WO2021124413A1 - フィルタ装置 - Google Patents
フィルタ装置 Download PDFInfo
- Publication number
- WO2021124413A1 WO2021124413A1 PCT/JP2019/049224 JP2019049224W WO2021124413A1 WO 2021124413 A1 WO2021124413 A1 WO 2021124413A1 JP 2019049224 W JP2019049224 W JP 2019049224W WO 2021124413 A1 WO2021124413 A1 WO 2021124413A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- signal
- coefficient
- value
- filter
- unit
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/06—Non-recursive filters
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0219—Compensation of undesirable effects, e.g. quantisation noise, overflow
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0223—Computation saving measures; Accelerating measures
- H03H17/0227—Measures concerning the coefficients
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0223—Computation saving measures; Accelerating measures
- H03H17/0227—Measures concerning the coefficients
- H03H17/023—Measures concerning the coefficients reducing the wordlength, the possible values of coefficients
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0223—Computation saving measures; Accelerating measures
- H03H17/0233—Measures concerning the signal representation
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H2017/0072—Theoretical filter design
- H03H2017/0081—Theoretical filter design of FIR filters
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H17/00—Networks using digital techniques
- H03H17/02—Frequency selective networks
- H03H17/0219—Compensation of undesirable effects, e.g. quantisation noise, overflow
- H03H2017/022—Rounding error
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03H—IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
- H03H2218/00—Indexing scheme relating to details of digital filters
- H03H2218/02—Coefficients
Definitions
- the present invention relates to a filter device that performs filtering on a digital signal.
- a finite impulse response filter (hereinafter referred to as FIR (Finite Impulse Response) filter), which is a type of digital filter, outputs a signal obtained by convolving a filter coefficient into an input signal.
- FIR Finite Impulse Response
- Patent Document 1 describes an invention for reducing the quantization noise of the filter coefficient.
- each filter coefficient of the digital filter is multiplied by an adjustment magnification value to perform adjustment, and the adjusted filter coefficient is multiplied by the input data.
- the input data multiplied by the adjusted filter coefficient is divided by the multiplicand to reduce the error caused by the rounding of the filter coefficient in a long digital filter.
- the input signal to the FIR filter is x (n)
- the output signal is y (n)
- the input signal x (n) and each filter coefficient h n include quantization noise, respectively.
- the output signal y (n) is given by the following equation (1). It is represented by.
- the filter coefficient h k is a large value among all the filter coefficients (h 0 , h 1 , ..., H N-1 ) and the range in which the filter coefficient is expressed is small
- the preliminary magnification adjustment value c k Cannot be a large value
- the second term representing the calculation error due to multiplication in the equation (2) cannot be a small value. That is, the calculation error cannot be reduced.
- the present invention has been made in view of the above, and an object of the present invention is to obtain a filter device capable of reducing deterioration of filter characteristics due to calculation error.
- the filter device includes a plurality of delay units connected in series, which give a delay when a signal is input and output it as a delay signal, in advance. It includes a plurality of multiplication units for multiplying a delay signal by a filter coefficient generated based on a predetermined value and a magnification adjustment value. Further, the filter device is a quotient obtained by dividing a multiplication result exceeding the maximum value by the maximum value when the multiplication result obtained by multiplying a predetermined value by a magnification adjustment value exceeds the maximum value of the expression range of the filter coefficient. Is provided as a coefficient adjustment unit that outputs the above as a coefficient adjustment value.
- the filter device is a signal conversion unit that adds and outputs a signal after multiplication of filter coefficients output by a plurality of multiplication units and an adjusted signal obtained by adjusting a corresponding delay signal using a coefficient adjustment value.
- a division unit that generates an output signal by dividing the signal output by the signal conversion unit by the magnification adjustment value.
- the filter device according to the present invention has an effect that deterioration of filter characteristics due to calculation error can be reduced.
- the figure which shows an example of the circuit structure of the filter apparatus which concerns on Embodiment 1 of this invention The figure which shows the structural example of the signal conversion part which includes the filter apparatus which concerns on Embodiment 1.
