JP2669291B2 - Method and apparatus for identifying unknown system using adaptive filter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for identifying an unknown system by means of an adaptive filter for identifying a system such as a transmission line or a spatial acoustic coupling path.

[0002]

2. Description of the Related Art An echo canceller, a noise canceller, a howling canceller, an adaptive equalizer, and the like are known as applications to which unknown system identification using an adaptive filter is conventionally applied. In the following, a technique for identifying an unknown system using an adaptive filter will be described for an example of an echo canceller that removes an echo leaking from the transmission side to the reception side on the 4-wire side of the 2-wire / 4-wire conversion circuit.

An echo canceller generates a pseudo echo (echo replica) waveform corresponding to a transmission signal waveform by using an adaptive filter having the number of taps capable of covering a period longer than the impulse response length of an echo path. It operates so as to cancel the echo leaking from the transmission side to the reception side on the 4-wire side of the 2-wire / 4-wire conversion circuit. At this time, the magnitude of each tap coefficient of the adaptive filter is adaptively determined by correlating an error signal obtained by subtracting an echo replica waveform from a signal waveform in which an echo and a received signal coexist with a transmission signal. Will be modified to A typical example of such an adaptive filter coefficient correction algorithm is a “LSM (Least Mean Square) algorithm (LMS Algorithm)” proceeding.
Of Ieeeeeeeeee (Proceedin
gs of IEEE) 63, 12, 1975, 16
The papers (hereinafter abbreviated as “Reference 1”) on pages 92 to 1716 and “Learning Identification Method; L.I.M.
ng IdentificationMethod; L
IM) ", IEE Transactions on Automatic Control" IEEE "
E Transactions on Automat
ic Control Vol. 12, No. 3, 1967, 28
A paper (hereinafter referred to as “Reference 2”) is known on pages 2-287.

[0004] 4 where an echo canceller is actually inserted
If there is a fixed delay between a point on the line line and a point on the 2-wire / 4-wire conversion circuit, the number of taps of the echo canceller is equal to the assumed maximum fixed delay and is substantially equal to the impulse response. It is only necessary to cover both periods of the response waveform portion. Therefore, especially when the fixed delay is large, the number of taps becomes enormous, the hardware scale increases,
The convergence time is increased by the mutual interference of the tap coefficients. In order to solve such difficulties, the position of the response waveform part obtained by removing the fixed delay from the impulse response of the echo path is estimated on the time axis, and the coefficient is set so that the tap coefficients of the adaptive filter are arranged only around the estimated position. A method of adaptively controlling the arrangement, for example, "A Fast Convergence Algorithm for Adaptive FIR Filters with Coarse-Located Taps ( A Fast Convergence)"
Algorithm for Adaptive FI
R Filters with Coarselly L
ocated Taps) "Proceeding of
International Conference on Acoustic, Speech and Signal Processing (Proceedingsof International)
al Conference on Acoustic
s, Speechand Signal Process
ssing) 1991, 1991, 1525-152
It is proposed in a dissertation (hereinafter referred to as “Reference 3”) on page 8. The main point of the method shown in "Reference 3" is that the convergence time is shortened by first estimating the approximate position of the response waveform portion and arranging the tap coefficients only in the vicinity thereof. is there. In this control method, the position estimation of the response waveform portion is performed using the maximum value of the tap coefficient absolute values, and only one range that limits the tap coefficient arrangement is specified. Therefore, when there are a plurality of response waveform parts (in the case of echo, it is called multi-echo), the tap placement limited range must be wide so as to cover all of them, and between the response waveform parts. When there is a large fixed delay, the effect of limiting the tap arrangement position decreases, and an increase in the convergence time cannot be avoided.

A method capable of quickly converging even with respect to such a multi-echo and arranging tap coefficients only in the waveform response section is disclosed in, for example, "A First Algorithm for Adaptive FIR."・ Filters with Coarsely Located Cohesents for Cancellation of Multiple Echoes (A Fast Algorithm f
or Adaptive FIR Filters w
it Coarsly-LocatedCoeff
clients for Cancellation
of Multiple Echoes), 1992
IEICE Autumn Conference, pp. A-93, pp. 1-93.

FIG. 15 is a block diagram showing the configuration of the echo canceller proposed in this "Document 4".
Adaptive filter shown in the figure, to sequentially delaying the transmission signal input from the input terminal 1 (N-1) number of delay elements 20 ₁ has a 20 _N-1, also tap delay is zero An adaptive filter having N taps is included. On the other hand, in order to generate the tap coefficient of this adaptive filter,
L coefficient generating circuits 50 ₁ to 50 _L are provided, and the total number N of taps of the adaptive filter and the coefficient generating circuit 50 _{1 are set.}
There is a relation of N> L with the number L of 50 to 50 _L. That is, the adaptive filter shown in FIG. 15 has a sufficient number of taps to realize a substantial response waveform portion excluding the fixed delay portion, and the tap coefficients are adaptively arranged in the substantial response waveform portion. Generates an echo replica. Therefore, to 50 ₁ each tap outputs and the coefficient generating circuit includes a path switch 7 for switching the connection between the 50 _L, further comprising a tap control circuit 9 for the connection switching control of the path switch 7 ing. The delayed signal output from each of the path switches 7 is supplied to the corresponding coefficient generating circuits 50 ₁ to 50 _L and the multipliers 40 ₁ to 40.
_L and supplied to. The multiplier 40 ₁ to 40 _L is to the coefficient generating circuit 50 ₁ multiplies the delayed signal output by the tap coefficient and the path switch 7 for outputting the 50 _L respectively,
The result of the multiplication is sent to the addition circuit 8 . The adder circuit 8 includes the multiplier 4
The multiplication results of 0 ₁ to 40 _L are added and output to the subtractor 5 as an echo replica. On the other hand, the transmission signal input to the input terminal 1 is transmitted from the output terminal 2 to the two-wire side via the two-wire / four-wire conversion circuit 3, but is transmitted as an echo caused by impedance mismatch or the like. Leak into 4. The echo input from the input terminal 4 is sent to the subtracter 5, where the echo replica output from the adder 8 is subtracted and transmitted to the output terminal 6. The subtraction result is also used as an error signal for updating the coefficient at the same time as the coefficient generating circuit 5
It is supplied from 0 ₁ to 50 _L.

When the LMS algorithm shown in "Document 1" is used as the coefficient correction algorithm, the configuration of the coefficient generation circuit 50 _i (i = 1, 2, ..., L) is as shown in FIG.
It looks like 16 . The delay signal and the error signal supplied to the coefficient generating circuit 50 _i are multiplied by the multiplier 31 and further multiplied by a predetermined constant μ by the multiplier 32 to store the coefficient value stored in the storage circuit 34. Is added by the adder 33, and the addition result is fed back to the storage circuit 34. The value stored in storage circuit 34 represents the tap coefficient value. The storage circuit 34 has a function of forcibly setting the held coefficient value to zero when the coefficient clear signal is input from the coefficient clear circuit 77 (FIG. 15).

As is clear from the above description, the tap coefficient multiplication path of the adaptive filter is connected only to a part of tap outputs selected by the path switch 7. Hereinafter, taps to which the tap coefficient multiplication path is connected are called valid taps, and taps that are not connected are called invalid taps. In the actual adaptive arrangement of tap coefficients, first, as an initial setting,
The tap coefficient multiplication paths are connected to the tap outputs at regular intervals and arranged. For example, delay element 20 _{kN / L} (k satisfies k <L
The output of a positive integer) is output to a coefficient generation circuit 50 _k and a multiplier 40 _k
Can be connected to This kind of connection
Connected to the coefficient generating circuit 50 _k and the multiplier 40 _k.
That the delay elements and a coefficient generation circuit 50 k _{+ 1} to the multiplier 40 k _{+ 1}
(KN / L) -1 between the connected delay elements.
Of delay elements may be present. These become effective taps, and the other taps become invalid taps.

Next, the control operation of the tap control circuit 9 will be described. The control sub-group storage circuit 83 stores sub-group numbers representing a tap control sub-group composed of a plurality of consecutive tap numbers in the order of control. The number of taps belonging to each tap control subgroup is set to be equal. For example, the total number of taps N is 2
Assuming that 0 and the number of tap control subgroups are 5, the number of taps belonging to each tap control subgroup is 4. In addition, the tap control subgroup is set to G (i) (i = 1,
2,. . . . . 5) and tap numbers belonging to G (i) are grouped with {}. G (1) = 1,2,3,4 G (2) = 5,6,7,8 G (3) = 9 , 10, 11, 12 G (4) = 13, 14, 15, 16 G (5) = 17, 18, 19, 20

The control subgroup storage circuit 83 sets the tap control subgroup numbers in ascending order of the group numbers at the time of initial setting. That is, the group number held by the control subgroup storage circuit 83 is Z (n) (n =
1, 2 ,. . . . . Expressed as 5), Z (1) = 1 Z (2) = 2 Z (3) = 3 Z (4) = 4 Z (5) = 5 is initialized. The control subgroup storage circuit 8
The address pointer defining the data read position of No. 3 is set at the head, and outputs the head group number, that is, Z (1) = 1 in the above example.

The memory circuit 72 is a FIFO (first in first out)
This is a system and stores (N−L) invalid tap numbers. However, N shows the total number of taps, and L shows the number of effective taps.
In addition, the tap calculation circuit 76 removes the invalid tap number stored in the storage circuit 72 from all the tap numbers,
That is, the effective tap number is calculated, and the tap number of the calculation result is used as a control signal for switching the taps.
To supply. The path switch 7 operates so as to select the tap output corresponding to the L effective tap numbers received from the tap calculation circuit 76 and transmit it to the coefficient generation circuits 50 ₁ to 50 _L.

The invalid tap number in the initial state, that is, the initial setting value of the storage circuit 72, is selected so that the valid tap numbers are continuously arranged from the smaller of all the tap numbers. For example, the total number of taps N = 20, the number of effective taps L
= 3 and the number of invalid taps NL = 17, all tap numbers are 1, 2, 3 ,. . . . . , 20. At this time, the valid tap numbers are 1, 2 and 3 from the smallest to 3
The tap is selected, and the invalid tap numbers held by the memory circuit 72 are 4, 5 ,. . . . . , 20 is initialized. After the above initial state is set, the coefficient of the tap (= valid tap) selected by the path switch 7 is corrected. Q
Every time the coefficient is modified (Q is a positive integer), the coefficient arrangement, that is, the effective tap position is updated. This effective tap position update is executed in the following procedure.

The minimum coefficient detecting circuit 71 is provided with the effective tap number output from the tap calculating circuit 76 and the coefficient generating circuit 50 _1.
Or 50 _L each receive the output tap coefficient,
The effective tap number corresponding to the coefficient having the smallest absolute value is supplied to the storage circuit 72 and the coefficient clear circuit 77. The coefficient clearing circuit 77 outputs a coefficient clearing signal to the coefficient generating circuit corresponding to the input tap number, thereby setting the coefficient having the minimum absolute value to zero. Since the storage circuit 72 uses the FIFO method, the input tap number is stored at the end of the queue, and the tap number at the head of the queue is transmitted to the determination circuit 73. The control tap range calculation circuit 84 receives the tap control subgroup number Z (n) output from the control subgroup storage circuit 83, and outputs the tap number belonging to the tap control subgroup, that is, G (Z (n)). The minimum tap number K _min and the maximum tap number K _max
Is supplied to the determination circuit 73. For example, in the above example, if Z (n) = 1, G (Z (n))
= G (1) = {1, 2, 3, 4}, and K _max =
4, supply K _min = 1. Input to the judgment circuit 73
K satisfies the determination condition represented by K _min <K <K _max
If not , the input tap number is fed back to the storage circuit 72, and the tap number of the storage circuit 72 is again taken out.
Input to the determination circuit 73. The repeated operation is continued until satisfying the above SL determination condition. When the above determination condition is satisfied, the invalid tap number held in the invalid tap storage circuit 72 is determined, and a new valid tap is determined. The above-described tap position control limited to one tap control subgroup enables concentrated arrangement of tap coefficients.

