signal, and adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the vibration after the interference based on the reference signal and the residual vibration signal, In an active vibration control device equipped with, a maximum and minimum value setting for obtaining the maximum and minimum values of each filter coefficient of the adaptive digital filter. Means, a monotonous decrease determining means for judging whether or not the value of each filter coefficient from the maximum value to the minimum value of the adaptive digital filter is monotonically decreasing, and the minimum value of the adaptive digital filter to the maximum value. When the value of each filter coefficient is increasing monotonically while approaching the value, a monotonous increase judging means for judging whether or not the value of each filter coefficient is monotonically increasing, and when the monotonic decreasing judgment means judges that the value does not monotonically decrease, of the adaptive digital filter When the filter coefficient is forcibly set to monotonically decrease from the maximum value toward the minimum value and the monotonic increase determination means determines that the monotonic increase has not occurred, the minimum value of the adaptive digital filter. To a maximum value, the filter coefficient forcibly setting means for forcibly setting the filter coefficient to monotonically increase.

In order to achieve the above object, the invention according to claim 3 is a control vibration source capable of generating a control vibration that interferes with a periodic vibration generated from the vibration source, and a vibration of the vibration source. Reference signal generating means for detecting a generation state and outputting as a reference signal, residual vibration detecting means for detecting the vibration after the interference and outputting as a residual vibration signal, and driving the control vibration source by filtering the reference signal. An adaptive digital filter for generating a driving signal, and adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the vibration after the interference based on the reference signal and the residual vibration signal, In an active vibration control device equipped with, a maximum and minimum value setting for obtaining maximum and minimum values of each filter coefficient of the adaptive digital filter Means and whether or not the value of each filter coefficient between the maximum value and the minimum value of the adaptive digital filter monotonically decreases, or the value of each filter coefficient between the minimum value and the maximum value monotonically increases. And a monotonic change determination means for determining at least one of the above, and when the monotonic change determination means determines that neither the monotonic decrease nor the monotonic increase is performed, the filter coefficient of the corresponding section of the adaptive digital filter is set. And a filter coefficient forced setting means for forcibly setting to monotonically decrease or monotonically increase.

An active vibration control device according to a fourth aspect of the present invention is the above-mentioned first, second or third aspect.
In the invention according to, the cycle detecting means for detecting the cycle of the periodic vibration, the cycle determining means for determining whether the cycle of the vibration detected by the cycle detecting means is a predetermined cycle or more, and the cycle A first processable state control unit that sets the filter coefficient forced setting unit to a process executable state when the determination unit determines that the vibration cycle is equal to or longer than the predetermined cycle.

Further, an active vibration control device according to a fifth aspect of the invention is the active vibration control device according to the first, second, third or fourth aspect of the invention, wherein the amplitude of the adaptive digital filter is a predetermined width. Second amplitude processing for determining whether or not the above, and second processing for setting the filter coefficient forced setting means in a process executable state when the amplitude determination means determines that the amplitude is equal to or larger than the predetermined width And a possible state control means.

On the other hand, in order to achieve the above object, the invention according to claim 6 is such that a control sound source capable of generating a control sound that interferes with the periodic noise emitted from the noise source, and a noise generation state of the noise source. A reference signal generating means for detecting and outputting as a reference signal, a residual noise detecting means for detecting the noise after interference and outputting as a residual noise signal, and a drive signal for filtering the reference signal to drive the control sound source. And an adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the noise after the interference based on the reference signal and the residual noise signal. In the active noise control device, a stable state determining means for determining whether or not the adaptive digital filter is in a stable state with no divergence tendency. The magnitude relationship between the predetermined threshold value, which is a value in the middle of or in the vicinity of the range of the filter coefficient of the adaptive digital filter in the stable state and each filter coefficient of the adaptive digital filter, is determined. Relationship determination means, storage means for storing the determination result of the magnitude relationship determination means in association with each filter coefficient when the stable state determination means determines that the stable state is determined, and the determination of the magnitude relationship determination means A filter coefficient forcing setting means for forcibly setting the corresponding filter coefficient to the threshold value is provided when the result is compared with the stored contents of the storage means and if the two do not match.

In order to achieve the above object, the invention according to claim 7 is a control sound source capable of generating a control sound that interferes with periodic noise emitted from a noise source, and a noise generation state of the noise source. And a residual noise detecting means for detecting the noise after interference and outputting it as a residual noise signal, and a drive signal for filtering the reference signal to drive the control sound source. And an adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the noise after the interference based on the reference signal and the residual noise signal. In the active noise control device, maximum value / minimum value setting means for obtaining the maximum value and the minimum value of each filter coefficient of the adaptive digital filter. , A monotonous decrease determining means for judging whether or not the value of each filter coefficient between the maximum value and the minimum value of the adaptive digital filter is monotonically decreasing, and the minimum value to the maximum value of the adaptive digital filter. When the value of each filter coefficient is increasing monotonically, the monotonous increase determining means determines whether or not the monotone decreasing determination means determines that the monotonic decrease does not decrease the maximum value of the adaptive digital filter. When the filter coefficient is forcibly set to monotonically decrease from the value toward the minimum value and the monotonic increase determination means determines that the monotonic increase does not occur, the adaptive digital filter determines the maximum value to the maximum value. And a filter coefficient compulsory setting means for compulsorily setting the filter coefficient so as to monotonically increase while going toward the value.

In order to achieve the above object, the invention according to claim 8 is a control sound source capable of generating a control sound interfering with periodic noise emitted from a noise source, and a noise generation state of the noise source. A reference signal generating means for detecting and outputting as a reference signal, a residual noise detecting means for detecting the noise after interference and outputting as a residual noise signal, and a drive signal for filtering the reference signal to drive the control sound source. And an adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the noise after the interference based on the reference signal and the residual noise signal. In the active noise control device, maximum value / minimum value setting means for obtaining the maximum value and the minimum value of each filter coefficient of the adaptive digital filter. , Whether or not the value of each filter coefficient from the maximum value to the minimum value of the adaptive digital filter monotonically decreases, or the value of each filter coefficient from the minimum value to the maximum value monotonically increases. Whether or not there is a monotone change determining means, and if the monotone change determining means determines that the monotone decrease or monotonic increase has not occurred, the filter coefficient in the corresponding section of the adaptive digital filter is monotonically decreased. Or a filter coefficient forced setting means for forcibly setting so as to increase monotonically,
It is equipped with.

The active noise control device according to claim 9 is the above-mentioned claim 6, claim 7 or claim 8.
In the invention according to, the cycle detecting means for detecting the cycle of the periodic noise, the cycle determining means for determining whether the cycle of the noise detected by the cycle detecting means is a predetermined cycle or more, and the cycle A first processable state control unit that sets the filter coefficient forced setting unit to a process executable state when the determination unit determines that the noise period is equal to or longer than the predetermined period.

Further, the active noise control device according to claim 10 is the above-mentioned claim 6, claim 7, and claim 8.
Alternatively, in the invention according to claim 9, when the amplitude determination means determines whether the amplitude of the adaptive digital filter is equal to or larger than a predetermined width, and the amplitude determination means determines that the amplitude is equal to or larger than the predetermined width. Second processable state control means for setting the filter coefficient forced setting means into a process executable state.

[0018]

Now, let us consider specifically the case where the periodic vibration is reduced by using an adaptive algorithm such as the LMS algorithm. For example, considering the case where the LMS algorithm is applied to a so-called active engine mount that cancels the periodic vibration generated in the vehicle engine before it is transmitted to the vehicle body side, a signal synchronized with the rotation of the engine crankshaft is used as a reference. Signal, vibration of the engine mount near the mounting point on the vehicle body side is used as a residual vibration signal, and the reference signal is filtered by an adaptive digital filter to generate a drive signal for the engine mount as a control vibration source. LMS based on residual vibration signal
The algorithm will be executed to update each filter coefficient of the adaptive digital filter.

In the situation where the control does not tend to diverge, the vibrations to be canceled are periodic.
As shown in (a), each filter coefficient of the adaptive digital filter draws a sinusoidal locus having the same period as the vibration. However, the adaptive digital filter is a finite impulse response (FIR) type digital filter.