- the figure which shows an example of the circuit structure of the filter apparatus which concerns on Embodiment 2 of this invention The figure which shows the structural example of the expression conversion part included in the filter apparatus which concerns on Embodiment 2.
- FIG. 1 is a diagram showing an example of a circuit configuration of a filter device according to a first embodiment of the present invention.
- the filter device 100 according to the first embodiment is a digital filter, specifically an FIR filter.
- the filter device 100 is realized by dedicated hardware, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) , Or a combination of these.
- the filter device 100 includes a plurality of delay units 1-1 to 1-N that delay an n-bit input signal by a predetermined time, and an input signal to the filter device 100 and a delay unit 1-.
- Multiplying units 2-0 to 2-N that multiply each delay signal, which is a signal output from 1 to 1-N, with the filter coefficient corresponding to each of these signals, and the mth (m is [0 to N]].
- the multiplication signal of each signal output from the coefficient adjusting unit 3 is a signal obtained by multiplying the signals output from the coefficient adjusting unit 3 with each other.
- a plurality of delay units 1-1 to 1-N are connected in series, and a signal x (n) is input to the delay unit 1-1.
- the delay signals output by the delay units 1-1 to 1- (N-1) are input to the delay units 1-2 to 1-N in the subsequent stage, respectively, and the multiplication units 2-1 to 2- (N-). It is input to 1).
- the delay signal output by the delay unit 1-N is input to the multiplication unit 2-N.
- the signal x (n) is input to the multiplication unit 2-0.
- a corresponding filter coefficient is input to each of the multiplication units 2-0 to 2-N.
- each filter coefficient to be multiplied by the input signal x (n) or the delay signal is a pre-designed value (hereinafter referred to as a design value) h 0 , h 1 , ..., H N and a magnification adjustment value A. It is a value generated based on. More specifically, each filter coefficient is a value obtained by multiplying a predetermined design value by a magnification adjustment value, or a value calculated based on a value obtained by multiplying a design value by a magnification adjustment value. Become. That is, the value of each filter coefficient is Ah 0 , Ah 1 , ..., Ah N or a value calculated based on these.
- the first value obtained by multiplying the design value by the magnification adjustment value is equal to or less than the maximum value C max of the expression range of the filter coefficient.
- the second value calculated based on the first value is used as the filter coefficient.
- Ah m % C max is the remainder of Ah m ⁇ C max.
- the first value Ah m exceeds the maximum value C max , the first value Ah m is divided by the maximum value C max to obtain the remainder which is the second value, and this is used as the filter coefficient.
- the remainder of Ah m ⁇ C max is equal to the first value Ah m. Therefore, it can be said that all the filter coefficients are remainders obtained by multiplying the design value by the magnification adjustment value and dividing the value obtained by dividing by the maximum value C max of the expression range of the filter coefficient.
- the coefficient adjustment value [Ah m ] is an integer value. That is, the coefficient adjustment value [Ah m ] is the quotient of Ah m ⁇ C max.
- FIG. 2 is a diagram showing a configuration example of a signal conversion unit 4 included in the filter device 100 according to the first embodiment.
- the signal conversion unit 4 includes addition units 41-1 to 41- (N + 1) that take the outputs from the multiplication units 2-0 to 2-N as inputs and add them while increasing the bit length.
- a multiplication unit 42 to be added, and an addition unit 43 for adding the output of the addition unit 41- (N + 1) and the output of the multiplication unit 42 are provided.
- the multiplication unit 42 and the addition unit 43 are not limited to one, and it is conceivable that there are more than the number of filter coefficients beyond the expression range by multiplying the magnification adjustment value A. In this way, the signal conversion unit 4 obtains the output represented by the following equation (3).
- the coefficient adjusting unit 3 duplicates the m-th delay output by the number equal to the coefficient adjusting value [Ah m ] and outputs it to the signal conversion unit 4, so that the multiplication in the signal conversion unit 4 is unnecessary and only addition is performed.