On the other hand, the change of the tap control subgroup
The following procedure is used. The counter 67 is a counter that counts the number of tap coefficient corrections, and supplies a control signal for changing the control group to the control subgroup storage circuit 83 and the counter 66 each time the number of coefficient corrections reaches a predetermined number. The control subgroup storage circuit 83
When the control signal for changing the control subgroup is received from the counter 67, the address pointer is advanced by one and the tap control subgroup number stored at the next address is output. The counter 66 counts the control signals for changing the control subgroup output from the counter 67, and outputs the control signal for updating the control subgroup order each time the total number of tap control subgroups is reached. The maximum coefficient detection circuit 81 receives the tap coefficient output from each of the coefficient generation circuits 50 ₁ to 50 _L and the effective tap number output from the tap calculation circuit 76, and determines the tap coefficient having the maximum absolute value as each tap control sub-value. The tap control subgroup numbers are detected for each group and the tap control subgroup numbers are output in descending order of the coefficient absolute value. Upon receiving the control signal for updating the control subgroup order from the counter 66, the control subgroup update circuit 82 stores the tap control subgroup numbers output from the maximum coefficient detection circuit 81 in the control subgroup storage circuit 83 in the order supplied. Writing from the head address changes the tap control subgroup order held by the control subgroup storage circuit 83.

As is clear from the above description, since the tap control range moves successively over all taps, it converges at a relatively high speed even when there are a plurality of substantial response waveform parts such as a multi-echo, and the response The tap coefficient multiplication path can be arranged only in the waveform portion.

[0016]

[0017]

SUMMARY OF THE INVENTION An object of the present invention is to provide a method and apparatus for identifying an unknown system by an adaptive filter having a short convergence time even for an unknown system having an impulse response composed of a plurality of response waveform parts. Is.

[0018]

According to a first aspect of the present invention, there is provided an identification method, wherein a part of all taps of an adaptive filter is stored as an effective tap used for a product-sum operation, and the number is stored.
The coefficient multiplication of the product-sum operation is applied only to the valid taps,
Taps that are not used for the product-sum operation are stored as invalid taps in the queue, and the coefficient corresponding to the number of the valid tap is corrected by a predetermined number of times. The tap whose absolute value of the coefficient is the smallest is set as an invalid tap and its number is stored at the end of the queue, and the invalid tap number at the head of the queue is taken out and used as a new valid tap number. Unknown system having an impulse response composed of a plurality of response waveform portions and a fixed delay by using a plurality of adaptive filters for adaptively controlling tap positions by using the adaptive filters and identifying each adaptive filter at least one or more response waveform portions It is characterized by identifying.

The identification device of the second invention uses a plurality of adaptive filters for correcting and updating the coefficient and tap position according to an error signal obtained by subtracting the identification signal output by the adaptive filter from the output signal of the unknown system. In an unknown system identification device using an adaptive filter for identifying an unknown system having an impulse response composed of a response waveform part and a fixed delay, at least one of the response waveforms is supplied with an input signal to the unknown system. A plurality of adaptive filters for identifying the parts, an adder for adding the respective outputs of the adaptive filters to output the identification signal, and a subtraction for subtracting the identification signal from the output signal of the unknown system to obtain the error signal The adaptive filters each having a fixed delay next to the fixed delay identified by itself. In addition to receiving the information about the fixed delay identified by the adaptive filter and the information of the identification completion, identifying the response waveform section having the next largest fixed delay after the fixed delay identified by itself. And supplying information on the identified fixed delay and information on completion of identification to the adaptive filter.

In the identification method of the third invention, a part of all taps of the adaptive filter is stored as an effective tap used for the product-sum operation, and the coefficient multiplication of the product-sum operation is performed only on the effective tap. The number of taps that are not used in the multiply-accumulate operation is stored in the queue as an invalid tap, and each time the coefficient corresponding to the number of the valid tap is modified a predetermined number of times, among the valid taps, The tap with the smallest absolute value of the corresponding coefficient is stored as an invalid tap at the end of the wait queue, and the invalid tap number at the head of the queue is extracted and tapped as a new valid tap number. When an unknown system is identified using one adaptive filter for adaptively controlling the position, the first step support is performed from the start of the adaptive operation to a predetermined number of coefficient corrections. Performs coefficient correction with reference to FIG,
After that, coefficient correction is performed using the second step size, and when the invalid tap number at the head of the queue is near the valid tap number determined to have the maximum coefficient absolute value, a new coefficient is newly added. The valid tap number is stored in the end of the queue without validating it when it is not in the vicinity, and the invalid tap number extracted from the head of the queue becomes valid until the new valid tap number Adaptive control of the tap position is performed by repeating the comparison between the tap number and the effective tap number corresponding to the maximum absolute coefficient value.

In the identification method according to a fourth aspect of the present invention, a part of all taps of the adaptive filter is stored as an effective tap used for a product-sum operation, and the coefficient multiplication of the product-sum operation is performed only on the effective tap. The taps that are not used for the product-sum operation are stored as invalid taps in the queue, and the coefficient corresponding to the valid tap number is corrected a predetermined number of times. The tap having the smallest absolute value of the coefficient to be stored is stored at the end of the queue as an invalid tap, and the tap is operated by taking out the invalid tap number at the head of the queue and setting it as a new valid tap number. When identifying an unknown system using one adaptive filter that adaptively controls the position, the first step from the start of adaptive operation to the predetermined number of coefficient corrections Performs coefficient correction using the size,
Thereafter, a second step size is used, and the queue is further composed of two queues, the effective tap number having the smallest absolute value of the coefficient and the effective tap number having the largest absolute value of the coefficient. If it is near the effective tap number determined to be present, it is stored at the end of the first queue; if it is not near, it is stored at the end of the second queue; The tap position is adaptively controlled by extracting the invalid tap number at the head of the queue and newly setting it as a valid tap number.

According to a fifth aspect of the present invention, an unknown system is identified using an adaptive filter that corrects tap coefficients using an error signal obtained by subtracting an identification signal output from an adaptive filter from an output signal of an unknown system. In the apparatus, a plurality of cascaded delay elements for delaying an input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and the path switch Receiving the output signal, the error signal, and the coefficient clear signal, multiplying a plurality of coefficient generation circuits for generating tap coefficients, and multiplying each of the tap coefficients output by the coefficient generation circuit by the output signal of the path switch. A plurality of multipliers, an adder that adds the outputs of the plurality of multipliers and outputs the identification signal, and subtracts the identification signal from the output signal of the unknown system A subtractor for obtaining the error signal Te,
Receiving a tap coefficient output by the coefficient generating circuit, a tap control circuit for generating a control signal of the path switch and the coefficient clear signal, the coefficient generating circuit,
A first multiplier that multiplies the delay signal and the error signal, a second multiplier that multiplies an output signal of the first multiplier by a predetermined constant, and the second multiplier An adder for adding the output signal of the adder and the feedback signal, and holding the output signal of the adder as the tap coefficient, feeding back the adder as the feedback signal, and resetting to zero by the coefficient clear signal. The tap control circuit receives a counter that outputs a control signal after counting a predetermined number of times from the start of the adaptive operation, a storage circuit that stores an invalid tap number, and a storage circuit that receives the output of the storage circuit. And a first switch that receives the output of the determination circuit and selectively returns to the storage circuit in response to the control signal of the counter. And a tap number calculating circuit for calculating the remaining tap numbers excluding the tap numbers held by the storage circuit group from all the tap numbers of the adaptive filter and outputting the calculated tap numbers to the path switch, and an output of the tap number calculating circuit. A minimum coefficient detection circuit that receives a tap number to be output and a coefficient value output from the plurality of coefficient generation circuits, obtains a tap number corresponding to the minimum absolute value coefficient value, and supplies the tap number to the storage circuit; A coefficient clearing circuit that supplies the coefficient clearing signal to a coefficient generating circuit corresponding to a tap number output by a tap number output circuit and a tap number output by the plurality of coefficient generating circuits. A maximum coefficient detection circuit that outputs an effective tap number having a maximum absolute coefficient value, an effective tap number received from the maximum coefficient detection circuit, and a predetermined constant. And a control tap range calculation circuit for calculating a control tap range as the determination condition using the first and second step sizes. And a second switch for switching to the step size of and outputting.

The identifying apparatus of the sixth invention is an apparatus for identifying an unknown system by using an adaptive filter which corrects a coefficient by using an error signal obtained by subtracting the identification signal output by the adaptive filter from the output signal of the unknown system. In, a plurality of cascaded delay elements for delaying the input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and an output of the path switch A plurality of coefficient generation circuits that receive a signal, the error signal, and the coefficient clear signal, and generate tap coefficients, and a plurality of coefficient multipliers that multiply each tap coefficient output by the coefficient generation circuit and the output signal of the path switch. A multiplier, an adder that adds the outputs of the multipliers and outputs the identification signal, and a signal that subtracts the identification signal from the output signal of the unknown system to obtain the error signal. And a tap control circuit that receives a tap coefficient output from the coefficient generation circuit and generates a control signal of the path switch and the coefficient clear signal, wherein the coefficient generation circuit includes the delay circuit. A first multiplier that multiplies a signal by the error signal, a second multiplier that multiplies the output of the first multiplier by a predetermined constant, and an output signal of the second multiplier And a feedback signal, and a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. The tap control circuit comprises a counter that outputs a control signal after counting a predetermined number of times from the start of the adaptive operation, and a tap control circuit in the vicinity of an effective tap number that is determined to have a maximum coefficient absolute value. A first storage circuit that stores an invalid tap number, a second storage circuit that stores an invalid tap number that is not in the vicinity, and a determination condition given by receiving the outputs of the first and second storage circuits A determination circuit that outputs one and discards the others, and calculates the remaining tap numbers from all tap numbers of the adaptive filter, excluding the tap numbers held by the first and second storage circuits, and by using the route switch. A tap number calculation circuit for outputting, a tap number output by the tap number calculation circuit, and a tap coefficient output by the plurality of coefficient generation circuits, and a minimum coefficient detection for obtaining a tap number corresponding to a coefficient value having a minimum absolute value. A circuit, a synthesizing circuit for synthesizing the output signal of the judging circuit and the output signal of the minimum coefficient detecting circuit, and An evaluation circuit that evaluates whether or not the tap number supplied from the synthesis circuit belongs to the neighborhood, and a tap number supplied from the synthesis circuit according to the evaluation result of the evaluation circuit. A distribution circuit for distributing to the circuit, a coefficient clear circuit for supplying the coefficient clear signal to a coefficient generation circuit corresponding to the tap number output by the minimum coefficient detection circuit, a tap number output by the tap number calculation circuit, and the coefficient A maximum coefficient detection circuit that receives the tap coefficient output by the generation circuit and outputs an effective tap number having a maximum absolute coefficient value, an effective tap number received from the maximum coefficient detection circuit, and a predetermined constant. A control tap range calculation circuit for specifically calculating the range in the vicinity thereof, and a control signal from the counter upon receiving first and second step sizes Characterized in that it is composed of a first switch from the step size is switched to the second step size is output from.