However, particularly when the engine speed is low or idling, the residual vibration signal has a frequency (20) of periodic vibration generated in the engine due to the influence of mechanical resonance of the controlled object. ˜30 Hz), it tends to diverge at higher frequencies, and when each filter coefficient of the adaptive digital filter is updated based on the residual vibration signal having the divergence tendency, as shown in FIG. In addition to the frequency of the reference signal, high-order frequency components appear in the digital filter, and the control tends to diverge. If left as it is, the high-order frequency components become excessive as shown in FIG. Will diverge. For example, since mechanical elastic resonance of a general member to which an engine mount is attached exists at 100 to 200 Hz, control is particularly performed in a situation where such a frequency becomes a high-order frequency. It is more likely to diverge.

On the other hand, according to the first aspect of the present invention, whether or not the stable state judging means is in a state in which the adaptive digital filter is not in a diverging tendency (a state as shown in FIG. 9A). On the other hand, the magnitude relationship determining means determines an intermediate value in the range in which the value of the filter coefficient of the adaptive digital filter can be taken in the stable state (specifically, a sine wave shape drawn by the filter coefficient of the adaptive digital filter). Is an intermediate value between the maximum value and the minimum value of the waveform.
For example, when the filter coefficient swings both positive and negative across 0, the value is “0”. ) Or a predetermined threshold value which is a value in the vicinity thereof, and the magnitude relationship between each filter coefficient of the adaptive digital filter is determined, and the determination result of the magnitude relationship determining means is stored in the storage means. That is, in the example of FIG. 9A, whether each filter coefficient is above or below the threshold value (more specifically, the polarity of each filter coefficient) is large or small. It will be judged by the relationship judging means.

Even if the adaptive digital filter is not in a stable state, the magnitude relationship determining means determines the magnitude relationship, and the determination result and the stored contents of the storage means are compared by the filter coefficient forcing setting means. If the result of the comparison indicates that the two do not match, the filter coefficient forcing setting unit forcibly sets the non-matching filter coefficient to the threshold value.

That is, assuming that the adaptive digital filter has a divergent tendency and the situation as shown in FIG. 9B is reached, the magnitude relationship in the case of FIG. 9A is stored in the storage means. 9A and 9B, the filter coefficients having different polarities are forcibly set to the size "0". As a result, as shown in FIG.
Even in the situation as shown in FIG. 9B, the higher-order frequency component does not grow any more, and therefore the divergent state as shown in FIG. 9C is avoided.

On the other hand, in the invention according to claim 2, the maximum value / minimum value judging means judges the maximum value and the minimum value of each filter coefficient of the adaptive digital filter. Since the drawn waveform is sinusoidal, each filter coefficient monotonically decreases from the maximum value to the minimum value and monotonically increases from the minimum value to the maximum value in the non-divergent state.

The monotonous decrease means that the relationship between the i-th filter coefficient W _i and the i + 1-th filter coefficient W _{i + 1} is always W _i ≧ W _{i + 1} , and the monotonous increase means It means that the relation of these filter coefficients is always W _i ≦ W _{i + 1} . Then, the section that should be monotonically decreasing by the monotonic decrease determination means (that is, from the maximum value to the minimum value)
It is determined whether or not the monotonic increase is maintained, and whether or not the monotonic increase is maintained in the section where the monotonic increase is determined (that is, from the minimum value to the maximum value) by the monotonic increase determination means. Is determined,
Each filter coefficient of the adaptive digital filter is shown in FIG.
In the normal state as shown in (1), since the monotonous decrease and the monotonic increase are maintained, the filter coefficient forcible setting unit does not particularly set the filter coefficient forcibly.

However, for example, when the divergence tends to be as shown in FIG. 9 (b), high-order frequency components appear in the waveform, so that W is increased from the maximum value to the minimum value.
There is a portion where _i <W _{i + 1,} and there is a portion where W _i > W _{i + 1} even though it is from the minimum value to the maximum value. Then, the filter coefficient forced setting means forcibly sets the filter coefficient so as to monotonically decrease in the section where the monotonic decrease cannot be maintained, and monotonically sets in the section where the monotonic increase cannot be maintained. Since the filter coefficient is forcibly set so as to increase, even if the situation as shown in FIG. 9B is reached, the higher-order frequency component does not grow any more, and therefore,
The divergent state as shown in FIG. 9C is avoided.

The setting of the filter coefficient in the filter coefficient forced setting means is not particularly limited as long as a monotonous decrease and a monotonous increase can be maintained as a result. For example, a filter coefficient that is contrary to a monotonic decrease or monotonic increase may be forcibly set to the same value as the previous filter coefficient, or such a filter coefficient may be set to the next previous filter coefficient. You may make it set to the average value with the filter coefficient of.

Here, in the invention according to claim 2, it is determined whether or not both the monotonous decrease and the monotonic increase are maintained, and if at least one is not maintained, the filter coefficient compulsory setting is performed. Although the means is configured to forcibly set the filter coefficient as described above, at least one of a monotone decrease and a monotone increase is set as the determination target for the reason of, for example, reducing the calculation load, and the like. If the monotonic change that was the target of the judgment is not maintained,
Even if the filter coefficient is forcibly set so that the monotonic change is maintained, some effect is expected.

Therefore, in the invention according to claim 3,
The monotonic change determination means determines at least one of whether it is monotonically decreasing from the maximum value to the minimum value or monotonically increasing from the minimum value to the maximum value. If it is determined that the change is not maintained, the filter coefficient compulsory setting means forcibly sets the filter coefficient of the corresponding section of the adaptive digital filter so as to monotonically decrease or monotonically increase.
At least one of the monotonous decrease and the monotonic increase is maintained, and the divergent state as shown in FIG. 9C is avoided.

The invention according to claim 3 (claim 8)
Also in the invention according to the above), the "corresponding section" in which the filter coefficient compulsory setting means forcibly sets the filter coefficient is a section determined by the content of the judgment in the monotonic change judgment means, If it is determined whether or not is monotonically decreasing, the corresponding section is between the maximum value and the minimum value of the filter coefficient, and the monotonic change determination means determines whether or not monotonically increasing. If so, the corresponding section is a section from the minimum value to the maximum value of the filter coefficient.

According to the invention of claim 4,
When the cycle of the vibration is detected by the cycle detecting means, it is judged by the cycle judging means whether or not the cycle of the vibration is a predetermined cycle or more, and when it is judged that the cycle of the vibration is the predetermined cycle or more. The first processable state control unit brings the filter coefficient forced setting unit into the process executable state.

That is, since the filter coefficient forced setting means in the invention according to claim 1, claim 2 or claim 3 can be processed only when the period of vibration is long to a certain extent, the residual vibration signal is controlled. When the cycle of vibration that is less likely to diverge at higher frequencies of the frequency of vibration due to the influence of the mechanical resonance of the object is short, the processing necessary for the determination by the filter coefficient forced setting means (for example, claim 1 In the invention according to claim 2, the magnitude relationship determination processing in the magnitude relationship determination means, and in the invention according to claim 2, the monotonic decrease determination processing in the monotone reduction determination means, etc.) are not required to be executed.

Further, in the invention according to claim 5,
The amplitude determining means determines whether or not the amplitude of the adaptive digital filter is greater than or equal to a predetermined width, and when it is determined that the amplitude is greater than or equal to the predetermined width, the second processable state control means forces the filter coefficient. The setting means is ready for processing. That is, claim 1, claim 2, only when it is determined that the output is large enough to easily diverge.
Since the filter coefficient forced setting means in the invention according to claim 3 or 4 is set in a processable state, the filter coefficient forced setting is performed especially in a situation where the amplitude of the adaptive digital filter is small and control is unlikely to diverge. It becomes unnecessary to execute the processing required for the determination by the means.

