JPH06149266A - Active sound eliminating device - Google Patents

Abstract

PURPOSE:To set the parameters of coefficient update algorithm to values which are proper from the point of view of the whole system and to take an artificial countermeasure against un unforeseen unstable state. CONSTITUTION:This active sound eliminating device is provided with a parameter part 71 which sets the parameters of the coefficient update algorithm according to the state of a reference signal (d) detected by a 1st mechanical-to-electric transducer 21 and a mu adjustment command block 72 which forcibly indicates the adjustments of the parameters to the parameter setting part 71 from outside. Consequently, the parameters which are fixed before are varied according to the state of the reference signal and set to proper values viewed from the whole system, and an adjustment command for the parameters is generated from outside through the mu adjustment command block 72 according to the transition of a sound elimination state, thereby taking an artificial countermeasure against an unexpected unstable state.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an active silencer used for canceling noise generated during the operation of a mechanical device, and more particularly to a variable parameter active silencer.

[0002]

2. Description of the Related Art FIG. 3 shows a conventional configuration. In FIG. 3, a first mechanical → electrical converter 21 is installed near the sound source or the vibration source 1 or in the course of propagation of a wave at a control point. The first mechanical-to-electrical converter 21 detects a signal highly correlated with the sound source or the vibration source 1 and outputs it to the amplifier 2
It is output to the signal processing circuit 6 via 2 as the reference signal d. The signal processing circuit 6 calculates the control signal S ′ for silencing based on the reference signal d. This control signal S ′ is sent to the control electric → mechanical converter 41 via the amplifier 42.
As a result, in order to cancel the sound or vibration (hereinafter, generally referred to as wave) in the control area 5, a control wave having a phase shifted by 180 degrees is generated.

On the other hand, in the control area 5, a second mechanical → electrical converter 31 is installed. The second mechanical-electrical converter 31 detects the residual sound 8 and outputs its error signal ε to the signal processing circuit 6 via the amplifier 32.

Here, the signal processing circuit 6 calculates and outputs the above-mentioned control signal S'using the reference signal d and the error signal ε. This signal processing circuit 6 is usually composed of a microcomputer and performs signal processing based on a known algorithm such as LMS shown in FIG.

In the signal processing circuit 6, 63 is a block for correcting the feedback of the control signal S'to the reference signal d, and 64 is a delay correction block.
Reference numeral 62 denotes a coefficient updating unit that updates the coefficient of the muffling control filter 61, and sets an optimum coefficient that minimizes the error signal ε to realize the muffling control.

[0006]

Here, the LMS algorithm for updating the coefficient A (n) in the silence control filter 61 of FIG. 4 to A (n + 1) is shown in the equation (1). A (n + 1) = A (n) -2 με (n) D ′ (n) (1) where A (n): silence control filter μ: parameter ε (n): error signal D ′ (n): delay Correction value

As can be seen from the above equation (1), the second right side
As the term approaches zero, the optimum coefficient is obtained and it approaches convergence. In this case, the only adjustable is the parameter μ. For this parameter μ, a tentative maximum value can be determined by the following equation (2). However, in reality, it is difficult to obtain the root mean square sum of the delay correction values D ′ (n) in advance, and conventionally, the value of the parameter μ was determined by trial and error based on the actual noise reduction effect. .

[0008]

[Equation 1]

Further, it is empirically known that it is necessary to further reduce the above equation (2) to 1/10 to 1/100 due to a calculation error or the like. It was impossible to raise or lower.

The present invention has been made in view of the above points, and it is possible to set the parameters of the coefficient updating algorithm to appropriate values from the viewpoint of the entire system, and it is possible to manually set the parameters for sudden instability. It is an object of the present invention to provide an active silencer that can cope with the problem.

[0011]

The active silencer of the present invention detects a signal having a high correlation with a sound source or an excitation source as a reference signal and generates a control signal with a phase shift of 180 degrees from a silence control filter. In the active muffling device that mutes, detects residual sound, and updates the coefficient of the muffling control filter based on a predetermined coefficient updating algorithm, sets the parameter of the coefficient updating algorithm according to the state of the reference signal. A parameter setting means and an adjustment command means for forcibly instructing the parameter setting means to adjust the parameters from the outside are provided.

[0012]

According to the above construction, by changing the parameter which is conventionally fixed according to the state of the reference signal,
It can be set to an appropriate value for the entire system. In addition, since a parameter adjustment command can be issued from the outside according to the transition of the muffling situation, it is possible to deal with a sudden unstable situation manually.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An active silencer according to an embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the structure.
In FIG. 1, the same parts as those in FIG. 4 are designated by the same reference numerals, and the description thereof will be omitted here. In this embodiment, the signal processing circuit 70 is provided with a parameter setting section 71 and a μ adjustment command block 72. Parameter setting section 7
1 executes the arithmetic processing as shown in FIG. 2 every time the reference signal d is input, and sets the successive LMS parameter μ to the optimum value. The μ adjustment command block 72 gives a parameter μ adjustment command 73 to the parameter setting unit 71 based on an empirical rule when the parameter μ needs to be corrected. To issue a command from the μ adjustment command block 72, a switch operation is performed from the outside.