- the configuration to be performed is also conceivable.
- the division unit 5 divides the output of the signal conversion unit 4 represented by the equation (3) by the above-mentioned magnification adjustment value A which is a multiplicand, and finally rounds the output so that the bit length is the same as x (n). By doing so, the output signal y (n) of the filter device 100 represented by the following equation (4) is generated.
- the division unit 5 can also be realized by bit shifting by setting the magnification adjustment value A to be multiplied by the filter coefficient as a power of 2.
- the output signal y (n) of the filter device 100 is as shown in the following equation (5).
- equation (5) em k indicates the calculation error at the time of multiplication.
- the total calculation error can be divided by the same magnification adjustment value A, and the calculation error can be reduced. Further, since the value of the filter coefficient multiplied by the magnification adjustment value A is allowed to exceed the maximum value of the expression range, it is possible to improve the filter characteristics while keeping the number of quantization bits small.
- the filter device 100 is composed of the plurality of delay units 1-1 to 1-N for delaying the input signal, the input signals x (n), and the delay units 1-1 to 1-N.
- Multiplying units 2-0 to 2-N that multiply each output delay signal by the design value and the filter coefficient obtained by multiplying the magnification adjustment value A, and the filter coefficient exceeds the maximum value C max of the expression range.
- the coefficient adjustment unit 3 that generates a coefficient adjustment value based on the magnification adjustment value A and the maximum value C max of the expression range and outputs the generated coefficient adjustment value and the corresponding delay signal, and the multiplication unit 2-0 to 2
- It includes a signal conversion unit 4 that adds and converts the expression, and a division unit 5 that divides the expression-changed signal output from the signal conversion unit 4 by the magnification adjustment value A. According to the filter device 100, it is possible to reduce the deterioration of the filter characteristics due to the calculation error.
- Embodiment 2 Next, the filter device according to the second embodiment, specifically, the filter device realized by using the probabilistic calculation will be described.
- the stochastic calculation has a feature that the circuit scale of the product-sum calculation can be reduced, and the filter device can be miniaturized by using the stochastic calculation.
- the correlation of the random number sequence between the numerical values input to the arithmetic circuit becomes a factor and the calculation error. Occurs. Reducing this calculation error is a problem when using stochastic calculation.
- FIG. 3 is a diagram showing an example of the circuit configuration of the filter device 100a according to the second embodiment.
- the filter device 100a according to the second embodiment includes an expression conversion unit 6 (expressed as SNG (Stochastic Number Generator) 6 in FIG. 3) that converts an input signal x (n) of binary expression n bits into a probabilistic expression. , A plurality of delay units 1-1 to 1-N that delay the input signal from the expression conversion unit 6 by a predetermined time, and output signals from the expression conversion unit 6 and delay units 1-1 to 1-N.
- SNG Stochastic Number Generator
- a multiplication unit 7-0 to 7-N that multiplies each delay signal and a filter coefficient corresponding to each of these signals, and a coefficient adjustment unit 8 that outputs the m-th delay signal as an input to a signal conversion unit 9 described later.
- Each of the multiplication units 7-0 to 7-N and the multiplication signal of each signal output from the coefficient adjustment unit 8 are added, and the signal conversion unit 9 for converting the expression and the output signal from the signal conversion unit 9 are combined. It is composed of a division unit 10 that divides by a multiplicand.
- the stochastic expression signal output by the expression conversion unit 6 is input to the head delay unit 1-1 among the plurality of delay units 1-1 to 1-N.
- FIG. 4 is a diagram showing a configuration example of the expression conversion unit 6 included in the filter device 100a according to the second embodiment.
- the expression conversion unit 6 generates a random number generation unit 11 that generates a random number, an input signal x (n) to the filter device 100a that is an input value to the expression conversion unit 6, and a random number value input from the random number generation unit 11.
- a comparator 12 for comparison is provided. A random value is input to the terminal Z of the comparator 12, and a signal x (n) is input to the terminal W.