The identifying apparatus of the seventh invention is an apparatus for identifying an unknown system by using an adaptive filter whose coefficient is corrected by using an error signal obtained by subtracting the identification signal output by the adaptive filter from the output signal of the unknown system. , A plurality of cascaded delay elements for delaying an input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and an output of the path switch. A plurality of coefficient generation circuits for receiving the signal, the error signal, and the coefficient clear signal and generating tap coefficients; and A multiplier, an adder for adding the output of the multiplier to output the identification signal, and subtracting the identification signal from an output signal of the unknown system to generate the error signal. And a tap control circuit that receives a tap coefficient output by the coefficient generation circuit and generates a control signal of the path switch and the coefficient clear signal, and the coefficient generation circuit includes a delay signal And a first multiplier for multiplying the error signal, a second multiplier for multiplying an output signal of the first multiplier by a predetermined constant, and an output signal of the second multiplier And a feedback signal, and a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. The tap control circuit stores a control subgroup number corresponding to a tap control subgroup composed of a plurality of consecutive tap numbers on a one-to-one basis in a selection order. And a storage circuit group including a plurality of storage circuits, and taps for calculating the remaining tap numbers excluding the tap numbers held by the storage circuit group from all the tap numbers of the adaptive filter and outputting the tap numbers to the path switch. A number calculation circuit, a minimum coefficient detection circuit that receives a tap number output by the tap number calculation circuit and a tap coefficient output by the coefficient generation circuit, and obtains a tap number corresponding to a coefficient value having a minimum absolute value; An evaluation circuit for obtaining a tap control subgroup corresponding to the tap number supplied from the minimum coefficient detection circuit; and the minimum value in one of the storage circuits corresponding to the tap control subgroup supplied from the evaluation circuit. The distribution circuit for transmitting the tap number supplied from the detection circuit, and the coefficient generation circuit corresponding to the tap number output by the minimum coefficient detection circuit A coefficient clearing circuit for supplying a coefficient clearing signal; a processing circuit for taking out and discarding a tap number from one of the storage circuit groups corresponding to the control subgroup number output by the first storage circuit; In response to the tap numbers output by the number calculation circuit and the coefficient values output by the plurality of coefficient generation circuits, control subgroup numbers arranged in the order of the magnitude of the maximum absolute coefficient value for each control subgroup are output. A maximum coefficient detection circuit and a first coefficient circuit for supplying an instruction signal for changing the control subgroup number to the first storage circuit each time the number of times of coefficient correction reaches a predetermined number
And a second counter that outputs a control subgroup order change signal each time the instruction signal of the first counter is output a predetermined number of times, and a change signal output by the second counter. And a control sub-group updating circuit for writing the control sub-group number output from the maximum coefficient detecting circuit into the first storage circuit.

In the identification method of the eighth invention, a part of all the taps of the adaptive filter is stored as an effective tap used for the product-sum operation, and the coefficient multiplication of the product-sum operation is stored as the effective tap. The number of taps not used for the product-sum operation is stored in a queue as an invalid tap number, and a coefficient corresponding to the number of the valid tap is corrected by a predetermined number of times. , The tap with the smallest absolute value of the corresponding coefficient is stored as an invalid tap at the end of the queue, the invalid tap number at the head of the queue is extracted, and the invalid tap number has the largest absolute coefficient value. If it is in the vicinity of the effective tap number determined to be a valid tap number, it is newly set as an effective tap number. The tap position is adapted by repeating the operation of repeatedly comparing the tap number at the head of the queue with the valid tap number corresponding to the maximum absolute coefficient value until the invalid tap number extracted from the head of the queue becomes new valid. When performing identification of an unknown system using an adaptive filter to be controlled, all taps of the adaptive filter are divided into a plurality of tap control subgroups each of which has an equal number of consecutive taps. The tap number belonging to the selected one tap control subgroup is defined as the neighborhood, and the selected tap control subgroup is changed according to the selection order for each predetermined coefficient correction number, and the selection order and the coefficient correction number Adaptively controls the tap position by determining it using the information of the effective tap coefficient in each tap control subgroup. Characterized in that it.

The identification device of the ninth invention is a device for identifying an unknown system by using an adaptive filter which modifies coefficients using an error signal obtained by subtracting the identification signal output by the adaptive filter from the output signal of the unknown system. In, a plurality of cascaded delay elements for delaying the input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and an output of the path switch A plurality of coefficient generation circuits that generate a tap coefficient by receiving a signal, the error signal, and the coefficient clear signal;
A plurality of multipliers each multiplying each tap coefficient output by the coefficient generation circuit and an output signal of the path switch,
An adder for adding the output of the multiplier and outputting the identification signal; a subtracter for subtracting the identification signal from the output signal of the unknown system to obtain the error signal; and a tap coefficient output by the coefficient generating circuit. And a tap control circuit for generating a control signal for the path switch and the coefficient clear signal, wherein the coefficient generation circuit includes a first multiplier for multiplying the delay signal and the error signal, A second multiplier that multiplies the output signal of the first multiplier by a predetermined constant, an adder that adds the output signal of the second multiplier and a feedback signal, and an adder of the adder. The tap control circuit has a first-in first-out structure, which holds an output signal as the tap coefficient, is fed back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. A first memory circuit, and a tap number calculation circuit that calculates the remaining tap numbers excluding the tap numbers held by the first memory circuit from all tap numbers of the adaptive filter and outputs the tap numbers to the path switch; A minimum coefficient detection circuit that receives the tap number output from the tap number calculation circuit and the tap coefficient output from the coefficient generation circuit and transmits the tap number corresponding to the coefficient value having the smallest absolute value to the first storage circuit. , A one-to-one correspondence with a coefficient clear circuit that supplies the coefficient clear signal to the coefficient generation circuit corresponding to the tap number output by the minimum coefficient detection circuit, and a tap control subgroup composed of a plurality of consecutive tap numbers. A second storage circuit for storing a control subgroup number to be controlled in accordance with the selection order, and a control subgroup number output from the second storage circuit, A control tap range calculation circuit that outputs the upper limit and the lower limit of the tap numbers belonging to the group number, and whether the tap numbers extracted from the first storage circuit are within the upper limit and the lower limit, If the tap number is not within the range, the determination circuit for returning the tap number to the first storage circuit, the tap number output by the tap number calculation circuit and the tap coefficient output by the coefficient generation circuit are received, A maximum coefficient detection circuit that detects a coefficient value having the maximum absolute value for each control subgroup and outputs the control subgroup number and the maximum absolute coefficient value arranged in the order of the magnitude of the maximum absolute coefficient value, and the maximum coefficient detection circuit A first supply signal for changing the control subgroup number to the second memory circuit each time the maximum number of coefficient corrections is received from the circuit and the number of coefficient corrections reaches a predetermined number. Counter, a second counter that outputs a control subgroup order change signal every time the instruction signal from the first counter is output a predetermined number of times, and a change signal output by the second counter. And a control subgroup updating circuit for writing the control subgroup number output by the maximum coefficient detecting circuit to the second storage circuit, and the first counter controls each control received from the maximum coefficient detecting circuit. The number of counts for supplying the instruction signal is determined using the maximum absolute coefficient value corresponding to the subgroup number.

An identification apparatus according to a tenth aspect of the present invention is an apparatus for identifying an unknown system using an adaptive filter that corrects coefficients using an error signal obtained by subtracting an identification signal output from an adaptive filter from an output signal of an unknown system. At
A plurality of cascaded delay elements for delaying an input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and an output signal of the path switch A plurality of coefficient generating circuits that receive the error signal and the coefficient clear signal and generate tap coefficients, and a plurality of multipliers that multiply the tap coefficients output by the coefficient generating circuits and the output signal of the path switch, respectively. An adder for adding the output of the multiplier to output the identification signal, a subtractor for subtracting the identification signal from the output of the unknown system to obtain the error signal, and a tap coefficient output by the coefficient generating circuit And a tap control circuit for generating the control signal of the path switch and the coefficient clear signal. The coefficient generation circuit multiplies the delay signal by the error signal. A first multiplier, a second multiplier that multiplies the output signal of the first multiplier by a predetermined constant, and an addition that adds the output signal of the second multiplier and the feedback signal And a storage circuit which holds the output signal of the adder as the tap coefficient, and which is fed back to the adder as the feedback signal and reset to zero by the coefficient clear signal. The circuit calculates a first memory circuit having a first-in first-out structure and all tap numbers of the adaptive filter excluding the tap numbers held by the first memory circuit, and outputs the tap numbers to the path switch. Receiving the tap number calculation circuit, the tap number output by the tap number calculation circuit, and the tap coefficient value output by the coefficient generation circuit, and the tap number corresponding to the coefficient value having the smallest absolute value is stored in the first memory. A minimum coefficient detecting circuit for transmitting to the channel, a coefficient clear circuit for supplying the coefficient clear signal to a coefficient generating circuit corresponding to the tap number output by the minimum coefficient detecting circuit, and a tap composed of a plurality of consecutive tap numbers A second storage circuit that stores a control subgroup number corresponding to the control subgroup in a one-to-one correspondence in accordance with a selection order, and a control subgroup number output from the second storage circuit are received and belong to the control subgroup number. The control tap range calculation circuit for outputting the upper limit and the lower limit of the tap number and the tap number fetched from the first storage circuit are judged to be within the range of the upper limit and the lower limit, and if they are not within the range. Is a determination circuit for returning the tap number thus taken out to the first storage circuit, the tap number output by the tap number calculation circuit, and the coefficient generation time. The maximum coefficient detection circuit that receives the tap coefficient output from the path, detects the maximum absolute coefficient value for each control subgroup, and outputs the control subgroup number arranged in the order of the magnitude of the maximum absolute coefficient value. And a control subgroup coefficient information extraction circuit that outputs coefficient information for each control subgroup in response to a tap number output by the tap number calculation circuit and a tap coefficient output by the coefficient generation circuit, and the control subgroup A first counter that receives the coefficient information from the coefficient information extraction circuit and supplies an instruction signal for changing the control subgroup number to the second storage circuit each time the number of times the coefficient is modified reaches a predetermined number; A second counter that outputs a control subgroup order change signal each time the instruction signal from the first counter is output a predetermined number of times, and the second counter And a control subgroup updating circuit for writing the control subgroup number output from the maximum coefficient detecting circuit to the second storage circuit in response to a change signal applied to the second storage circuit. The number of counts for supplying the instruction signal is determined by using coefficient information corresponding to each control subgroup number received from the extraction circuit.

The identifying apparatus of the eleventh invention is an apparatus for identifying an unknown system by using an adaptive filter which modifies coefficients by using an error signal obtained by subtracting the identification signal output by the adaptive filter from the output signal of the unknown system. At
A plurality of cascaded delay elements for delaying an input signal of the unknown system, a path switch for selecting and outputting a part of the delay signal output from the delay element, and an output signal of the path switch A plurality of coefficient generating circuits that receive the error signal and the coefficient clear signal and generate tap coefficients, and a plurality of multipliers that multiply the tap coefficients output by the coefficient generating circuits and the output signal of the path switch, respectively. An adder for adding the output of the multiplier and outputting the identification signal; a subtractor for subtracting the identification signal from the output signal of the unknown system to obtain the error signal; and a tap output by the coefficient generation circuit A tap control circuit that receives a coefficient value and generates a control signal for the path switch and the coefficient clear signal is provided, and the coefficient generation circuit includes the delay signal and the error signal. A first multiplier for multiplication, a second multiplier for multiplying an output signal of the first multiplier by a predetermined constant, and an output signal of the second multiplier and a feedback signal And a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. The tap control circuit calculates a first memory circuit having a first-in first-out structure, and a remaining tap number from all tap numbers of the adaptive filter, excluding tap numbers held by the first memory circuit,
The tap number calculation circuit for outputting to the path switch, the tap number output by the tap number calculation circuit, and the tap coefficient output by the coefficient generation circuit are received, and the tap number corresponding to the coefficient value having the smallest absolute value is received. No. 1 storage circuit, a coefficient clear circuit that supplies the coefficient clear signal to the coefficient generation circuit corresponding to the tap number output by the minimum coefficient detection circuit, and a plurality of consecutive tap numbers. A second storage circuit that stores a control subgroup number corresponding to the tap control subgroup in a one-to-one correspondence with the tap control subgroup, and a control subgroup number output from the second storage circuit, and receive the control subgroup number. Control tap range calculation circuit for outputting upper and lower limits of tap numbers belonging to the number, and tap numbers extracted from the first storage circuit A judgment circuit for judging whether the tap number is within the range of the upper limit and the lower limit, and returning the tap number thus taken out to the first memory circuit when not within the range, and a tap output by the tap number calculation circuit A control subgroup coefficient information extraction circuit that outputs a control subgroup number in the order determined by the coefficient information for each control subgroup in response to the number and the tap coefficient output by the plurality of coefficient generation circuits, and the tap number. Receiving the tap number output from the calculation circuit and the tap coefficient output from the coefficient generation circuit, the coefficient value with the maximum absolute value is detected for each control subgroup, and the maximum absolute coefficient value and control subgroup number are output. The maximum coefficient detection circuit that receives the maximum coefficient absolute value and the control subgroup number from the maximum coefficient detection circuit Said every that the second
A first counter that supplies an instruction signal for changing the control subgroup number to the storage circuit of the second storage circuit, and the control subgroup order is changed every time the instruction signal of the first counter is output a predetermined number of times. A second counter that outputs a signal, and a control subgroup update that writes the control subgroup number output by the control subgroup coefficient information extraction circuit to the second storage circuit by the change signal output by the second counter Wherein the first counter comprises:
It is characterized in that the maximum coefficient absolute value corresponding to each control subgroup number received from the maximum coefficient detection circuit is used to determine the count number for supplying the instruction signal.