FIG. 2 is a diagram showing a calculation flow in the parameter setting section 71. In the parameter setting unit 71, the moving average of the delay correction values D ′ (n) {σ for each input of the reference signal d
(N)} ² Is calculated by the following equation (3) and is substituted into the equation (4) to obtain Sum. Here, I is the silence control filter 61
Sum is equivalent to the sum of average power of tap inputs, which is the denominator on the right side of Expression (2). {Σ (n)} ² = Α {D '(n)} ² + (1-α) {σ (n-1)} ² (3) Sum = I × {σ (n)} ² (4) However, it is stable within the range of 0 <α <1. I: Number of taps of the muffling control filter Next, the maximum value μmax of the parameter μ can be obtained by the equation (5), The range (6) formula is determined. μmax = 1 / Sum = 1 / I × {σ (n)} ² (5) 0 <μ = μ / I × {σ (n)} ² <Μmax (6) where 0 <μ <1

Further, it is known from experience that the value of the parameter μ is set to 1/10 to 1/100 due to a calculation error or the like. Therefore, when there are many residual sounds 8 and effective silencing cannot be obtained. , .Mu. Adjustment command block 72 gives an instruction to set the parameter .mu. Value to 1/10 to 1/100. That is, the value of μ is multiplied by β as in the expression (7) to adjust. μ = μ × β = [μ / I × {σ (n)} ² ] × β (7) However, 0.01 ≦ β ≦ 0.1 If there is no command from the μ adjustment command block 72, it is assumed that no adjustment is necessary, and the calculation of the formula (7) is performed. To continue to the next.

In the muffler thus constructed, a signal having a high correlation with the sound source or the vibration source 1 is detected by the first mechanical → electrical converter 21. This signal is
After being amplified by the amplifier 22, the A / D converter 23
Is converted into a digital signal and is input to the delay correction filter 64 of the signal processing circuit 70 as the reference signal d.

The delay correction filter 64 processes the input reference signal d according to the flow shown in FIG. 2 to obtain a delay correction value D '(n). The parameter setting unit 71 is
From equation (3), the moving average of D '(n) {σ (n)} ²
Is calculated, and the parameter μ of the LMS algorithm is obtained from the equations (4), (5) and (6). The coefficient updating unit 62 uses μ
According to the signal from the adjustment command block 72, the parameter μ
Is multiplied by β (equation (7)). At this time, it is judged from the μ adjustment command block 72 whether or not there is a command, and if there is no command, the parameters in the LMS algorithm remain unchanged from the expression (6).
Used as a data.

On the other hand, the residual sound 8 is detected by the second mechanical → electrical converter 31 installed in the control area 5, and the error signal ε is detected via the amplifier 32 and the A / D converter 33.
Is entered. The coefficient updating unit 62 uses the error signal ε to calculate the coefficient of the silence control filter 61 based on a known algorithm such as LMS. The muffling control filter 61 generates a control signal S ′ in which the phase for canceling the wave in the control region 5 is shifted by 180 degrees based on this coefficient. The control signal S ′ is sent to the control electrical → mechanical converter 41 via the D / A converter 43 and the amplifier 42, and a control wave is generated.

The flow chart of FIG. 2 is entirely calculated by the signal processing software. Further, the μ adjustment command block 72 is realized by hardware, and when the switch is turned on,
It is possible to switch to n steps (n ≧ 2) up to 0.01,
The value for β in equation (7) is specified by operating each switch. Further, when the switch is off, there is no μ adjustment command 73, and the calculation of equation (7) is passed and the process proceeds to the next process.

As described above, by changing the conventionally fixed parameter μ in accordance with the fluctuation of the input level of the reference signal d, it becomes possible to set an appropriate μ in view of the entire system, and it is faster. A stable muffling state can be obtained by setting the convergence state.

Further, depending on the transition of the mute condition, a command for correcting the parameter μ can be given from the outside, so that it is possible to manually cope with a sudden unstable condition. In other words, when there is a large amount of residual sound 8 in the control region 5 although it has converged, by giving a command to reduce the value of the parameter μ, stable convergence can be achieved and the oscillation tendency can be suppressed.

[0023]

As described above, according to the present invention, the conventionally fixed parameter is changed according to the state of the reference signal, so that it can be set to an appropriate value viewed from the whole system. Since the parameter adjustment command is issued from the outside, it is possible to handle the sudden unstable situation manually.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a calculation process of the embodiment.

FIG. 3 is a block diagram showing a conventional configuration.

FIG. 4 is a flowchart showing conventional arithmetic processing.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Sound source or added sound, 21 ... 1st machine-> electric converter, 22 ... Amplifier, 23 ... A / D converter, 31 ... 2nd machine-> Electric converter, 32 ... Amplifier, 33 ... A / D converter, 41 ... Controlling electrical-to-mechanical converter, 42 ... Amplifier, 43 ... D / A converter, 5 ... Control area, 61 ... Silence control filter, 62 ... Coefficient updating unit, 63 ... Feedback correction block , 64 ... Delay correction block, 70 ... Signal processing circuit, 71 ... Parameter setting section, 72 ... μ adjustment command block, 73 ... μ adjustment command, 8 ... Residual sound.

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] February 4, 1993

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 1

[Correction method] Change

[Correction content]

[Figure 1]

Claims (1)

[Claims] 1. A signal having a high correlation with a sound source or an excitation source is detected as a reference signal, a control signal having a phase shifted by 180 degrees is generated by a muffling control filter to muffle the sound, and a residual sound is detected to determine a predetermined value. In an active muffling device that updates the coefficient of the muffling control filter based on a coefficient updating algorithm, a parameter setting unit that sets a parameter of the coefficient updating algorithm according to the state of the reference signal, and an external parameter setting unit On the other hand, an active muffling device comprising an adjustment commanding means for forcibly instructing adjustment of the above parameters.