Here, while all the inventions according to claims 1 to 5 are directed to vibration, claim 6
The invention according to claim 10 is directed to noise. Therefore, the operation of the invention according to claims 5 to 8 is substantially the same as that of the above-mentioned claim 1 although there is a difference between vibration and sound.
It is the same as the invention according to claim 4. That is, claim 6
In the invention according to the above aspect, the stable state determining means determines whether or not the adaptive digital filter is in a state in which there is no tendency to diverge, while the magnitude relationship determining means determines the filter of the adaptive digital filter when in the stable state. The magnitude relationship between a predetermined threshold value, which is an intermediate value in the range of possible coefficient values or a value in the vicinity thereof, and each filter coefficient of the adaptive digital filter is determined, and the determination result of the magnitude relationship determining means is stored. Stored in the means.

Even when the adaptive digital filter is not in a stable state, the magnitude relationship determining means determines the magnitude relationship, and the determination result and the stored contents of the storage means are compared by the filter coefficient forcing setting means. If the result of the comparison indicates that they do not match, the filter coefficient forced setting means forcibly sets the filter coefficient that does not match to the above threshold value.

As a result, even if the adaptive digital filter tends to diverge, the higher-order frequency components do not grow any more, and it is possible to avoid a full-scale divergent state. In the invention according to claim 7,
The maximum value / minimum value determining means determines the maximum value and the minimum value of each filter coefficient of the adaptive digital filter, and the monotonic decrease determining means determines whether or not the monotonic decrease is maintained in the section which should be the monotonic decrease. At the same time, the monotonic increase determination means determines whether or not the monotonic increase is maintained in the section that should be a monotonic increase.

Therefore, when the adaptive digital filter tends to diverge, higher-order frequency components appear in the waveform.
_Since there is a portion where _i <W _{i + 1} and there is a portion where W _i > W _{i + 1} even during the period from the minimum value to the maximum value, the filter coefficient is forced. For the section where the monotonous decrease cannot be maintained, the setting means
The filter coefficient is forcibly set to monotonically decrease, and the filter coefficient is forcibly set to monotonically increase in the section where the monotonic increase cannot be maintained, so higher-order frequency components grow further. There is no need to do so, and it is possible to avoid reaching a full-scale divergent state.

According to the invention of claim 8,
The monotonic change determination means determines whether or not the monotonic decrease or monotonic increase is maintained, and when it is determined that such a monotonic change is not maintained, the filter coefficient compulsory setting means causes the filter coefficient of the corresponding section. Is forcibly set to monotonically decrease or monotonically increase, so that at least one of monotonic decrease or monotonic increase is maintained, and a full-scale divergence state is avoided.

Further, in the invention according to claim 9 or claim 10, the processing for suppressing the divergence as described above is executed only when the possibility of divergence is high.

[0040]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing an embodiment of the present invention. In this embodiment, a so-called active engine mount for actively reducing vibration transmitted from an engine to a vehicle body is provided in an active vibration control device according to the present invention. It has been applied to.

First, the structure will be described. As shown in FIG. 1, the engine mount 1 is integrally provided with a mounting bolt 2a for mounting on an engine 30 as a vibration source, and has a hollow inside and a lower part. The mounting member 2 has an opening, and the upper end of the inner cylinder 3 is caulked to the outer surface of the lower portion of the mounting member 2. Inside the inner cylinder 3, a diaphragm 4 is disposed so as to vertically divide the space inside the mounting member 2 and the inner cylinder 3 into two parts, which are sandwiched by the caulking prevention portions of the mounting member 2 and the inner cylinder 3. Of the space divided by the diaphragm 4, the space above the diaphragm 4 communicates with the atmospheric pressure, and the space below the diaphragm 4 is provided with the orifice structure 5.

On the other hand, on the outer peripheral surface of the inner cylinder 3, the inner peripheral surface of the cylindrical support elastic body 6 which is formed such that the axial positions of the inner peripheral surface and the outer peripheral surface are higher on the inner peripheral side is vulcanized. The outer peripheral surface of the supporting elastic body 6 is vulcanized and adhered to the inner peripheral surface of the outer cylinder 7. The lower end of the outer cylinder 7 is caulked to the upper part of the cylindrical actuator holding member 8, and the lower end surface of the actuator holding member 8 is provided with a mounting bolt 9a for mounting on the member 35 side. A disc-shaped mounting member 9 provided integrally is fixed.

A cylindrical member 10 fixed to the lower end portion of the outer cylinder 7 by caulking is fixed to the upper end surface of the actuator holding member 8, and the inner peripheral surface of the cylindrical member 10 is further fixed. The movable member 12 is held by the cylindrical elastic body 11 so as to be vertically displaceable with a predetermined clearance from the upper end surface of the actuator holding member 8. The movable member 12 is made of a magnetizable material and is formed into a disk shape having a concave upper surface.

Inside the actuator holding member 8, there is provided an electromagnetic actuator 13 including an electromagnetic coil and the like, which displaces the movable member 12 in the vertical direction according to a control signal supplied from the outside. There is. Further, in this embodiment, the lower surface of the support elastic body 6 and the movable member 1 are
2, a main fluid chamber 15 is formed in a portion defined by the upper surface of the second fluid, and a sub fluid chamber 16 is formed in a portion defined by the diaphragm 4 and the orifice structure 5. The chambers 16 communicate with each other via the orifice 5a formed in the orifice structure 5. A fluid such as oil is enclosed in the main fluid chamber 15, the sub fluid chamber 16 and the orifice 5a.

The characteristic of the fluid mount determined by the shape of the flow path of the orifice 5a is that when the engine shake occurs during traveling, that is, when the engine mount 1 is 5 to 15 Hz.
Is adjusted so that it exhibits a high dynamic spring constant and high damping force when is excited. The electromagnetic actuator 13 is connected to the controller 20 and generates a predetermined electromagnetic force according to the drive signal y supplied from the controller 20.

The controller 20 includes a microcomputer, a necessary interface circuit, an A / D converter, and a D / A.
It is composed of a converter, an amplifier, etc., and is a vibration in a frequency band in which a fluid cannot move between the main fluid chamber 15 and the sub-fluid chamber 16 through the orifice 5a, that is, a vibration of a higher frequency than the engine shake described above. When idle vibration, muffled sound vibration, or vibration during acceleration is input, the transmission force to the mounting member 9 becomes “0” in synchronization with the vibration (specifically, the support elastic body 6 The drive signal y is generated and supplied to the electromagnetic actuator 13 so that the input due to the elastic deformation and the input due to the volume change of the main fluid chamber 15 can be canceled out.

Here, idle vibration and muffled sound vibration are
For example, in the case of a reciprocating 4-cylinder engine, the engine rotation 2
The main cause is that the engine vibration of the next component is transmitted to the member 35 via the engine mount 1. Therefore, if the drive signal y is generated and output in synchronization with the secondary component of the engine rotation, the vibration transmissibility is increased. Can be reduced. Therefore, in the present embodiment, an impulse signal that is synchronized with the rotation of the crank angle of the engine 30 (for example, in the case of a reciprocating four-cylinder engine, one impulse is generated every 180 degrees of rotation of the crank) is generated and output as the reference signal x. A pulse signal generator 21 as a reference signal generating means is provided, and the reference signal x is supplied to the controller 20 as a signal indicating a vibration generation state in the engine 30.

On the other hand, the member 35 is provided with an acceleration sensor 22 as a residual vibration detecting means for detecting the vibration state of the member 35 in the form of acceleration and outputting it as a residual vibration signal e in the vicinity of the mounting position of the engine mount 1. The residual vibration signal e, which is fixed, is supplied to the controller 20 as a signal representing the vibration after the interference. Then, the controller 20 is based on the reference signal x and the residual vibration signal e, and is a Filtered-X LMS algorithm which is one of the adaptive algorithms of the successive update type.
More specifically, the synchronous Filtered-X LMS
The drive signal y is generated and output according to the algorithm.

That is, the controller 20 has an adaptive digital filter W having a variable filter coefficient W _i (i = 0, 1, 2, ..., I-1: I is the number of taps), and the latest reference signal x Is output from the engine 30 via the engine mount 1 at a predetermined sampling clock interval and the filter coefficient W _i of the adaptive digital filter W is sequentially output as the drive signal y.
A process for appropriately updating the filter coefficient W _i of the adaptive digital filter W is executed based on the reference signal x and the residual vibration signal e so that the vibration transmitted to the device is reduced.