[0029]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the first and second inventions of the present invention. In the figure, K adaptive filters (ADF ₁ , ADF ₂ ...
_{K) 900 1, 900 2,} . . . . . , 900 _K are all connected in parallel, and each output signal is supplied to the adder 400. The adder 400 includes adaptive filters 900 ₁ , 90
0 ₂ ,. . . . . , 900 _K outputs are all added and supplied to the subtractor 5 as an echo replica. The echo supplied to the subtractor 5 via the input terminal 4 is subjected to subtraction of an echo replica by the subtractor 5. The error signal obtained as the result of the subtraction is the adaptive filters 900 ₁ , 900
₂ ,. . . . . , Returned to 900 _K. The adaptive filters 900 ₁ , 900 ₂ ,. . . . . , 900 _K input signal is provided via input terminal 1.

FIG. 2 is a signal waveform of an impulse response exemplifying a multi-echo having two response waveform parts. This A, B, 2 two shows the case where a response waveform portion, these two adaptive as K = 2 in FIG. 1 filter 900 ₁
An example of the case of identification using 900 and 900 ₂ will be described.
That is, the adaptive filters 900 ₁ and 900 ₂ are controlled to generate echo replicas corresponding to the response waveform portions A and B, respectively. First, the adaptive filter 900 ₁
Is a fixed delay T ₁ corresponding to the response waveform portion A, and then the adaptive filter 900 ₂ is a fixed delay T ₁ corresponding to the response waveform portion B.
_It is used to estimate ₂ respectively, but at this time,
The information for estimating the fixed delays T ₁ and T ₂ is actually represented by I _max1 and I _max2 which are the tap numbers of the maximum coefficient absolute values within the respective response waveform portions A and B. ₁ to adaptive filter 900 ₂
The tap number I _max1 is transmitted to. Adaptive filter 900
₂ sets I _max1 supplied from the adaptive filter 900 ₁ to I
_In comparison with _max2 , _Imax2 is determined by a method described later so that _Imax1 < _Imax2 is always satisfied. The adaptive filters are 900 ₁ , 900 ₂ ,. . . . . , 900 _K , the same method is used, but between I _{max (j )} and I _{max (j + 1)} corresponding to the j-th adaptive filter 900 _j ,
Imax _(j) <Imax _{(j + 1)}
Control _{max (j + 1)} . In addition, the adaptive filter 900 ₁
upon detecting a _max1, the control signal is transmitted from the adaptive filter 900 ₁ to the adaptive filter 900 _2. FIG. 3 shows a configuration example of the adaptive filters 900 ₁ and 900 ₂ .

The adaptive filter of FIG. 3, to sequentially delaying the transmission signal input from the input terminal 100 (N-1) number of delay elements 20 ₁ has a 20 _N-1, also tap delay is zero An adaptive filter having N taps is configured. Meanwhile, in order to generate the tap coefficients of the adaptive filter, to the L coefficient generator circuits 30 ₁ no 30 _L is provided. The connection between each tap output and the coefficient generating circuits 30 ₁ to 30 _L can be switched by the path switch 7, and a tap control circuit 9 for controlling connection switching of the path switch 7 is provided. Delayed signals which are the outputs of path switch 7, the corresponding coefficient generator circuits 30 ₁ to 3
0 _L and the multipliers 40 ₁ to 40 _L. The multipliers 40 ₁ to 40 _L include coefficient generation circuits 30 ₁ to 3
The tap coefficient output by O _L is multiplied by the delay signal output by the path switch 7, and the result is supplied to the addition circuit 8. The addition circuit 8 adds all the multiplication results of the multipliers 40 ₁ to 40 _L and outputs the result as an echo replica at the output terminal 2.
Output to 00. The error signal supplied from the input terminal 600, to 40 ₁ coefficient generation circuit for coefficient correction is supplied to the 40 _L.

FIG. 4 shows a coefficient generating circuit 30 _{i according to} this embodiment.
FIG. 4 is a block diagram illustrating a configuration of (i = 1, 2,..., L). The difference between this coefficient generation circuit and the conventional coefficient generation circuit shown in FIG. 16 is the signal representing the step size supplied from the switch 78 to the multiplier 32. That is, in the circuit of FIG. 4, the step size can be changed from outside.

In the adaptive filter of FIG. 3, the "first state" is set.
There are two states, and “second state”. The counter 86 determines the operation state based on the number of coefficient corrections since the adaptive filter started the adaptive operation, and supplies the result to the switch 78 and the evaluation circuit 80. Counter 8
The determination result of No. 6 is also transmitted to the adjacent adaptive filter via the output terminal 710. The start of the adaptive operation is recognized by the counter control signal supplied via the input terminal 700. Now, it is assumed that the “first state” is an initial state, and the output of the counter 86 at that time is “0”. This
At this time, the evaluation circuit 80 unconditionally stores the distribution circuit 79.
Generating a control signal for selecting the path 70 _i ,
Transfer to 79. The counter 80 is the total tattoo of the adaptive filter.
Number N, number of effective taps L, coefficient value required to update coefficient position
The number of times the coefficient is modified to a predetermined number using the number of updates Q
When reaches, change the output from “0” to “1” and adapt
Set the filter to "second state". Evaluation times at this time
The path 80 connects the minimum coefficient detection circuit 71 to the synthesis circuit 187.
Corresponding tap number of the smallest absolute coefficient value supplied via
Is a control tap supplied from the control tap range calculation circuit 84.
Evaluation range, that is, whether it is included in I _max ± L
Then, a control signal for the distribution circuit 79 is generated. The switch 78 is configured to select the step size μ ₁ in the “first state” and select the step size μ ₂ in the “second state”. μ ₁ and μ ₂ are positive constants that satisfy μ ₁ <μ ₂ . Therefore, in the initial state, the coefficient generation circuit 30 _i
The step size supplied to will be μ ₁ . The reason for providing the "first state" is that in the initial growth stage of effective taps, coefficient correction using a small step size is continuously performed, and the tap position corresponding to the coefficient having the maximum absolute value is accurately detected. Then, the position of the substantial response waveform portion is estimated.

The updating of the effective tap position is executed in the following procedure every time the coefficient is corrected Q times (Q is a positive constant). The minimum coefficient detection circuit 71 receives the effective tap number output from the tap calculation circuit 76 and the tap coefficient output from each coefficient generation circuit, and passes the effective tap number corresponding to the coefficient having the smallest absolute value via the synthesis circuit 187. Distribution circuit 79, evaluation circuit 8
0 to the coefficient clear circuit 77. The coefficient clearing circuit 77 outputs a coefficient clearing control signal to the coefficient generating circuit corresponding to the input tap number, thereby setting the tap coefficient having the smallest absolute value to zero. Distribution circuit 79, in response to a control signal from the evaluation circuit 80 selects one of the storage circuits 70 ₁ or 70 _2. When the tap number supplied from the synthesis circuit 187 is included in the control tap range selects the memory circuit 70 _1, when not included select 70 _2, transmits the tap number to each selected memory circuit. Memory circuit 70 ₁
Both and 70 ₂ are FIFO (First-In-
First-Out method, and the one selected by the distribution circuit 79 stores the invalid tap number supplied from the distribution circuit 79. Also, the tap calculation circuit 76
The total tap number from the memory circuit 70 ₁ and 70 ₂ of the holding tap number, excluding the invalid tap number to, i.e. to calculate the effective tap number, the path switch 7 the calculated tap number as a control signal for the tap changer Supply. The path switch 7 selects the tap output corresponding to the L effective tap numbers in response to this, and performs a connection switching operation so as to transmit it to the coefficient generation circuits 30 ₁ to 30 _L. In the initial state, that is, the “first state”, the memory circuit 7
Since 0 ₁ is used, the invalid tap number, that is, the initial setting value of the storage circuit 70 ₁ is selected so that the valid tap numbers are evenly spaced. After setting the initial state, the coefficient of the tap (= effective tap) selected by the path switch 7 is updated. Judging circuit 182 controls the tap range in response to a control signal supplied from the calculation circuit 84, tap number or discarding retrieved by selecting the memory circuit 70 _1, or 70 ₁ and 70 ₂ of Izure or select one The two types of states, that is, whether to supply the extracted tap number to the combining circuit 187, are switched. Maximum coefficient detection circuit 81
Is to no coefficient generating circuit 30 ₁ for each coefficient correction 30 _L is subjected to the effective tap number to the tap coefficient and the tap calculation circuit 76 for each output and outputs the effective tap number I that the absolute value corresponding to the coefficient which is the maximum _max is supplied to the control tap range calculation circuit 84. The maximum coefficient detection circuit 81 also has an input I810
_max ± L also supplied, in calculating the supply I _max ± L to control the tap range calculation circuit 84 is controlled so as not to exceed the I _max ± L supplied through the input terminal 810. Control tap range calculation circuit 8
4 calculates I _max ± L using I _max supplied from the maximum coefficient detection circuit 81 and supplies it to the evaluation circuit 80. The calculated I _max ± L is also transmitted via output terminal 800 to the adjacent adaptive filter. The control tap range calculation circuit 84 also generates a control signal for the determination circuit 182. When the I _max varies beyond a predetermined threshold value, the control signal for feeding back the tap number taken out determination circuit 182 selects the memory circuit 70 ₁ and 70 ₂ alternately to the synthesis circuit 87 Occurs over a period of time. When no change in I _max is detected,
Judging circuit 182 generates a control signal for discarding the tap numbers taken out by selecting the memory circuit 70 _1. I
Fluctuations in _max can be evaluated, for example, by detecting that I _max is continuously greater than a predetermined number of times and has a value different from the previous I _max . The tap number supplied from the determination circuit 182 to the synthesizing circuit 187 is distributed by the distribution circuit 79 to the control tap range calculation circuit 8.
4 depending on the stipulated _{I max} storage circuit 70 ₁ or 7
0 ₂ is fed back to the. By the above procedure, when the I _max varies, an invalid tap number stored in one of the memory circuits 70 ₁ and 70 _2, can be moved mutually to the other memory circuit.

With the above operation, the adaptive filter 9 is first
00 ₁ can estimate the fixed delay T ₁ corresponding to the response waveform portion A, and then the adaptive filter 900 ₂ can estimate the fixed delay T ₂ corresponding to the response waveform portion B. At this time, the information about the fixed delay T ₁ is expressed as I _max1 ± L by the adaptive filter 9.
Output from 00 ₁ of the output terminal 800, the filter 900
It is transmitted to the _second input terminal 810. Also, the fixed delay T
₁ , that is, when the identification of I _max1 is completed, the output signal of the counter 86 of the adaptive filter 900 ₁ is output to the output terminal 71.
Output through 0 and input terminal 7 of the adaptive filter 900 _2.
It is transmitted via 00. Adaptive filters are 900 ₁ , 90
0 ₂ ,. . . . . , 900 _K and 3 or more,
Similar transfer is performed between the j-th adaptive filter 900 _j and the j + 1-th adaptive filter 900 _{j + 1} .

In the present embodiment, the plurality of response waveform portions of the multi-echo are distributed to the plurality of adaptive filters for identification, whereby the convergence time required for the identification can be considerably shortened as compared with the prior art.