- the comparator 12 outputs 1 if the input value of the binary representation is larger than the random number value input from the random number generator 11, and 0 if the input value is equal to or less than the random number value.
- the bit length used for stochastic expression hereinafter referred to as SN length
- the input value is 0.3
- the random number value is a uniform random number in the range of 0 to 1, it is obtained.
- the value with the largest expected value as the value of the stochastic expression to be obtained is a numerical sequence in which 3 bits out of 10 bits are 1 and the remaining 7 bits are 0.
- the expression conversion unit 6 operates as described above.
- the expression conversion unit 6 may be configured to generate a stochastic expression value corresponding to a binary expression value in advance, hold it in a memory as a table, and read the value corresponding to the input value directly from the memory.
- the filter coefficients are multiplied in the multiplication units 7-0 to 7-N.
- the filter coefficient used at this time also needs to be converted into a stochastic expression.
- the filter coefficients Ah 0 , Ah 1 , ..., ⁇ Ah 0m ⁇ , ..., Ah N-1 , and Ah N shown in FIG. 3 are values that have been converted into stochastic expressions in advance, but the same circuit as the expression conversion unit 6 May be separately provided as a coefficient conversion unit, and the coefficient conversion unit may convert the filter coefficient into a stochastic expression.
- the multiplication units 7-0 to 7-N are expressed by the AND operation, but this is the case of the unipolar expression in which the range of the stochastic expression is 0 to 1, and the range of the stochastic expression is set.
- the bipolar expression set to -1 to 1 it is expressed by the XNOR operation.
- the signal conversion unit 9 performs parallel addition with the output from the multiplication units 7-0 to 7-N and the output from the coefficient adjustment unit 8 as inputs.
- the parallel addition is an output value expressed in binary by dividing the number of "1" included in all the input probabilistic expression values by the SN length. For example, when the SN length is 10, and the input probabilistic expression values are three values of "0000100100” (0.2), “1010111110” (0.7), and "0010101111” (0.6), The output value is 1.5, which is obtained by dividing 15 which is the number of "1” by 10 which is the SN length.
- the division unit 10 divides the output of the signal conversion unit 9 by the above-mentioned magnification adjustment value A which is a multiplicand, and finally performs rounding so that the bit length is the same as x (n). ) Is generated as the output signal y (n) of the filter device 100.
- the division unit 10 can also be realized by a bit shift by setting the magnification adjustment value A to be multiplied by the filter coefficient as a power of 2 as in the division unit 5 of the filter device 100 according to the first embodiment.
- the calculation error at the time of parallel addition is only the error when dividing the number of "1" by the SN length, and the calculation error can be reduced. Further, especially in the probabilistic expression, since it is necessary to double the SN length in order to double the resolution, the circuit scale is reduced by reducing the SN length, and the calculation error is the same as in the first embodiment. It is possible to reduce the deterioration of the filter characteristics due to the above. Further, since the multiplication units 7-1 to 7-N can be realized by a simple circuit, the overall circuit scale can be reduced.
- Embodiment 3 In the filter device according to the second embodiment, the error in the stochastic expression is between the conversion error derived from the conversion to the stochastic expression and the random number sequence used when converting each input of the operation into the stochastic expression. It is divided into the correlation error derived from the correlation. In this embodiment, a configuration for reducing conversion error and correlation error will be described.
- Conversion error occurs when the random number sequence is biased.
- the random number generation unit 11 converts a value of 0.6 into a stochastic expression. It is desirable to output a uniform value such as 6 times for a value of 0.6 or less and 4 times for a number larger than 0.6.
- the conversion error becomes large because the output distribution is biased.
- the calculation error is derived from the correlation between the random number series used when converting each input to the calculation unit into a stochastic expression.
- the calculation error at the time of parallel addition is small, and therefore, a method for reducing the calculation error in the multiplication unit will be described below.
- each delay signal input to each multiplication unit 7-0 to 7-N and each filter coefficient are converted into probabilistic representations of random number sequences.