The update formula of the adaptive digital filter W is F
The following equation (1) follows the iltered-X LMS algorithm. _{W i (n + 1) =} W i (n) -αR T e (n) ...... (1) where, indicates that a value at term time n stick is (n), also, alpha is a convergence factor It is a coefficient that is called and is related to the speed of convergence of the filter coefficient W _i and its stability. R ^T is theoretically a value (reference signal or Filtered-X signal) obtained by filtering the reference signal x with a transfer function filter C ^ representing the transfer function C between the engine mount 1 and the acceleration sensor 22. In this embodiment, since the reference signal x is an impulse train as a result of applying the synchronized Filtered-X LMS algorithm, those impulses generated when the impulse response of the transfer function filter C ^ is generated one after another in synchronization with the reference signal x. It matches the sum of the response waveforms at time n.

Further, theoretically, the reference signal x is filtered by the adaptive digital filter W to generate the drive signal y, and the filtering process corresponds to the convolutional calculation in the digital calculation. Since it is an impulse train, even if each filter coefficient W _i of the adaptive digital filter W is sequentially output as the drive signal y at a predetermined sampling clock interval from the time when the latest reference signal x is input as described above. , The same result as when the filtering result is the drive signal y.

Further, the controller 20 determines whether or not the vibration cycle is fluctuating, determines whether or not the vibration cycle is equal to or longer than a predetermined cycle, and the adaptive digital filter. A process for determining whether or not the amplitude of W is equal to or greater than a predetermined width is performed, and the cycle of vibration is stable by each of these processes, and the cycle of vibration is equal to or greater than the predetermined cycle. Moreover, when it is determined that the amplitude of the adaptive digital filter W is equal to or greater than the predetermined amplitude, the point where the waveform drawn by each filter coefficient W _i in one cycle of the adaptive digital filter W and the straight line of the amplitude 0 intersect (hereinafter , Referred to as zero cross point ZP.)
When the number of zero cross points ZP is 2, the process of determining that the adaptive digital filter W is in a stable state with no divergence tendency is executed.

When the controller 20 determines that the adaptive digital filter W is in a stable state with no divergence tendency because the number of zero cross points ZP is 2, the controller 20 uses these two zero cross points ZP as a reference. Each filter coefficient W of the adaptive digital filter W
It is configured to determine whether each _i is in a positive region having a size of 0 or more or a filter region having a size of less than 0, and store the determination result. Specifically, all the filter coefficients W _i are 0 or more in one region sandwiched by the zero cross points ZP, and the zero cross points ZP
Since all the filter coefficients W _i are 0 or less in the other region sandwiched by, the two zero cross points Z _i
It is determined whether P is a point from positive to negative or a point from negative to positive, and based on those determination results, a zero cross point (hereinafter referred to as positive / negative zero cross point ZP ₁₎ from positive to negative. .) And a zero-cross point that shifts from negative to positive (hereinafter, negative-positive zero-cross point Z
It is called P ₂ . ) And are set and stored respectively.

On the other hand, when it is determined that the number of zero cross points ZP is 3 or more, the controller 20 does not execute the process of storing the positive and negative zero cross points ZP ₁ and the negative positive zero cross points ZP ₂ . For each filter coefficient W _i of the adaptive digital filter W, it is determined whether or not the filter coefficient W _i between the negative positive zero cross point ZP ₂ and the positive negative zero cross point ZP ₁ is a negative value. However, when it is negative, the filter coefficient W _i is forcibly set to the size “0”, and the filter coefficient W _i is between the positive and negative zero cross points ZP ₁ and ZP _2. to determine also whether it is a positive value regardless, if is positive executes processing for setting its filter coefficient W _i forcibly size "0" It has become the jar.

Here, in the present embodiment, "0", which is an intermediate value of the values of the filter coefficients W _i of the adaptive digital filter W in the stable state, is set as a predetermined threshold value. There is. Next, the operation of this embodiment will be described.
That is, when an engine shake occurs, the flow path shape of the orifice 5a is appropriately selected. As a result, the engine mount 1 functions as a support device for high dynamic spring constant and high damping force. Is damped by the engine mount 1, and the vibration level on the member 35 side is reduced. In such a case, it is not particularly necessary to displace the movable member 12.

On the other hand, when the vibration in the frequency higher than the idle vibration frequency is input, which makes the fluid in the orifice 5a stick and the fluid cannot move between the main fluid chamber 15 and the sub-fluid 16, 20 executes a predetermined arithmetic processing, outputs a drive signal y to the electromagnetic actuator 13, and causes the engine mount 1 to generate an active control force capable of reducing vibration.

This will be specifically described with reference to FIG. 2, which is a flow chart showing the outline of the processing executed in the controller 20 when the idle vibration and the muffled sound vibration are input.
First, after predetermined initialization is performed in step 101, the process proceeds to step 102, and the reference signal R ^T is calculated based on the transfer function filter C ^.
In this step 102, the reference signals R ^{T for} one cycle are collectively calculated.

Then, in step 103, the counter i is cleared to zero, and then in step 104, the i-th filter coefficient W _{i of} the adaptive digital filter W is output as the drive signal y. When the drive signal y is output in step 104, the process proceeds to step 105, the residual vibration signal e is read, the counter j is cleared to zero in step 106, and then the process proceeds to step 107, where the j-th filter of the adaptive digital filter W. The coefficient W _j is updated according to the above equation (1).

When the updating process in step 107 is completed, the process proceeds to step 108, and a divergence suppressing process is executed to suppress each filter coefficient W _j of the adaptive digital filter W before it is in a diverging state. The specific operation when this divergence suppressing process is executed will be described later. When the process of step 108 is completed, the process proceeds to step 109, and it is determined whether or not the next reference signal x is input. If it is determined that the reference signal x is not input, the adaptive digital filter is used. In order to update the filter coefficient next to W or output the drive signal y, the process proceeds to step 110.

In step 110, it is determined whether or not the counter j has reached the output count T _y (more precisely, since the counter j starts from 0, a value obtained by subtracting 1 from the output count T _y ). In this determination, it is determined whether or not the filter coefficient W _i of the adaptive digital filter W is output as the drive signal y in step 104, and then the filter coefficient of the adaptive digital filter W is updated by the necessary number as the drive signal y. It is for. Therefore, if the determination in step 110 is “NO”, the counter j is incremented in step 111, and then step 107 is performed.
Then, the above-mentioned processing is repeated.

However, the determination in step 110 is "YE
In the case of “S”, it can be determined that the updating process of the necessary number of filter coefficients as the drive signal y among the filter coefficients of the adaptive digital filter W has been completed.
Move to 2. Then, after the counter i is incremented in step 112, it waits until the time corresponding to the interval of the predetermined sampling clock has elapsed after executing the processing of step 104, and when the time corresponding to the sampling clock has elapsed. , Step 104 above
Then, the above-mentioned processing is repeated.

However, when it is determined in step 109 that the reference signal x is input, the process proceeds to step 113, and the counter i (to be precise, since the counter i starts from 0, 1 is added to the counter i). Value) is stored as the latest output count T _y , and then the process returns to step 102 and the above-described processing is repeatedly executed. As a result of repeatedly executing such processing, as shown in FIG. 4, which represents the relationship among the reference signal x, the drive signal y, and the transfer function filter C ^,
From the controller 20 to the engine mount 1,
From the time when the reference signal x is input, the filter coefficients W _i of the adaptive digital filter W are sequentially supplied as the drive signal y at the sampling clock interval. Each filter coefficient W _i of the adaptive digital filter W is Synchronous F
Since it is sequentially updated by the above (1) according to the iltered-X LMS algorithm, each filter coefficient W of the adaptive digital filter W is passed after a certain amount of time has passed.
_{After i} has converged to the optimum value, the drive signal y is supplied to the engine mount 1 to reduce idle vibration and muffled sound vibration transmitted from the engine 30 to the member 35 side via the engine mount 1. Like The control force in the engine mount 1 is controlled by the electromagnetic force generated from the electromagnetic actuator 13 to move the movable member 1.
2 vibrates, and the vibration is obtained by acting as a force between the inner cylinder 3 and the outer cylinder 7 via the fluid in the main fluid chamber 15 and the expansion spring of the support elastic body 6.