FIG. 5 is a block diagram showing another second configuration example of the adaptive filter in this embodiment. The difference from the configuration example of FIG. 3 is that an ERLE evaluation circuit 188 is used instead of the counter 86 of FIG. The ERLE evaluation circuit 188 uses the echo supplied from the input terminal 300 and the error signal obtained via the input terminal 600 to calculate an echo suppression amount (ERLE) obtained by dividing the former by the latter. By comparing the resultant ERLE with a predetermined threshold value, a control signal similar to that of the counter 86 is generated. However, since the ERLE increases, instead of indicating that the control signal supplied from the counter 86 has not completed the predetermined number of coefficient corrections since the start of the coefficient correction, ERLE is used.
Control signal is generated corresponding to the case where is smaller than a predetermined value.

The case in which there are a plurality of adaptive filters (for example, two) as shown in FIG. 1 has been described above with reference to FIG. 3 (or FIG. 5).
(Or FIG. 5) input terminals 700 and 81
0 and the output terminals 710 and 800 are unnecessary, and the configuration may be made as in the embodiments of the fourth and sixth inventions shown in FIG. 6 (or FIG. 7).

FIG. 8 is a block diagram of one embodiment of the third and fifth aspects of the present invention. This embodiment has a problem with the adaptive filter unit of the echo canceller proposed in "Document 3", that is, the range of effective taps (I _max ±
In calculation of L), an excessive computing capacity is often required, and there is a high possibility that an instruction cycle in which tap replacement cannot be executed occurs due to restrictions on the available computing capacity, and the convergence time increases. An example of a configuration for solving the drawback will be described. In FIG. 8, when the coefficient correction is completed a predetermined number of times after the coefficient correction is started, the counter 86 supplies a control signal to the switches 78 and 190. The switch 78 selects μ ₁ as the step size and selects the coefficient generation circuits 30 ₁ , 30 ₂ ,. . . . . 30 _L
Although supplied to, after receiving the control signal from the counter 86 selects the step size mu _2. Where μ
Set ₁ <μ ₂ . Similarly, the switch 190 is used for the counter 8
The circuit is changed from the open state to the closed state by the control signal from 6, and the signal from the determination circuit 73 is fed back to the memory circuit 72. From the memory circuit 72 to the determination circuit 73 and the switch 1
The tap number returned to the storage circuit 72 via 90 is stored again as an invalid tap in the storage circuit 72. In the determination circuit 73, since invalid taps that do not satisfy the determination condition of I _max ± L are supplied to the switch 190, only tap numbers that satisfy I _max ± L are discarded from the storage circuit 72, and thus become valid taps. means. In other words, from the start of coefficient correction until the completion of a predetermined number of times coefficient correction, all invalid taps are sequentially activated,
After completion, only tap numbers that satisfy I _max ± L are selectively used as valid taps. In addition, from the start of coefficient correction to the completion of coefficient correction a predetermined number of times, stable growth of coefficients is realized with a small step size μ ₁ , and the accurate growth obtained at the time of completion is achieved. The maximum coefficient detection circuit 81 determines I _{max according} to the magnitude relationship of the coefficient values. After that, the normal step size is used to achieve fast convergence.

FIG. 9 is a block diagram showing a first embodiment of the seventh invention of the present invention. The configuration of the prior art shown in FIG. 15 includes an evaluation circuit 80, a distribution circuit 74, and storage circuits 70 ₁ and 7 7.
0 ₂ ,. . . . . The operation is the same except that the _70M and the processing circuit 75 are provided. The evaluation circuit 80 calculates the control subgroup number to which the tap number of the minimum absolute coefficient value supplied from the minimum coefficient detection circuit 71 belongs, and supplies it to the distribution circuit 74. Distribution circuit 74, of the to storage circuit 70 ₁ 70 _M, transmits the tap number supplied from the minimum coefficient detector circuit 71 selects one that corresponds to the control sub-group number supplied from the evaluation circuit 78. Storage circuit 70 ₁ through 70 _M has become a FIFO method, and stores the tap number supplied from the distribution circuit 74. The processing circuit 75 stores the storage circuit 70 according to the currently selected control subgroup number supplied from the control subgroup storage circuit 83.
_{One of 1} to 70 _M is selected, and the tap number stored at the head thereof is taken out and discarded. Moreover, the tap number calculating circuit 76, 70 ₁ memory circuits for storing an invalid tap from all tap number to the tap number without any holding invalid tap number to the 70 _M, i.e. to calculate the effective tap number, The calculated tap number is supplied to the path switch 7 as a control signal for tap switching.

In the conventional configuration shown in FIG. 15, the tap number extracted from the invalid tap storage circuit 72 is the determination circuit 7.
3, the invalid tap storage circuit 7 is evaluated with a certain probability.
However, this does not mean that all the tap numbers taken out change from invalid taps to valid taps. On the other hand, in the present embodiment, the tap number extracted from one of the memory circuit groups 70 ₁ to 70 _M becomes 100% effective taps, and there is no waste as in the conventional case.

FIG. 10 is a block diagram showing a second embodiment of the seventh invention of the present invention. The difference from the embodiment of FIG. 9 is that a coefficient absolute value calculation circuit 184 is used instead of the maximum coefficient detection circuit 81 in FIG. The maximum coefficient detection circuit 81 determines the order of the tap control subgroup numbers to be output according to the maximum absolute coefficient value in each tap control subgroup. The order of tap control subgroup numbers to be output is determined using the sum of absolute coefficient values.
The coefficient absolute value calculation circuit 184 receives each tap coefficient value and the effective tap number output from the tap number calculation circuit 76,
The total sum of absolute coefficient values in each tap control subgroup is calculated, and the tap control subgroup numbers are output in the descending order of the total sum.

FIG. 11 is a block diagram showing a third embodiment of the seventh invention of the present invention. The difference from the embodiment of FIG. 9 is that a coefficient square value calculation circuit 85 is used instead of the maximum coefficient detection circuit 81. Coefficient square value calculation circuit 8
5 determines the order of the tap control subgroup numbers to be output using the sum of squared coefficient values in each tap control subgroup. The coefficient square value calculation circuit 85 receives each tap coefficient value and the valid tap number output from the tap number calculation circuit 76, calculates the absolute sum of absolute square values in each tap control subgroup, and taps in the order of larger sum. Output the control subgroup number.

In the embodiments of FIGS. 9 to 11 described above, the control subgroup coefficient information includes a maximum absolute coefficient value for each tap control subgroup, a sum of coefficient absolute values for each tap control subgroup, or each tap control subgroup. Although the sum of coefficient square values for each subgroup is used, coefficient information other than these can also be used. Further, in the embodiment, the number of coefficients to be rearranged by one tap control is assumed to be 1, but it may be two or more.

FIG. 12 is a block diagram showing one embodiment of the eighth and ninth aspects of the present invention. Since the only difference from the conventional configuration of FIG. 15 is the operation of the maximum coefficient detection circuit 89 and the counter 88, the operation of the maximum coefficient detection circuit 89 and the counter 88 will be mainly described below. The maximum coefficient detection circuit 89 receives each tap coefficient value and the effective tap number output by the tap number calculation circuit 76, and detects the tap coefficient value having the maximum absolute value for each tap control subgroup.
The maximum absolute coefficient value is output to the counter 88, and the tap control subgroup number in which the maximum absolute coefficient value is rearranged in the order of the magnitude of the corresponding maximum absolute coefficient value is output to the control subgroup updating circuit 82. The counter 88 is a counter for counting the number of tap coefficient corrections, and outputs a control signal for changing the control group every time the number of times reaches the number determined by each maximum absolute coefficient value supplied from the maximum coefficient detection circuit 89. It is supplied to the group storage circuit 83 and the counter 86. That is, the control subgroup having a larger maximum absolute coefficient value has a longer time interval at which the counter 88 outputs the control group change signal.

FIG. 13 is a block diagram showing one embodiment of the eighth and tenth aspects of the present invention. The difference from the conventional configuration of FIG. 15 is that the control subgroup coefficient information extraction circuit 90
Since they are the same except for the counter and the counter 88, the operation of both will be mainly described below. The control subgroup coefficient information extraction circuit 90 receives each tap coefficient value and the effective tap number output by the tap number calculation circuit 76, obtains the sum of coefficient absolute values for each tap control subgroup, and outputs the corresponding tap control subvalue. It is supplied to the counter 88 together with the group number. The counter 88 is a counter that counts the number of tap coefficient corrections, and controls each time the number of times reaches the number determined by the sum of coefficient absolute values for each tap control subgroup supplied from the control subgroup coefficient information extraction circuit 90. Control signal for group change is stored in control subgroup storage circuit 8
3 and the counter 86. In other words, the control subgroup having a larger total coefficient absolute value has a longer time interval at which the counter 88 outputs the control group change signal.

FIG. 14 is a block diagram showing one embodiment of the eighth and eleventh aspects of the present invention. The only difference from the embodiment of FIG. 13 is the operation of the maximum coefficient detecting circuit 89 and the control subgroup coefficient information extracting circuit 90. The operation of both will be mainly described below. The control subgroup coefficient information extraction circuit 90 receives each tap coefficient value and the effective tap number output by the tap number calculation circuit 76, obtains the sum of coefficient absolute values for each tap control subgroup, and calculates the coefficient absolute values. The tap control subgroup numbers rearranged in the order of the sum are output to the control subgroup update circuit 82. The maximum coefficient detection circuit 89 detects the tap coefficient value having the maximum absolute value for each tap control subgroup,
The maximum absolute coefficient value and the corresponding tap control subgroup number are supplied to the counter 88. The counter 88 is a counter that counts the number of tap coefficient corrections, and controls a control signal for changing a control group every time the number of coefficient corrections reaches the number determined by each maximum absolute coefficient value supplied from the maximum coefficient detection circuit 89. The data is supplied to the subgroup memory circuit 83 and the counter 86. That is, the control subgroup having a larger maximum absolute coefficient value has a longer time interval at which the counter 88 outputs the control group change signal.

In each of the embodiments shown in FIGS. 12 to 14 described above, the time until the limited subgroup is changed from one subgroup to the next subgroup is determined according to the importance of each subgroup. With tap position control, the more important control subgroups become limited subgroups over time. Therefore, the tap coefficients are preferentially arranged for the response waveform portion of the impulse response, and high-speed convergence is achieved.

In the embodiments of FIGS. 13 and 14, the sum of coefficient absolute values for each tap control subgroup and the sum of coefficient square values for each tap control subgroup are used as the control subgroup coefficient information. Other coefficient information can also be used. Further, in each embodiment, the number of coefficients to be rearranged by one tap control is assumed to be 1, but 2
The above can be applied. Although the embodiment of the present invention has been described in detail above with respect to the example of the echo canceller, the present invention can be applied to a noise canceller, a howling canceller, an adaptive equalizer, etc. based on the same principle. Further, an algorithm other than the LMS algorithm used as an example can be applied to the tap coefficient correction algorithm.

[0051]

As described above, according to the present invention, the convergence time required for identifying an unknown system having a fixed delay or having a plurality of response waveform portions can be considerably shortened as compared with the prior art.

[Brief description of the drawings]

FIG. 1 is a block diagram showing first and second embodiments of the present invention.

2 is a signal timing diagram illustrating a response waveform portion to be identified in the embodiment of FIG.

FIG. 3 is a block diagram showing a configuration example of an adaptive filter of the embodiment of FIG.

FIG. 4 is a block diagram showing a configuration example of a coefficient generation circuit of the embodiment of FIG.

FIG. 5 is a block diagram showing another embodiment of the first and second inventions of the present invention.

FIG. 6 is a block diagram showing one embodiment of the fourth and sixth inventions of the present invention.

FIG. 7 is a block diagram showing another embodiment of the fourth and sixth inventions of the present invention.

FIG. 8 is a block diagram showing an embodiment of third and fifth inventions of the present invention.

FIG. 9 is a block diagram showing a first embodiment of the seventh invention of the present invention.