- a random number sequence generated by the random number generation unit 11 of the expression conversion unit 6 and a random number sequence for converting the filter coefficient into a stochastic expression are used as a super-uniform distribution sequence such as a Sobol sequence. It is effective to realize it.
- the Sobol columns may be realized by reading them out from the conversion table in order, for example, a method as disclosed in the literature "Notes on generating Sobol'sequences", Stephen Joe and Frances Y.Kuo, August 2008. It may be realized by.
- the ultra-lowdiscrepancy sequence can be selected from a plurality of series, two series with a small multiplication error are selected, one is used for the representation conversion of the delay signal, and the other is used for the representation conversion of the filter coefficient. Is effective.
- the configuration shown in the above-described embodiment shows an example of the content of the present invention, can be combined with another known technique, and is one of the configurations without departing from the gist of the present invention. It is also possible to omit or change the part.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Complex Calculations (AREA)
- General Physics & Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
Abstract
Description
図1は、本発明の実施の形態1にかかるフィルタ装置の回路構成の一例を示す図である。実施の形態1にかかるフィルタ装置100は、ディジタルフィルタであり、具体的にはFIRフィルタである。フィルタ装置100は、専用のハードウェアで実現される場合、例えば、単一回路、複合回路、プログラム化したプロセッサ、並列プログラム化したプロセッサ、ASIC(Application Specific Integrated Circuit)、FPGA(Field Programmable Gate Array)、またはこれらを組み合わせたものである。
次に、実施の形態2にかかるフィルタ装置、具体的には、確率的演算を用いて実現するフィルタ装置について説明する。
実施の形態2の構成のフィルタ装置において、確率的表現における誤差は、確率的表現への変換に由来する変換誤差と、演算の各入力を確率的表現へ変換する際に用いた乱数系列間の相関に由来する相関誤差とに分けられる。本実施の形態では、変換誤差および相関誤差を低減する構成について記載する。
Claims (9)
- 信号が入力されると遅延を与えて遅延信号として出力する、直列に接続された複数の遅延部と、
予め定められた値および倍率調整値に基づき生成されたフィルタ係数を、前記遅延信号に乗算する複数の乗算部と、
前記予め定められた値に前記倍率調整値を乗算して得られる乗算結果が前記フィルタ係数の表現範囲の最大値を超える場合に、前記最大値を超える前記乗算結果を前記最大値で除算した商を係数調整値として出力する係数調整部と、
前記複数の乗算部が出力するフィルタ係数乗算後の信号と、前記係数調整値を用いて対応する遅延信号を調整して得られる調整後信号とを加算して出力する信号変換部と、
前記信号変換部が出力する信号を前記倍率調整値で除算して出力信号を生成する除算部と、
を備えることを特徴とするフィルタ装置。 - 前記乗算結果を前記最大値で除算した剰余を前記フィルタ係数とする、
ことを特徴とする請求項1に記載のフィルタ装置。 - 前記信号変換部は、前記係数調整値を対応する前記遅延信号に乗算して前記調整後信号を得る、
ことを特徴とする請求項1または2に記載のフィルタ装置。 - 前記係数調整部は、前記係数調整値に対応する前記遅延信号を前記係数調整値と同じ数だけ出力し、
前記信号変換部は、前記係数調整部が出力する、前記係数調整値に対応する前記遅延信号のそれぞれを加算して前記調整後信号を得る、
ことを特徴とする請求項1または2に記載のフィルタ装置。 - 前記倍率調整値を2の累乗数とし、
前記除算部が行う除算処理をビットシフトにより実現する、
ことを特徴とする請求項1から4のいずれか一つに記載のフィルタ装置。 - 入力信号を確率的表現の信号に変換して複数の前記遅延部の中の先頭の遅延部に入力する表現変換部、
を備え、
前記フィルタ係数を確率的表現のフィルタ係数とし、
前記係数調整部および前記信号変換部は、確率的表現の信号を対象として演算処理を行う、
ことを特徴とする請求項1から5のいずれか一つに記載のフィルタ装置。 - 前記フィルタ係数を前記確率的表現のフィルタ係数に変換する係数変換部、