On the other hand, the specific flow of the divergence suppressing process in step 108 is as shown in FIG. 3, and first in step 201, the change rate Δ of the output count T _y.
It is determined whether T _y is substantially zero. In step 201, for example, during the past predetermined time (e.g., 1 second), when the output times T _y is not changed, the change rate [Delta] T _y of output times T _y is substantially zero determination To be done.

If the determination in step 201 is "NO", it can be determined that each filter coefficient W _i of the adaptive digital filter W is changing so as to follow the periodic change of vibration, and will be described later. Such a filter coefficient W _i
However, since the problem that the convergence of the filter coefficient W _{i to} the optimum value is hindered occurs when the setting of the above is performed, the processing of step 202 and thereafter is not executed, and the processing of FIG. The process will return to 2.

However, the determination in step 201 is "YE
In the case of "S", since such a problem does not occur, the process proceeds to step 202. In step 202, it is determined whether the output count T _y is larger than the predetermined value β. Here, if the sampling clock is fixed, the output count T _y directly represents the period of vibration,
The determination in step 202 is to determine whether or not the vibration cycle is longer than the predetermined cycle.

Then, when the vibration cycle is long, as shown in FIG.
Since the possibility of shifting to the divergence tendency as shown in (b) is high, it is considered that the possibility of shifting to the divergence tendency is extremely small especially when the vibration cycle is short. The process of FIG. 3 is terminated and the process returns to the process of FIG. 2 without executing the process. The predetermined value β
The size of is not particularly limited, and actually, FIG.
The cycle of vibrations in which divergence as shown in (c) is likely to be obtained experimentally, and the calculated cycle is divided by the interval of the sampling clock.

On the other hand, the determination in step 202 is "YES".
In the case of, the process proceeds to step 203 and the maximum value W of the filter coefficient W _i representing the amplitude of the adaptive digital filter W is
_It is determined whether _MAX is larger than the predetermined value γ. Here, in a situation in which the value of the filter coefficient W _i of the adaptive digital filter W is excessively large, it can be determined that a large output in which control is easily diverged is made. It is considered that the forced setting process of the coefficient W _i is necessary, so that the processes from step 204 onward are executed. Therefore, when the determination in step 203 is "NO", the processing of step 204 and subsequent steps is not executed, and the processing of FIG. 3 is terminated and the processing returns to the processing of FIG.
When there is a divergence tendency, the filter coefficient W _i of the adaptive digital filter W swings in the positive and negative directions in substantially the same manner. It is determined whether or not the amplitude of is greater than or equal to a predetermined width.

Then, the determination in step 203 is "YE
In the case of “S”, it is determined that there is a high possibility that a divergence tendency will occur in the near future, the process proceeds to step 204, and each filter coefficient W _i
The position where the polarity of is reversed is searched and the zero cross point ZP is set, and then the routine proceeds to step 205, where it is judged whether the number of the set zero cross points ZP is two or not.

When the processes of these steps 204 and 205 are first executed, it means that the judgments in steps 202 and 203 are both "YES" for the first time.
The control should not be divergent yet. If the control does not tend to diverge, the filter coefficients W _i of the adaptive digital filter W have periodic vibrations, and therefore, as shown in FIG. The waveform drawn by W _i is sinusoidal, and its cycle matches the cycle of vibration, that is, the input cycle of the reference signal x.
Is two.

That is, at least when the process of step 205 is executed for the first time, the determination is "YES", and therefore the process proceeds to step 206. And
In step 206, of the two existing zero cross points ZP, a zero cross point that shifts from a region where the filter coefficient W _i has a positive value to a region where the filter coefficient W _{i has} a negative value is searched, and the searched zero cross point is a positive / negative zero cross. The zero cross point which is stored as the point ZP _{1 and} which moves from the negative region to the positive region on the contrary is searched, and the searched zero cross point is stored as the negative positive zero cross point ZP ₂ . Incidentally, in the example of FIG. 5A, if the position of the zero cross point is represented by the filter coefficient immediately before the zero cross point ZP, the filter coefficient W ₅ is positive and the filter coefficient W ₆ is negative. Since the filter coefficient W ₁₂ is negative and the filter coefficient W ₁₃ is positive, the value of the positive / negative zero cross point ZP ₁ becomes “5” and the value of the negative positive zero cross point ZP ₂ becomes “13”.

When the process of step 206 is completed, the process of FIG. 3 is terminated and the process of FIG. 2 is resumed.
Then, if the processing of FIG. 2 is repeatedly executed in such a situation, as described with reference to FIGS. 9A to 9C, the higher order frequency component of the engine vibration appears in the residual vibration signal e. Therefore, the influence thereof is expressed in each filter coefficient W _i of the adaptive digital filter W. Then, the sinusoidal waveform as shown in FIG. 5A drawn by each filter coefficient W _i of the adaptive digital filter W collapses,
Since the waveform of the higher-order frequency component appears at a certain point as shown in FIG. 5B, in the process of step 204,
Two or more zero cross points ZP will be set.

Therefore, the determination in step 205 is "N
Since it becomes "O", the routine does not proceed to step 206 but proceeds to step 207. In step 207, whether or not the negative filter coefficient W _i exists in the positive region, that is, in the region before the positive / negative zero cross point ZP ₁ and in the region after the negative / positive zero cross point ZP ₂ is determined. To be judged. Specifically, if the positive and negative zero cross points ZP ₁ and the negative positive zero cross points ZP ₂ are set to the values as described above, one of the filter coefficients W _{0 to} W ₅ and W ₁₃ that has a negative value is found. It is determined whether or not it exists.

When it is determined by the processing of step 207 that the filter coefficient W _i having a negative value exists in the positive region, the process proceeds to step 208, and the value of the filter coefficient W _i is changed. It is forcibly set to a predetermined threshold value "0". For example, in the example of FIG.
Since the filter coefficients W ₄ at step 207 is determined to be applicable, the filter coefficient W ₄ Step 20
It will be set to "0" at 8.

Then, the process proceeds to step 209, this time within the negative region, that is, the positive / negative zero cross point ZP ₁
It is determined whether or not the positive filter coefficient W _i is present in the region after and the region before the negative / positive zero cross point ZP ₂ . Specifically, if the positive and negative zero cross points ZP ₁ and the negative positive zero cross points ZP ₂ are set to the values as described above, the filter coefficients W _{6 to} W _{12 are set.}
It is determined whether or not there is a positive value in.

When it is determined by the processing of step 209 that the filter coefficient W _i having a positive value exists in the negative region, the process proceeds to step 210, and the value of the filter coefficient W _i is changed. It is forcibly set to a predetermined threshold value "0". For example, in the example of FIG.
Since the filter coefficients W ₇ is determined to be appropriate in step 209, the filter coefficient W ₇ Step 21
At 0, it will be set to "0".

When the processing of 210 is completed or when the determination of step 209 is "NO", at that time, the process shown in FIG.
The process of step 1 is ended and the process returns to the process of FIG. That is,
When the divergence suppressing process as shown in FIG. 3 is executed, the filter coefficient W _i which tends to diverge is forcibly set to “0”, and thus further divergence is suppressed,
Thus, it is possible to avoid a full-scale divergent state as shown in FIG. 9 (c).

That is, even if the divergence tendency shown in FIG. 5B is obtained, the filter coefficients W ₄ and W ₇ having the divergence tendency at this point are forcibly set to “0”. Since the state as shown in FIG. 5 (c) is reached, the divergence is suppressed at this point, the filter coefficient W _i of the adaptive digital filter W is appropriately updated thereafter, and the high level in the residual vibration signal e, which is the divergent factor, is increased. When the next frequency component becomes smaller, the normal state is restored again as shown in FIG.