FIG. 10 is a block diagram showing a second embodiment of the seventh invention of the present invention.

FIG. 11 is a block diagram showing a third embodiment of the seventh invention of the present invention.

FIG. 12 is a block diagram showing an embodiment of the eighth and ninth inventions of the present invention.

FIG. 13 is a block diagram showing one embodiment of the eighth and tenth aspects of the present invention.

FIG. 14 is a block diagram showing an eighth and eleventh embodiment of the present invention.

FIG. 15 is a block diagram showing a configuration example of a conventional device.

FIG. 16 is a block diagram showing a configuration example of a coefficient generation circuit of FIG. 15;

[Explanation of symbols]

1,4 Input terminal 2,6 Output terminal 3 2-wire / 4-wire conversion circuit 5 Subtractor 7 Path switch 8,33,400 Adder 9,93,95,96,97,98,99,910,9
11, 912, 913, 914 Tap control circuit 20 _{1 to} 20 _N-1 delay element 30 ₁ to 30 _L , 50 ₁ to 50 _L coefficient generation circuit 31, 32, 40 ₁ to 40 _L multiplier 34, 72, 70 ₁ ˜70 _M memory circuit 71 minimum coefficient detection circuit 73,182 determination circuit 75 processing circuit 76 tap number calculation circuit 77 coefficient clear circuit 78,190 switch 74,79 distribution circuit 80 evaluation circuit 81,89 maximum coefficient detection circuit 82 control subgroup Update circuit 83 Control sub-group storage circuit 84 Control tap range calculation circuit 85 Coefficient square value calculation circuit 66, 67, 86, 87, 88 Counter 90 Control sub-group coefficient information extraction circuit 184 Coefficient absolute value calculation circuit 187 Synthesis circuit 188 ERLE evaluation Circuit 900 _{1 to} 900 _K Adaptive filter

Claims (23)