As described above, according to this embodiment, it is possible to suppress the symptom before the control diverges at the time when the symptom appears. Therefore, for example, the vibration reduction control is stopped after the control diverges. Unlike the conventional solutions, it is a very effective solution that does not require stopping the vibration reduction control. Further, in this embodiment, the number of times of output T _y corresponding to the period of vibration is counted, and only when the number of times of output T _y is larger than the predetermined value β, step 203 is performed.
The subsequent processing is executed, and the filter coefficient W
The maximum value W _{MAX of i} is _calculated , and the maximum value W _MAX is a predetermined value γ.
Since the processing of step 204 and subsequent steps is executed only when the control value is larger than the above, eventually, the processing of step 204 and subsequent processing is executed only when the control is likely to diverge. Therefore, when the possibility that the control will diverge is extremely small, the processing after step 204 is not executed, so that there is an advantage that the increase of the calculation load is suppressed to the minimum.

[0079] Further, in this embodiment, each filter coefficient W _i of the adaptive digital filter W when it is determined that the stable state, the polarity of the magnitude relation that is, each filter coefficient W _i of the threshold "0" rather than individually stored vibration by paying attention to that a periodic vibration that varies sinusoidally, set the positive and negative zero-crossing point ZP ₁ and negative positive zero crossing point ZP _2, their positive and negative zero crossing point ZP ₁ Also, since the position of the negative and positive zero cross point ZP ₂ is stored, there is an advantage that the process for storing is simple and the storage capacity required for it is small. That is, with such a method, even if the number of taps (the number of filter coefficients) of the adaptive digital filter W increases, the positions of the two points of the positive / negative zero cross point ZP ₁ and the negative positive zero cross point ZP ₂ are stored. This is because the polarities of all the filter coefficients W _i are stored.

In this embodiment, since the case where each filter coefficient W _i of the adaptive digital filter W swings positively and negatively around the size "0", the predetermined threshold value is set to "0". However, the present invention is not limited to this, and, for example, in the case of a system in which it is clear that the vibration oscillates in the positive direction or the negative direction, the threshold value is set to 0 or more or 0 or less in accordance with the direction of the vibration. Alternatively, when the filter coefficient is composed of only positive values, the filter coefficient may be set to an intermediate value in the range that can be normally taken. However, it is not necessary to strictly set it to an intermediate value between the maximum value and the minimum value, and even if it is a value close to it, the same operational effect as that of the present embodiment can be obtained.

Here, in this embodiment, the engine mount 1 constitutes a controlled vibration source, and step 20
The stable state determination means is constituted by the processing of 4, 205, and the positive / negative zero cross point Z is determined in step 206.
The storage means is configured by the process of storing P ₁ and the negative positive zero cross point ZP _2, and the process of searching the positive and negative zero cross point ZP ₁ and the negative positive zero cross point ZP ₂ in step 206 and steps 207 and 20.
The processing of 9 constitutes a magnitude relation determining means, the processing of steps 208 and 210 constitutes a filter coefficient forced setting means, and steps 103, 112 and 113.
The cycle detecting means is constituted by the processing of
The process of step 2 constitutes the cycle determining means, and step 2
The first processable state control means is configured by the processing of skipping the processing of step 203 and subsequent steps when the determination of 02 is “NO”, the amplitude determination means is configured by the processing of step 203, and the determination of step 203 is performed. "N
In the case of "O", the second processable state control means is configured by the processing of skipping the processing of step 204 and subsequent steps, and the adaptive processing means is configured by the processing of step 107.

FIG. 6 is a diagram showing a second embodiment of the present invention. This second embodiment also applies the present invention to an engine mount similar to that of the first embodiment, and FIG. ,
4 is a flowchart showing a specific flow of divergence suppression processing, similar to FIG. 3 of the first embodiment. Since the other construction is the same as that of the first embodiment, its illustration and description are omitted.

That is, in this embodiment, the contents of the processing of steps 201 to 203 of the divergence suppression processing are the same as those of the first embodiment.
Same as the embodiment, but the determination in step 203 is “YE
In the case of "S", the process proceeds to step 301, where the maximum value W _MAX and the minimum value W _MIN of each filter coefficient W _i of the adaptive digital filter W are set.

Then, when the maximum value W _MAX and the minimum value W _MIN are set in step 301, the process _proceeds to step 302, in which each filter coefficient W _i is constantly maintained from the maximum value W _MAX to the minimum value W _MIN. It is determined whether or not the relationship “W _i ≧ W _{i + 1} ” is satisfied, that is, whether or not the value of each filter coefficient W _i is monotonically decreasing. Here, when the processing of step 301 is executed for the first time, step 202
Since both of the determinations of 203 and 203 are “YES” for the first time, the control should not yet have a diverging tendency.

Therefore, the waveform drawn by each filter coefficient W _i of the adaptive digital filter W should maintain the sine wave shape as shown in FIG. 7A, and therefore the maximum value W _MAX.
From the minimum value W _MIN to the filter coefficient W _{2 to} W
₉ should be monotonically decreasing. Therefore, at this time, the determination in step 302 is “YES”, and thus the process in step 303 is not executed and the process proceeds to step 304.

In step 304, this time, the minimum value W
_While going from _MIN to the maximum value W _MAX , each filter coefficient W _i
, Always satisfies the relation of "W _i ≤W _{i + 1} ", that is, whether the value of each filter coefficient W _i monotonically increases. Here, for the first time, step 303
Since the control is not yet in the diverging tendency even when the processing of (1) is executed, as shown in FIG.
Each filter coefficient W ₉ between _MIN and maximum value W _MAX
~W _13, W ₀ ~W ₄ should have a monotonically increasing.

Therefore, at this time, the determination in step 304 is "YES", so the processing in step 305 is not executed, the processing in FIG. 7 is terminated, and the same as in FIG. 2 described in the first embodiment. It will return to processing. Then, if the processing similar to that shown in FIG. 2 is repeatedly executed in such a situation, as described with reference to FIGS. 9A to 9C, the residual vibration signal e has a higher-order frequency component of the engine vibration. Appears, the influence appears in each filter coefficient W _i of the adaptive digital filter W. Then, each filter coefficient W _i of the adaptive digital filter W is drawn in FIG.
Since the sinusoidal waveform as shown in (a) collapses and a higher-order frequency component waveform appears as shown in FIG. 7 (b) at some point, the above-mentioned monotonous decrease and monotonic increase cannot be maintained. Will end up.

Then, for example, in the example of FIG. 7B,
Since there is a portion where “W _i <W _{i + 1} ”, there is a portion where “W _i <W _{i + 1} ”, from the maximum value W _MAX to the minimum value W _MIN , which is the section that should have decreased monotonically, so the determination in step 302 is “NO”. Then, the process proceeds to step 303. And
In step 303, the filter coefficient W _{i + 1} satisfying “W _i <W _{i + 1} ” is the previous filter coefficient W _i.
It is forced to the same value as _i , so that the maximum value W
Each filter coefficient W _i from _MAX to the minimum value W _MIN
The monotonic decrease of will be recovered. For example, FIG. 7 (b)
In the case of the example, since the monotonous decrease is broken between the portion of W ₄ <W _{5 and} the portion of W ₆ <W ₇ , the process proceeds from step 302 to step 303, and the filter coefficient W ₅ is changed to the filter coefficient W ₄ And the filter coefficient W ₇ is set to the same value as the filter coefficient W ₆ . Therefore, step 3
After performing the processing of 03, the adaptive digital filter W
Will draw a waveform as shown in FIG.

When the process of step 303 is completed, the process proceeds to step 304. Then, if it is assumed in step 304 that there is a portion where “W _i > W _{i + 1} ”, while going from the minimum value W _MIN , which should be a monotonically increasing section, to the maximum value W _MAX , step S Moved to 305,
The filter coefficient W _{i + 1 for} which "W _i > W _{i + 1} "
Is forcibly set to the same value as the previous filter coefficient W _i , whereby the minimum value W _MIN to the maximum value W _MAX.
The monotonic increase of each filter coefficient W _i is recovered while going to

As described above, according to the present embodiment, as shown in FIG.
When the divergence tendency as shown in (b) is reached, the filter coefficient W _{i + 1} which has a divergence tendency in which the monotonous decrease and the monotonic increase are collapsed.
Is forcibly set to the same value as the previous filter coefficient W _i, and the monotonous decrease and monotonic increase are recovered, so that further divergence is suppressed, and as shown in FIG. 9C. It is possible to avoid such a serious divergence.