(57) [Claims] 1. A method according to claim 1, wherein a part of all taps of the adaptive filter is stored as an effective tap used for a product-sum operation, and a number of the effective tap is multiplied to the effective tap only. The unused taps are stored as invalid taps in the queue, and the coefficient corresponding to the number of the valid tap is corrected a predetermined number of times. Adapt the tap position such that the smallest tap is set as an invalid tap and the number is stored at the end of the queue, and the invalid tap number at the head of the queue is taken out and set as a new valid tap number. By using a plurality of adaptive filters to be controlled, and by making each of the adaptive filters identify at least one or more response waveform parts, it is possible to configure an image composed of a plurality of response waveform parts and a fixed delay. An unknown system identification method using an adaptive filter, which is characterized by identifying an unknown system having an impulse response. 2. The adaptive filter starts from an adaptive operation start.
Adaptive filter total taps, effective taps, coefficient position change
Predetermined coefficient using the number of new coefficient value updates required
Coefficient correction is performed using the first step size up to the number of times of correction, and thereafter, coefficient correction is performed using the second step size. The invalid tap number at the head of the queue has the maximum coefficient absolute. If it is in the vicinity of the valid tap number determined to be a value, it is set as a new valid tap number, and if it is not in the neighborhood, it is stored as the last in the queue without being validated, and from the head of the queue. The tap position adaptive control is performed by repeatedly comparing the tap number at the head of the queue with the valid tap number corresponding to the maximum coefficient absolute value until the fetched invalid tap number becomes new valid. A method for identifying an unknown system using the adaptive filter described in. 3. The adaptive filter comprises an adaptive motion of the filter.
From the start of the work, the total number of taps of the adaptive filter, the number of effective taps,
Predetermined using the number of coefficient value updates required to update the coefficient position
The first step size is used to perform the coefficient correction up to the number of times the coefficient is corrected, and thereafter, the second step size is used, and the queue is further divided into the first and second queues.
When the effective tap number having the smallest absolute value of the coefficient is in the vicinity of the effective tap number determined to have the maximum coefficient absolute value, the tail end of the first queue is formed. When the tap position is not in the vicinity, the tap position is added to the end of the second queue, and the invalid tap number at the head of the first queue is extracted and set as a new valid tap number. The method for identifying an unknown system using an adaptive filter according to claim 1, wherein the adaptive control is performed. 4. A coefficient correction is performed using the first step size from the start of the adaptive operation until a predetermined echo suppression amount is reached, and thereafter, a coefficient correction is performed using the second step size. The method for identifying an unknown system using the adaptive filter according to claim 2 or 3, wherein 5. A plurality of response waveform portions and a fixed delay using an adaptive filter that corrects and updates coefficients and tap positions according to an error signal obtained by subtracting an identification signal output by an adaptive filter from an output signal of an unknown system. An apparatus for identifying an unknown system by an adaptive filter for identifying an unknown system having an impulse response comprising: a plurality of adaptive systems for identifying at least one response waveform portion by receiving an input signal supplied to the unknown system. A filter, an adder that adds each output of the adaptive filter and outputs the identification signal, and a subtractor that subtracts the identification signal from an output signal of the unknown system to obtain the error signal, Each of the adaptive filters is adapted to identify a response waveform section having a fixed delay next to the fixed delay identified by itself. Receiving the information on the fixed delay identified by the adaptive filter and the information on the completion of identification, and identifying the response waveform part having the next largest fixed delay to the adaptive delay identified by the other adaptive filter identified by the self-identified fixed delay. An apparatus for identifying an unknown system by an adaptive filter, which supplies information about the fixed delay and identification completion information. 6. The adaptive filter selects a plurality of cascade-connected delay elements that provide a delay to an input signal of an unknown system, and selects and outputs a part of the plurality of delay signals output by the delay elements. A path switch, receiving the output signal of the path switch, the error signal and the coefficient clear signal,
A plurality of coefficient generating circuit for generating the updated tap coefficients, a plurality of multipliers for multiplying the output signal of the path switch and the tap coefficients respectively output from the coefficient generating circuit, the output signal of the multiplier An adder that adds and outputs the identification signal as an identification signal, a subtractor that subtracts the identification signal from the output signal of the unknown system to obtain the error signal, and receives the tap coefficient output by the coefficient generation circuit, A tap control circuit that generates a control signal of a path switch and the coefficient clear signal, wherein the coefficient generation circuit multiplies the delay signal and the error signal by a first multiplier; A second multiplier that multiplies the output signal of the multiplier by a predetermined constant, an adder that adds the output signal of the second multiplier and a feedback signal, and outputs the output signal of the adder. Tap coefficient Holds as, returned as the feedback signal to the adder is constituted by a storage circuit which is reset to zero by the coefficient clear signal, said tap control circuit is applied from the adaptive operation start of the filter
Update the total number of taps, effective taps, and coefficient positions of the adaptive filter
A predetermined number of times is calculated using the
A counter for changing the one control signal after und, the tap number storage circuit for storing an invalid tap number, the data
A decision circuit which receives the output of the step number storage circuit and outputs a signal satisfying a given decision condition and discards the others, and receives the output signal of the decision circuit and responds to a control signal from the counter to produce the storage circuit. A first switch that selectively returns feedback to a tap number calculation circuit that calculates the remaining tap numbers excluding the tap numbers held by the storage circuit from all the tap numbers of the adaptive filter, and outputs the calculated tap numbers to the path switch. Receiving a tap number output from the tap number calculation circuit and a tap coefficient output from the coefficient generation circuit, obtaining a tap number corresponding to a coefficient value having a minimum absolute value, and supplying the tap number to the storage circuit A coefficient clear circuit that supplies the coefficient clear signal to the coefficient generation circuit corresponding to the tap number output by the minimum coefficient detection circuit; and the tap number calculation circuit A maximum coefficient detection circuit that receives an output tap number and a tap coefficient output by the coefficient generation circuit and outputs an effective tap number having a maximum absolute coefficient value, and an effective tap number received from the maximum coefficient detection circuit and a determined constant and controlled tap range calculating circuit for calculating a control tap range as the determination conditions using the external
Or stored in the step size storage circuit.
A second switch that receives the stored first and second step sizes and switches from the first step size to the second step size according to a control signal from the counter, and outputs the second step size. An apparatus for identifying an unknown system using an adaptive filter according to claim 5. 7. The adaptive filter selects a plurality of cascade-connected delay elements that provide a delay to an input signal of an unknown system, and selects and outputs a part of a delay signal that is an output of the plurality of delay elements. receiving a path switch, and said output signal of said path switching error signal and the coefficient clear signal, further
A plurality of coefficient generating circuits for generating a new tap coefficient;
A plurality of multipliers each multiplying each tap coefficient output by the coefficient generation circuit and an output signal of the path switch,
An adder for adding the output signals of the multipliers and outputting as the identification signal, a subtracter for subtracting the identification signal from the output signal of the unknown system to obtain the error signal, and an update output by the coefficient generating circuit receiving said tap coefficients
After that, use the tap number corresponding to the coefficient value with the smallest absolute value.
Te, comprising a tap control circuit for generating a control signal and the coefficient clear signal of the path switches, the coefficient generating circuit includes a first multiplier for multiplying the error signal and the delay signal, said A second multiplier for multiplying the output signal of the first multiplier by a predetermined constant;
While holding the output signal of the multiplier of the as the tap coefficient, is fed back to the adder as the feedback signal, and is configured by a storage circuit reset to zero by the coefficient clear signal, the tap control circuit, From the start of the adaptive operation of the filter
Update the total number of taps, effective taps, and coefficient positions of the adaptive filter
A predetermined number of times is calculated using the
A counter that changes the control signal once after counting, a first storage circuit that stores an invalid tap number near an effective tap number whose coefficient absolute value is determined to be the maximum,
A second storage circuit for storing an invalid tap number that is not in the vicinity thereof, and a determination circuit for receiving an output signal of the first and second storage circuits, outputting a value included in a given value range, and discarding the others. And a tap number calculation circuit for calculating the remaining tap numbers from all tap numbers of the adaptive filter excluding the tap numbers held by the first and second storage circuits, and outputting the tap numbers to the path switch, and the tap numbers. A minimum coefficient detection circuit that receives the tap number output by the calculation circuit and the tap coefficient output by the coefficient generation circuit and obtains the tap number corresponding to the coefficient value with the smallest absolute value, the output signal of the determination circuit, and the minimum value. a combining circuit for combining an output signal of the coefficient detection circuit, before in response to a control signal from the counter
It is supplied from the minimum coefficient detection circuitry through the serial combining circuits
An evaluation circuit that evaluates whether or not the tap number belonging to the neighborhood that is determined by the value range, and the tap number that is supplied from the combining circuit is stored in the first or second memory according to the evaluation result of the evaluation circuit. Distribution circuit for distribution to circuits
When receives the tap number to an output of said minimum coefficient detector circuit
A coefficient clear circuit that supplies the coefficient clear signal only to a coefficient generating circuit corresponding to the tap number, and receives a tap number output by the tap number calculating circuit and a tap coefficient value output by the coefficient generating circuit, A maximum coefficient detection circuit that outputs an effective tap number having a maximum absolute coefficient, and a control tap range calculation that specifically calculates the neighborhood using the effective tap number received from the maximum coefficient detection circuit and a predetermined constant Circuit and externally supplied or
Switch for receiving the first and second step sizes stored in the step size storage circuit and switching from the first step size to the second step size in response to a control signal from the counter, and outputting the switch 6. The apparatus for identifying an unknown system using an adaptive filter according to claim 5, comprising: 8. An ERLE (echo suppression amount) evaluation circuit is provided in place of the counter, and the ERLE evaluation circuit receives an output signal of the unknown system and the error signal to calculate an echo suppression amount to calculate the echo suppression amount. The apparatus for identifying an unknown system by an adaptive filter according to claim 6 or 7, which generates a control signal when the amount of suppression reaches a predetermined threshold value. 9. A part of all the taps of the adaptive filter is stored as an effective tap used for the product-sum operation, and the coefficient multiplication of the product-sum operation is performed only on the effective tap to perform the product-sum operation. The number of unused taps is stored in a queue as an invalid tap, and each time the coefficient corresponding to the number of the valid tap is modified a predetermined number of times, the absolute value of the corresponding coefficient among the valid taps is changed. One tap that adaptively controls the tap position by storing the smallest tap as an invalid tap at the end of the wait queue, extracting the invalid tap number at the head of the queue, and setting it as a new valid tap number. When identifying an unknown system using an adaptive filter, the adaptive
Total number of taps, effective taps, and coefficients of the adaptive filter from the beginning
Predetermined using the number of coefficient value updates required for position update
Up to the coefficient correction circuit, the coefficient correction is performed using the first step size, and thereafter, the coefficient correction is performed using the second step size. The invalid tap number at the head of the queue is the coefficient absolute. If the value is in the vicinity of the effective tap number determined to be the maximum, it is set as a new effective tap number, and if it is not in the vicinity, it is stored as the last in the queue without being validated, Until the invalid tap number extracted from the top becomes new valid, the tap number at the top of the queue is repeatedly compared with the valid tap number corresponding to the maximum absolute coefficient value to perform adaptive control of the tap position. A method for identifying an unknown system using an adaptive filter, characterized by performing the following. 10. A part of all the taps of the adaptive filter is stored as an effective tap used for the product-sum operation, and the coefficient multiplication of the product-sum operation is applied only to the effective tap to perform the product-sum operation. Unused taps are stored as invalid taps in the queue, and the coefficient corresponding to the effective tap number is corrected a predetermined number of times each time, the absolute value of the corresponding coefficient among the effective tap numbers is minimized. One of the taps is stored as the invalid tap at the end of the queue, and the invalid tap number at the head of the queue is taken out to operate as a new valid tap number to adaptively control the tap position. Adaptive behavior of the filter when identifying an unknown system using the adaptive filter
From the start, the total number of taps, effective taps,
Predetermined using the number of coefficient value updates required for
The coefficient correction is performed using the first step size up to the specified coefficient correction circuit , the second step size is used thereafter, and the queue is further stored in the first and second 2nd step sizes.
The first queue having the smallest absolute value of the coefficient is located near the effective tap number determined to have the largest coefficient absolute value. And if it is not in the vicinity, it is added to the end of the second queue, and the invalid tap number at the head of the first queue is taken out and set as a new valid tap number, thereby setting the tap position. A method for identifying an unknown system by an adaptive filter, characterized by performing adaptive control of the system. 11. The coefficient correction is performed using the first step size from the start of the adaptive operation until a predetermined echo suppression amount is reached, and thereafter the coefficient correction is performed using the second step size. Claim 9 or 1 which performs
A method for identifying an unknown system using the adaptive filter described in 0. 12. An apparatus for identifying an unknown system using an adaptive filter that corrects a tap coefficient using an error signal obtained by subtracting an identification signal output from an adaptive filter from an output signal of the unknown system, wherein A plurality of cascade-connected delay elements that give a delay to an input signal, a path switch that selects and outputs a part of the delay signal that is the output of the delay element, an output signal of the path switch, and the error signal in response to the coefficient clear signal and, further
A plurality of coefficient generating circuits for generating a new tap coefficient;
A plurality of multipliers each multiplying each tap coefficient output by the coefficient generation circuit and an output signal of the path switch,
An adder for outputting the identification signal adding multiplier output the plurality of, a subtracter from said output signal of the unknown system by subtracting the identification signal to obtain the error signal, the said coefficient generating circuit outputs After receiving the updated tap coefficient,
Absolute value using the minimum tap number corresponding to the coefficient value, comprising a tap control circuit for generating a control signal and the coefficient clear signal of the path switches, the coefficient generating circuit, the delay signal and the error A first multiplier for multiplying the signal by a signal; a second multiplier for multiplying an output signal of the first multiplier by a predetermined constant; an output signal of the second multiplier and a feedback signal And a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. And the tap control circuit starts the adaptive operation from the start of the adaptive operation of the filter.
Update the total number of taps, effective taps, and coefficient positions of the adaptive filter
A predetermined number of times is calculated using the
A counter for changing the one control signal after und, the tap number storage circuit for storing an invalid tap number, the data
A decision circuit which receives the output of the check number storage circuit and outputs the one which satisfies the given decision condition and discards the others, and selects the storage circuit according to the control signal of the counter upon receiving the output of the decision circuit A first switch that is fed back, and the remaining tap numbers except the tap numbers held by the memory circuit group from all tap numbers of the adaptive filter are calculated,
The tap number calculation circuit for outputting to the path switch, the tap number output by the tap number calculation circuit, and the coefficient value output by the plurality of coefficient generation circuits are received, and the tap number corresponding to the coefficient value having the smallest absolute value is received. The tap coefficient calculating circuit outputs the minimum coefficient detecting circuit for obtaining and supplying the storage circuit, the coefficient clear circuit for supplying the coefficient clear signal to the coefficient generating circuit corresponding to the tap number output by the minimum coefficient detecting circuit, and the tap number calculating circuit. Receiving a tap number and a tap coefficient output from the plurality of coefficient generating circuits, and outputting a valid tap number having a maximum absolute coefficient value; a valid tap number received from the maximum coefficient detecting circuit; A control tap range calculation circuit for calculating a control tap range as the determination condition using a predetermined constant, and first and second step servers. Supplied from outside by a control signal from said counter receives's, or steps Sai
Device for identifying an unknown system by an adaptive filter, which comprises a second switch for switching from a first step size stored in a memory circuit to a second step size and outputting the second step size. 13. An apparatus for identifying an unknown system using an adaptive filter, the coefficient of which is corrected by using an error signal obtained by subtracting an identification signal output from an adaptive filter from an output signal of the unknown system. A plurality of cascaded delay elements that give a delay to the signal, a path switch that selects and outputs a part of the delay signal that is the delay element output, an output signal of the path switch, the error signal, and a coefficient. A plurality of coefficient generation circuits that receive the clear signal and generate updated tap coefficients, and a plurality of coefficient multipliers that multiply each tap coefficient output by the coefficient generation circuit and the output signal of the path switch. A multiplier, an adder that adds the outputs of the multipliers and outputs the identification signal, and subtracts the identification signal from the output signal of the unknown system to obtain the error signal. And adder, the coefficient generation circuit outputs
A tap control circuit that receives the updated tap coefficient and generates a control signal of the path switch and the coefficient clear signal is provided, and the coefficient generation circuit multiplies the delay signal and the error signal. A first multiplier, a second multiplier that multiplies the output of the first multiplier by a predetermined constant, and an adder that adds the output signal of the second multiplier and the feedback signal And a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. Open the adaptive operation of the filter
Total number of taps, effective taps, and coefficients of the adaptive filter from the beginning
Predetermined using the number of coefficient value updates required for position update