That is, even if the divergence tendency is as shown in FIG. 7B, the filter coefficients W ₅ and W ₇ having the divergence tendency at this point are forcibly forced to the filter coefficients W one before each. _Since it is set to the same value as ₄ and W ₆ ,
Since the state as shown in (c) is reached, the divergence is suppressed at this point, the filter coefficient W _i of the adaptive digital filter W is appropriately updated thereafter, and the higher-order frequency in the residual vibration signal e that was the divergent factor When the component becomes smaller, the normal state is restored again as shown in FIG. 7 (a).
Thus, even in the present embodiment, as in the first embodiment, it is possible to suppress this at the time when the symptom appears before the control diverges. Unlike the conventional solution in which the reduction control is stopped, it is a very effective solution in which it is not necessary to stop the vibration reduction control. Other actions and effects are the same as the first
It is similar to the embodiment.

FIG. 8 shows the vibration reduction effect of this embodiment,
FIG. 7 is a frequency characteristic diagram showing a vibration reduction effect by the conventional active vibration control device that does not execute the divergence suppressing process as shown in FIG. 6, in which the horizontal axis represents frequency and the vertical axis represents vibration acceleration at the foot of the vehicle. I am taking a level. That is, in the conventional active vibration control device, a divergence tendency is seen especially in the sixth-order frequency of the fundamental frequency, and if the control is continued as it is, there is a high possibility that the control will diverge at a certain point. According to this, it is understood that the tendency of divergence at the sixth-order frequency of the fundamental frequency is suppressed without deteriorating the vibration level at other frequencies.

Here, in this embodiment, step 3
The processing of 01 constitutes the maximum and minimum value setting means,
The processing of step 302 constitutes a monotonic decrease determination means, the processing of step 304 constitutes a monotonic increase determination means, and the processing of steps 303 and 305 constitutes a filter coefficient compulsory setting means. In the second embodiment, when it is determined that the monotonous decrease or monotonic increase is not maintained, the corresponding filter coefficient W _{i + 1} is forced to the same value as the immediately previous filter coefficient W _i. However, the method of forcibly setting the filter coefficient to recover such a monotonic decrease or monotonous increase is not limited to this. . For example, for the corresponding filter coefficient W _{i + 1} , the average value (= (W _i + W _{i + 2} ) / of the filter coefficient W _i immediately before and the filter coefficient W _{i + 2 after} that
A method of forcibly setting to 2) may be used.

In the second embodiment, it is determined whether both the monotonous decrease and the monotonic increase are maintained, and if at least one is not maintained, the filter coefficient is forcibly set. However, for example, when there is a request to further suppress the increase in the calculation load, it is determined whether the monotonic decrease is maintained or whether the monotonic increase is maintained. According to the judgment result, the filter coefficient is compulsorily set so that the monotonic decrease or monotonic increase is maintained between the maximum value and the minimum value or between the minimum value and the maximum value. Good.

In each of the above embodiments, the case where the synchronous Filtered-X LMS algorithm is used as the adaptive algorithm has been described, but the applicable adaptive algorithm is not limited to this, and for example, a normal Filtered -X LMS algorithm etc. may be sufficient. In each of the above embodiments, the case where the active vibration control device according to the present invention is applied to a so-called active engine mount for reducing the vibration transmitted from the engine 30 of the vehicle to the member 35 has been described. The application target of the invention is not limited to this, and may be a device that reduces vibrations generated from other than the engine 30 as long as they are periodic vibrations. You may apply to the apparatus etc. which reduce the vibration transmitted to a floor.

Furthermore, the object of reduction is not limited to vibration, but may be, for example, an active noise control device for a vehicle that actively reduces periodic noise transmitted from the engine 30 into the vehicle interior. . In such a case, a loudspeaker (control sound source) that is driven by a drive signal supplied from the controller to generate a control sound in the vehicle compartment, and a microphone that detects residual noise in the vehicle compartment and outputs the residual noise signal to the controller ( Residual noise detecting means).

For example, as an adaptive algorithm, Filt
If the ered-X LMS algorithm is used, the reference signal x supplied from the pulse generator 21 is filtered by the adaptive digital filter W to generate the drive signal y for the loudspeaker, while the reference signal x is supplied to the microphone. Of the adaptive digital filter W by executing the Filtered-X LMS algorithm on the basis of the residual noise signal e supplied from the microphone and the value filtered by the transfer function filter C ^ that models the transfer function between the loudspeaker and the loudspeaker. Filter coefficient W
_i is sequentially updated, and after such update processing,
Alternatively, the divergence suppressing process as shown in FIG. 6 may be executed. Even if this noise is reduced, synchronous F
Of course, the iltered-X LMS algorithm can be applied.

[0098]

As described above, in the invention according to claim 1 or claim 6, the filter coefficient of the adaptive digital filter which tends to diverge is forcibly set to a predetermined threshold value. Therefore, since the divergence of the adaptive digital filter is suppressed, it is possible to prevent the adaptive digital filter from reaching a full-scale divergent state, that is, when the symptom appears before the control diverges, this is suppressed. Therefore, unlike the conventional solution in which the vibration reduction control is stopped after the control has diverged, there is an effect that it is not necessary to stop the vibration reduction control.

Further, in the invention according to claim 2 or claim 7, each filter coefficient of the adaptive digital filter is set to be a monotonically decreasing section and a monotonically increasing section, and monotonically decreasing in the monotonically decreasing section. If the monotonic increase collapses in a section that should be monotonically increasing, the monotonic decrease,
Since the filter coefficient of the adaptive digital filter is forcibly set so that the monotonous increase is restored, the divergence of the adaptive digital filter is suppressed as in the invention according to claim 1 or claim 5. It is possible to prevent the adaptive digital filter from reaching a full-scale divergence state, that is, when the symptom appears before the control diverges, this is suppressed, so for example, the vibration reduction control is stopped after the control diverges. Unlike the conventional solution that causes such a problem, there is an effect that it is not necessary to stop the vibration reduction control.

Further, in the invention according to claim 3 or claim 8, it is possible to obtain an effect similar to the effect of the invention according to claim 2 or claim 6 with a smaller increase in calculation load. The effect is that you can do it. Then, in the invention according to claim 4, claim 5, claim 9 or claim 10, since the process required for divergence suppression is executed only when the control is highly likely to diverge, There is an effect that an increase in calculation load can be minimized.

[Brief description of drawings]

FIG. 1 is a sectional view showing a configuration of a first embodiment of the present invention.

FIG. 2 is a flowchart showing an outline of processing executed in a controller.

FIG. 3 is a flowchart showing an outline of a divergence suppressing process executed in the controller of the first embodiment.

FIG. 4 is a waveform diagram showing a relationship among a reference signal, a drive signal, and a transfer function filter.

FIG. 5 is a waveform diagram for explaining the operation of the first embodiment.

FIG. 6 is a flowchart showing an outline of a divergence suppressing process executed in the controller of the second embodiment.

FIG. 7 is a waveform diagram for explaining the operation of the second embodiment.

FIG. 8 is a frequency characteristic diagram for explaining the effect of the second embodiment.

FIG. 9 is a waveform diagram for explaining the process of divergence of the adaptive digital filter.

[Explanation of symbols]

 1 Engine Mount (Controlled Vibration Source) 13 Electromagnetic Actuator 20 Controller 21 Pulse Generator (Reference Signal Generation Means) 22 Acceleration Sensor (Residual Vibration Detection Means) 30 Engine (Vibration Source) 35 Member x Reference Signal y Drive Signal e Residual Vibration Signal

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location G05B 13/02 S 7531-3H G10K 11/178

Claims (10)

[Claims] 1. A control vibration source capable of generating a control vibration that interferes with a periodic vibration emitted from the vibration source, and a reference signal generating means for detecting a vibration generation state of the vibration source and outputting the vibration as a reference signal. Residual vibration detecting means for detecting the vibration after the interference and outputting it as a residual vibration signal; an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the controlled vibration source; the reference signal and the An active vibration control device comprising: adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the vibration after the interference based on a residual vibration signal, wherein the adaptive digital filter diverges. Stable state determination means for determining whether or not the stable state is not prone, and the adaptive digital signal when in the stable state The filter coefficient of the filter has a magnitude relationship determining means for determining a magnitude relationship between a predetermined threshold value which is a value in the middle or in the vicinity thereof and each filter coefficient of the adaptive digital filter, and the stable state determining means. A storage unit that stores the determination result of the magnitude relationship determining unit in association with each filter coefficient when it is determined to be in the stable state, and the determination result of the magnitude relationship determining unit and the storage content of the storage unit are compared. If the two do not match, an active vibration control device is provided, which further comprises a filter coefficient forcing setting unit for forcibly setting the corresponding filter coefficient to the threshold value. 2. A control vibration source capable of generating a control vibration that interferes with a periodic vibration emitted from the vibration source, and a reference signal generating means for detecting a vibration generation state of the vibration source and outputting it as a reference signal. A residual vibration detecting means for detecting the vibration after the interference and outputting it as a residual vibration signal, an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the controlled vibration source, the reference signal and the An active vibration control device comprising: an adaptive processing unit that updates a filter coefficient of the adaptive digital filter according to an adaptive algorithm so that the vibration after the interference is reduced based on a residual vibration signal. Maximum value / minimum value setting means for obtaining the maximum value and the minimum value of the filter coefficient, and from the maximum value of the adaptive digital filter. A monotonous decrease determination means for determining whether or not the value of each filter coefficient decreases monotonically while going toward a small value, and a value of each filter coefficient between the minimum value and the maximum value of the adaptive digital filter monotonically decreases. A monotonous increase determining means for determining whether or not the filter coefficient increases, and a filter coefficient between the maximum value and the minimum value of the adaptive digital filter when the monotonic decrease determining means determines that the monotonic decrease does not occur. Is forcibly set to monotonically decrease and the monotonic increase determination means determines that the monotonic increase has not occurred, the filter coefficient between the minimum value and the maximum value of the adaptive digital filter increases monotonically. And a filter coefficient forcibly setting means for forcibly setting so that the active vibration control device. 3. A control vibration source capable of generating a control vibration that interferes with a periodic vibration emitted from the vibration source, and a reference signal generating means for detecting a vibration generation state of the vibration source and outputting it as a reference signal. Residual vibration detecting means for detecting the vibration after the interference and outputting it as a residual vibration signal; an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the controlled vibration source; the reference signal and the An active vibration control device comprising: an adaptive processing unit that updates a filter coefficient of the adaptive digital filter according to an adaptive algorithm so that the vibration after the interference is reduced based on a residual vibration signal. Maximum value / minimum value setting means for obtaining the maximum value and the minimum value of the filter coefficient, and from the maximum value of the adaptive digital filter. A monotone that determines at least one of whether the value of each filter coefficient decreases monotonically while moving toward a small value, or whether the value of each filter coefficient increases monotonically while moving from the minimum value to the maximum value When the change determination means and the monotonic change determination means determine that the monotonous decrease or monotonic increase has not occurred, the filter coefficient of the corresponding section of the adaptive digital filter is forcibly set to monotonically decrease or monotonically increase. An active vibration control device comprising: 4. A cycle detecting means for detecting a cycle of the periodic vibration, a cycle judging means for judging whether the cycle of the vibration detected by the cycle detecting means is a predetermined cycle or more, and the cycle. The first processable state control means for setting the filter coefficient forced setting means into a process executable state when the determination means determines that the cycle of the vibration is equal to or longer than the predetermined cycle. Item 5. The active vibration control device according to any one of items 3. 5. An amplitude determination means for determining whether the amplitude of the adaptive digital filter is equal to or larger than a predetermined width, and the filter coefficient when the amplitude determination means determines that the amplitude is equal to or larger than the predetermined width. The active type vibration control device according to any one of claims 1 to 4, further comprising a second processable state control unit that sets the forced setting unit to a process executable state. 6. A control sound source capable of generating a control sound that interferes with periodic noise emitted from a noise source, reference signal generating means for detecting a noise generation state of the noise source and outputting it as a reference signal, and the interference. Residual noise detecting means for detecting the subsequent noise and outputting it as a residual noise signal, an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the control sound source, the reference signal and the residual noise signal In the active noise control device, the adaptive digital filter is adapted to update the filter coefficient of the adaptive digital filter according to an adaptive algorithm so as to reduce the noise after the interference based on Stable state determination means for determining whether or not the stable state is present, and the adaptive digital filter when in the stable state. And a stable state determining means for determining a magnitude relationship between a predetermined threshold value that is a value in the middle of or in the vicinity of a range in which the filter coefficient can be taken and each filter coefficient of the adaptive digital filter. When it is determined to be in a stable state, a storage unit that stores the determination result of the magnitude relationship determining unit in association with each filter coefficient is compared with the determination result of the magnitude relationship determining unit and the storage content of the storage unit. An active noise control device comprising: a filter coefficient forcing setting unit for forcibly setting the corresponding filter coefficient to the threshold value when the two do not match. 7. A control sound source capable of generating a control sound that interferes with a periodic noise emitted from a noise source, a reference signal generating means for detecting a noise generation state of the noise source and outputting it as a reference signal, and the interference. Residual noise detecting means for detecting the subsequent noise and outputting it as a residual noise signal, an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the control sound source, the reference signal and the residual noise signal An active noise control device comprising: an adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so that the noise after the interference is reduced, Maximum value minimum value setting means for obtaining maximum value and minimum value, and the maximum value to minimum value of the adaptive digital filter The value of each filter coefficient is decreasing monotonically while it is decreasing, and the value of each filter coefficient is increasing monotonically from the minimum value to the maximum value of the adaptive digital filter. If the monotonous increase determination means determines whether or not the monotone decrease determination means determines that the monotonic decrease does not occur, the filter coefficient between the maximum value and the minimum value of the adaptive digital filter decreases monotonically. So that the filter coefficient is monotonically increased between the minimum value and the maximum value of the adaptive digital filter when the monotonic increase determination means determines that the monotonic increase has not occurred. An active noise control device comprising: a filter coefficient forcibly setting means for forcibly setting. 8. A control sound source capable of generating a control sound that interferes with periodic noise emitted from a noise source, reference signal generating means for detecting a noise generation state of the noise source and outputting it as a reference signal, and the interference. Residual noise detecting means for detecting the subsequent noise and outputting it as a residual noise signal, an adaptive digital filter for filtering the reference signal to generate a drive signal for driving the control sound source, the reference signal and the residual noise signal An active noise control device comprising: an adaptive processing means for updating the filter coefficient of the adaptive digital filter according to an adaptive algorithm so that the noise after the interference is reduced, Maximum value minimum value setting means for obtaining maximum value and minimum value, and the maximum value to minimum value of the adaptive digital filter A monotonic change determination means for determining at least one of whether the value of each filter coefficient is monotonically decreasing while moving toward the maximum value or whether the value of each filter coefficient is monotonically increasing while moving from the minimum value to the maximum value When the monotonic change determination means determines that the monotonic decrease or monotonic increase has not occurred, the filter coefficient forcibly setting the filter coefficient of the corresponding section of the adaptive digital filter to monotonically decrease or monotonically increase. An active noise control device comprising: a forced setting means. 9. A cycle detecting means for detecting the cycle of the periodic noise, a cycle determining means for determining whether the cycle of the noise detected by the cycle detecting means is a predetermined cycle or more, and the cycle. 7. A first processable state control unit for setting the filter coefficient forced setting unit to a process executable state when the determination unit determines that the noise period is equal to or longer than the predetermined period, and the first processable state control unit is provided. Item 9. The active noise control device according to any one of items 8. 10. An amplitude determining means for determining whether the amplitude of the adaptive digital filter is equal to or more than a predetermined width, and the filter coefficient when the amplitude determining means determines that the amplitude is equal to or more than the predetermined width. The active noise control device according to any one of claims 6 to 9, further comprising a second processable state control unit that sets the compulsory setting unit to a process executable state.