A counter that outputs a control signal after counting the number of times , a first memory circuit that stores an invalid tap number near the valid tap number that is determined to have the largest coefficient absolute value, and an invalid counter that is not in the neighborhood. A second memory circuit for storing a tap number; a decision circuit for receiving the outputs of the first and second memory circuits and outputting those included in a given value range ; discarding the others; A tap number calculation circuit that calculates the remaining tap numbers from the tap numbers excluding the tap numbers held by the first and second storage circuits and outputs the tap numbers to the path switch, and the tap numbers output by the tap number calculation circuit A minimum coefficient detecting circuit that receives the tap coefficient output from the plurality of coefficient generating circuits and obtains a tap number corresponding to the coefficient value having the smallest absolute value; an output signal of the determination circuit; Wherein the combining circuit for combining the output signal of the coefficient detection circuit, the output signal of the decision circuit and counter
In response to a control signal from the
The tap number supplied from the small coefficient detection circuit is in the above range.
An evaluation circuit that evaluates whether it belongs to the neighborhood determined by the above, and a distribution circuit that distributes the tap number supplied from the synthesis circuit to the first or second storage circuit according to the evaluation result of the evaluation circuit. And a tap number corresponding to the tap number received from the minimum coefficient detection circuit.
A coefficient clear circuit that supplies the coefficient clear signal only to the number generation circuit, a tap number output by the tap number calculation circuit, and the tap coefficient output by the coefficient generation circuit, and an effective value having a maximum absolute coefficient value. A maximum coefficient detection circuit that outputs a tap number, a control tap range calculation circuit that specifically calculates the range in the vicinity by using an effective tap number received from the maximum coefficient detection circuit and a predetermined constant, Is supplied externally by a control signal from the counter in response to the first and second step sizes , or
Is a switch for switching from a first step size stored in a step size storage circuit to a second step size for output, and an apparatus for identifying an unknown system by an adaptive filter. 14. An ERLE evaluation circuit is provided instead of the counter, and the ERLE evaluation circuit receives an output signal of the unknown system and the error signal to calculate an echo suppression amount, and the echo suppression amount is predetermined. An apparatus for unknown system identification by an adaptive filter according to claim 12 or 13, which generates a control signal when a predetermined threshold is reached. 15. An apparatus for identifying an unknown system by using an adaptive filter, the coefficient of which is corrected by using an error signal obtained by subtracting an identification signal output by the adaptive filter from an output signal of the unknown system. A plurality of cascaded delay elements that give a delay to the signal, a path switch that selects and outputs a part of the delay signal that is the delay element output, an output signal of the path switch, the error signal, and a coefficient. A plurality of coefficient generation circuits that generate tap coefficients in response to the clear signal, a plurality of multipliers that multiply each tap coefficient output by the coefficient generation circuit and the output signal of the path switch, and the multipliers. An adder for adding outputs to output the identification signal; a subtractor for subtracting the identification signal from the output signal of the unknown system to obtain the error signal; A tap control circuit that receives the tap coefficient output from the coefficient generation circuit and generates the control signal of the path switch and the coefficient clear signal, the coefficient generation circuit multiplying the delay signal and the error signal. And a second multiplier that multiplies the output signal of the first multiplier by a predetermined constant, and the output signal of the second multiplier and the feedback signal. An adder, and holding an output signal of the adder as the tap coefficient;
The tap control circuit is configured to include a memory circuit which is fed back to the adder as the feedback signal and is reset to zero by the coefficient clear signal, and the tap control circuit includes a tap control subgroup including a plurality of consecutive tap numbers. First to store control subgroup numbers corresponding to one-to-one according to a selection order
Of the memory circuit, a memory circuit group composed of a plurality of memory circuits, and the remaining tap numbers excluding the tap numbers held by the memory circuit group from all tap numbers of the adaptive filter are calculated, A tap number calculation circuit for output, a minimum coefficient detection circuit for receiving a tap number output by the tap number calculation circuit and a tap coefficient output by the coefficient generation circuit, and obtaining a tap number corresponding to a coefficient value having a minimum absolute value An evaluation circuit for obtaining a tap control subgroup corresponding to the tap number supplied from the minimum coefficient detection circuit, and one circuit in the memory circuit group corresponding to the tap control subgroup supplied from the evaluation circuit. A distribution circuit for transmitting the tap number supplied from the minimum value detection circuit, and a coefficient generation corresponding to the tap number output from the minimum coefficient detection circuit. A coefficient clear circuit that supplies the coefficient clear signal to a path, and a processing circuit that extracts a tap number from one circuit corresponding to the control subgroup number output from the first memory circuit in the memory circuit group and discards the tap number. , A control subgroup number arranged in the order of the magnitude of the maximum absolute coefficient value for each control subgroup, receiving the tap number output by the tap number calculation circuit and the coefficient value output by the plurality of coefficient generation circuits. And a first counter for supplying an instruction signal for changing the control subgroup number to the first storage circuit each time the number of times of coefficient correction reaches a predetermined number. A second counter that outputs a control subgroup order change signal each time the instruction signal of the first counter is output a predetermined number of times, and an output of the second counter Wherein the further signal maximum coefficient detector circuit device of the unknown system identification by the adaptive filter, characterized in that it is composed of a writing control subgroup updating circuit in the first storage circuit control sub group number output. 16. A coefficient absolute value calculation for calculating a sum of absolute values of coefficients for each of said control subgroups and outputting a control subgroup number arranged in the order of the magnitude of said sum instead of said maximum coefficient detection circuit. The apparatus for identifying an unknown system by an adaptive filter according to claim 15, comprising a circuit. 17. A coefficient square value calculation for calculating a sum of coefficient square values for each of the control sub groups and outputting control sub group numbers arranged in order of magnitude of the sum instead of the maximum coefficient detecting circuit. The apparatus for identifying an unknown system by an adaptive filter according to claim 15, comprising a circuit. 18. A part of all taps of an adaptive filter is stored as an effective tap used for a product-sum operation, and its number is stored. The number of unused taps is stored in a queue as an invalid tap number, and each time the coefficient corresponding to the number of the valid tap is corrected a predetermined number of times, the absolute value of the corresponding coefficient among the valid tap numbers The tap with the smallest value is stored as the invalid tap at the end of the queue, the invalid tap number at the head of the queue is extracted, and the invalid tap number is determined to have the maximum coefficient absolute value. If it is near the tap number, it is set as a new effective tap number.If it is not near the tap number, it is not valid and stored at the end of the queue and taken out from the head of the queue. An adaptive filter for adaptively controlling the tap position by operating so as to repeatedly compare the tap number at the head of the queue with the valid tap number corresponding to the maximum absolute coefficient value until the invalid tap number is newly valid. When performing identification of an unknown system using, all taps of the adaptive filter are divided into a plurality of tap control subgroups each of which has an equal number of consecutive taps, and one tap selected among the tap control subgroups. The tap number belonging to the tap control subgroup is defined as the neighborhood,
The selected tap control subgroup to its subgroup
The tap position is adaptively controlled by changing in accordance with the selection order for each coefficient correction number specified for each tap, and determining the selection order and the coefficient correction number using information on the effective tap coefficients in each tap control subgroup. A method for identifying an unknown system by means of an adaptive filter. 19. An apparatus for identifying an unknown system using an adaptive filter, the coefficient of which is corrected by using an error signal obtained by subtracting an identification signal output from the adaptive filter from an output signal of the unknown system. A plurality of cascaded delay elements for delaying a signal, a path switch for selecting and outputting a part of a delay signal which is an output of the delay element, an output signal of the path switch, the error signal, and a coefficient. A plurality of coefficient generation circuits that generate tap coefficients in response to the clear signal, a plurality of multipliers that multiply each tap coefficient output by the coefficient generation circuit and the output signal of the path switch, and the multiplier outputs And an adder that outputs the identification signal by subtracting the identification signal from the output signal of the unknown system to obtain the error signal. A tap control circuit that receives a tap coefficient output by a number generation circuit and generates a control signal of the path switch and the coefficient clear signal; and the coefficient generation circuit converts the delay signal and the error signal. A first multiplier for multiplication, a second multiplier for multiplying an output signal of the first multiplier by a predetermined constant, and an output signal of the second multiplier and a feedback signal And a storage circuit that holds the output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal, The tap control circuit has a first-in first-out structure.
And a tap number calculation circuit for calculating the remaining tap numbers excluding the tap numbers held by the first storage circuit from all tap numbers of the adaptive filter and outputting the tap numbers to the path switch, and the tap numbers. A minimum coefficient detection circuit that receives the tap number output from the calculation circuit and the tap coefficient output from the coefficient generation circuit and transmits the tap number corresponding to the coefficient value having the smallest absolute value to the first storage circuit, and the minimum coefficient detection circuit. A coefficient clearing circuit for supplying the coefficient clearing signal to the coefficient generating circuit corresponding to the tap number output by the coefficient detecting circuit, and a control sub-group corresponding one-to-one with a tap control sub-group composed of a plurality of continuous tap numbers A second storage circuit for storing the group numbers in accordance with the selection order, and a control subgroup number output from the second storage circuit, Control tap range calculation circuit that outputs the upper limit and the lower limit of the tap number belonging to the tap number, and whether the tap number extracted from the first storage circuit is within the range of the upper limit and the lower limit. If not, the control circuit receives the tap number returned to the first storage circuit, the tap number output by the tap number calculation circuit, and the tap coefficient output by the coefficient generation circuit, and each control A maximum coefficient detection circuit that detects the maximum absolute coefficient value for each subgroup and outputs the control subgroup number and the maximum absolute coefficient value arranged in the order of the size of the maximum absolute coefficient value, and the maximum coefficient detection circuit A first clock for supplying an instruction signal for changing a control subgroup number to the second storage circuit every time the coefficient correction number reaches a predetermined number in response to the maximum absolute coefficient value Data and, second the instruction signal from the first counter outputs a change signal of the control sub-group sequence each time it is number outputted predetermined
And a control subgroup updating circuit that writes a control subgroup number output by the maximum coefficient detection circuit into the second storage circuit in response to a change signal output by the second counter. The unknown system identification by an adaptive filter is characterized in that the counter determines the count number for supplying the indication signal by using the maximum absolute coefficient value corresponding to each control subgroup number received from the maximum coefficient detection circuit. Equipment. 20. An apparatus for identifying an unknown system using an adaptive filter that corrects a coefficient using an error signal obtained by subtracting an identification signal output by an adaptive filter from an output signal of the unknown system, wherein the input of the unknown system is provided. A plurality of cascaded delay elements that give a delay to the signal, a path switch that selects and outputs a part of the delay signal that is the delay element output, an output signal of the path switch, the error signal, and a coefficient. A plurality of coefficient generation circuits for generating tap coefficients in response to the clear signal; a plurality of multipliers for multiplying each of the tap coefficients output by the coefficient generation circuit with an output signal of the path switch; An adder that adds outputs and outputs the identification signal; a subtractor that subtracts the identification signal from the output of the unknown system to obtain the error signal; and the coefficient A tap control circuit that receives the tap coefficient output by the generation circuit and generates the control signal of the path switch and the coefficient clear signal, wherein the coefficient generation circuit multiplies the delay signal by the error signal A first multiplier, a second multiplier that multiplies an output signal of the first multiplier by a predetermined constant, and an addition that adds an output signal of the second multiplier and a feedback signal And a storage circuit that holds an output signal of the adder as the tap coefficient, feeds back to the adder as the feedback signal, and is reset to zero by the coefficient clear signal. The circuit has a first-in first-out structure.
A tap number calculating circuit for calculating the remaining tap numbers excluding the tap numbers held in the first storage circuit from all the tap numbers of the adaptive filter, and outputting the tap numbers to the path switch; A minimum coefficient detection circuit that receives a tap number output by a calculation circuit and a tap coefficient value output by the coefficient generation circuit, and transmits a tap number corresponding to a coefficient value having a minimum absolute value to the first storage circuit; A coefficient clearing circuit for supplying the coefficient clearing signal to a coefficient generating circuit corresponding to the tap number output by the minimum coefficient detecting circuit; A second storage circuit that stores the group numbers according to the selection order and a control subgroup number output from the second storage circuit are received, and the control subgroup numbers are received. A control tap range calculation circuit that outputs an upper limit and a lower limit of a tap number belonging to the number and a tap number extracted from the first storage circuit determine whether or not the tap number is within the range of the upper limit and the lower limit; When there is no tap number, a determination circuit that feeds back the extracted tap number to the first storage circuit, a tap number output by the tap number calculation circuit, and a tap coefficient output by the coefficient generation circuit are received. A maximum coefficient detection circuit that detects a coefficient value having a maximum absolute value for each control subgroup, and outputs a control subgroup number arranged in the order of the magnitude of the maximum absolute coefficient value; and a tap number output by the tap number calculation circuit. And a control subgroup coefficient information extraction circuit that outputs coefficient information for each control subgroup in response to the tap coefficient output by the coefficient generation circuit, and the control subgroup A first counter that receives the coefficient information from the number information extraction circuit and supplies an instruction signal for changing the control subgroup number to the second storage circuit each time the number of times the coefficient is modified reaches a predetermined number; A second counter that outputs a control subgroup order change signal each time the instruction signal from the first counter is output a predetermined number of times, and the maximum coefficient by the change signal output by the second counter A control subgroup updating circuit for writing a control subgroup number output from the detection circuit into the second storage circuit; Of unknown system identification by an adaptive filter characterized in that the number of counts for supplying the indication signal is determined by using coefficient information corresponding to the group number Location. 21. An apparatus for identifying an unknown system by using an adaptive filter that corrects a coefficient using an error signal obtained by subtracting an identification signal output by an adaptive filter from an output signal of the unknown system. A plurality of cascaded delay elements that give a delay to the signal, a path switch that selects and outputs a part of the delay signal that is the delay element output, an output signal of the path switch, the error signal, and a coefficient. A plurality of coefficient generation circuits that generate tap coefficients in response to the clear signal, a plurality of multipliers that multiply each tap coefficient output by the coefficient generation circuit and the output signal of the path switch, and the multipliers. An adder that adds the output and outputs the identification signal, and a subtractor that subtracts the identification signal from the output signal of the unknown system to obtain the error signal; A tap control circuit that receives a tap coefficient value output from a coefficient generation circuit and generates a control signal for the path switch and the coefficient clear signal, the coefficient generation circuit including the delay signal and the error signal. A first multiplier for multiplying by, a second multiplier for multiplying an output signal of the first multiplier by a predetermined constant, and an output signal of the second multiplier and a feedback signal. And an adder for adding and holding an output signal of the adder as the tap coefficient, feeding back to the adder as the feedback signal, and a storage circuit reset to zero by the coefficient clear signal, The tap control circuit has a first-in first-out structure.
And a tap number calculation circuit for calculating the remaining tap numbers excluding the tap numbers held by the first storage circuit from all tap numbers of the adaptive filter and outputting the tap numbers to the path switch, and the tap numbers. A minimum coefficient detection circuit that receives the tap number output from the calculation circuit and the tap coefficient output from the coefficient generation circuit and transmits the tap number corresponding to the coefficient value having the smallest absolute value to the first storage circuit, and the minimum coefficient detection circuit. A coefficient clearing circuit that supplies the coefficient clearing signal to a coefficient generating circuit corresponding to the tap number output by the coefficient detecting circuit;
A second storage circuit that stores a control subgroup number corresponding to the tap control subgroup composed of a plurality of consecutive tap numbers on a one-to-one basis according to a selection order, and a control subgroup number output by the second storage circuit In response, the control tap range calculation circuit that outputs the upper limit and the lower limit of the tap numbers belonging to the control subgroup number, and whether the tap numbers extracted from the first storage circuit are within the range of the upper limit and the lower limit. And a tap number output from the tap number calculation circuit and a tap number output from the plurality of coefficient generation circuits when the tap number is not within the range. Control subgroup coefficient information that outputs the control subgroup number in the order determined by the coefficient information for each control subgroup In response to the output circuit, the tap number output by the tap number calculation circuit and the tap coefficient output by the coefficient generation circuit, the maximum absolute coefficient value is detected for each control subgroup, and the maximum absolute coefficient value is detected. And a maximum coefficient detection circuit for outputting a control subgroup number, and the second storage each time the number of times of coefficient correction reaches a predetermined number by receiving the maximum absolute coefficient value and the control subgroup number from the maximum coefficient detection circuit. A first counter for supplying an indication signal for changing a control subgroup number to the circuit;
Each time the instruction signal of the first counter is output a predetermined number of times, a second counter that outputs a change signal of the control subgroup order, and a change signal that is output by the second counter And a control sub-group updating circuit for writing the control sub-group number output from the group coefficient information extracting circuit to the second storage circuit, wherein the first counter is provided for each control sub-group received from the maximum coefficient detecting circuit. An apparatus for identifying an unknown system using an adaptive filter, wherein a count number for supplying the instruction signal is determined using a maximum coefficient absolute value corresponding to a group number. 22. The apparatus for identifying an unknown system using an adaptive filter according to claim 20, wherein a sum of absolute values of coefficients is used as the coefficient information. 23. The apparatus for identifying an unknown system using an adaptive filter according to claim 20, wherein a sum of coefficient square values is used as the coefficient information.