を備えることを特徴とする請求項6に記載のフィルタ装置。 - 前記表現変換部は、超一様分布列を用いて入力信号を確率的表現の信号に変換する、
ことを特徴とする請求項6または7に記載のフィルタ装置。 - 前記フィルタ係数を、超一様分布列を用いて前記確率的表現のフィルタ係数に変換する係数変換部、
を備えることを特徴とする請求項6から8のいずれか一つに記載のフィルタ装置。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CA3158143A CA3158143A1 (en) | 2019-12-16 | 2019-12-16 | Filter device |
EP19956841.1A EP4044433B1 (en) | 2019-12-16 | 2019-12-16 | Finite impulse response filter (fir) |
PCT/JP2019/049224 WO2021124413A1 (ja) | 2019-12-16 | 2019-12-16 | フィルタ装置 |
JP2021558571A JP7076655B2 (ja) | 2019-12-16 | 2019-12-16 | フィルタ装置 |
US17/717,483 US11894822B2 (en) | 2019-12-16 | 2022-04-11 | Filter device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2019/049224 WO2021124413A1 (ja) | 2019-12-16 | 2019-12-16 | フィルタ装置 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/717,483 Continuation US11894822B2 (en) | 2019-12-16 | 2022-04-11 | Filter device |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021124413A1 true WO2021124413A1 (ja) | 2021-06-24 |
Family
ID=76477301
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2019/049224 WO2021124413A1 (ja) | 2019-12-16 | 2019-12-16 | フィルタ装置 |
Country Status (5)
Country | Link |
---|---|
US (1) | US11894822B2 (ja) |
EP (1) | EP4044433B1 (ja) |
JP (1) | JP7076655B2 (ja) |
CA (1) | CA3158143A1 (ja) |
WO (1) | WO2021124413A1 (ja) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0346813A (ja) | 1989-07-14 | 1991-02-28 | Sony Tektronix Corp | デジタル・フィルタ回路 |
JPH0360509A (ja) * | 1989-07-29 | 1991-03-15 | Sony Corp | ディジタル信号処理回路 |
JPH09199926A (ja) * | 1996-01-17 | 1997-07-31 | Atr Hikari Denpa Tsushin Kenkiyushiyo:Kk | 適応ディジタルビームフォーミング装置 |
JP2006093884A (ja) * | 2004-09-21 | 2006-04-06 | Oki Electric Ind Co Ltd | フィルタ装置 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0720047B2 (ja) * | 1985-11-01 | 1995-03-06 | ソニー株式会社 | デイジタルフイルタ |
US6505221B1 (en) * | 1999-09-20 | 2003-01-07 | Koninklijke Philips Electronics N.V. | FIR filter utilizing programmable shifter |
JP2009129302A (ja) | 2007-11-27 | 2009-06-11 | Panasonic Corp | 確率的演算法によるニューラルネットワーク素子の学習法 |
-
2019
- 2019-12-16 WO PCT/JP2019/049224 patent/WO2021124413A1/ja unknown
- 2019-12-16 EP EP19956841.1A patent/EP4044433B1/en active Active
- 2019-12-16 CA CA3158143A patent/CA3158143A1/en not_active Abandoned
- 2019-12-16 JP JP2021558571A patent/JP7076655B2/ja active Active
-
2022
- 2022-04-11 US US17/717,483 patent/US11894822B2/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH0346813A (ja) | 1989-07-14 | 1991-02-28 | Sony Tektronix Corp | デジタル・フィルタ回路 |
JPH0360509A (ja) * | 1989-07-29 | 1991-03-15 | Sony Corp | ディジタル信号処理回路 |
JPH09199926A (ja) * | 1996-01-17 | 1997-07-31 | Atr Hikari Denpa Tsushin Kenkiyushiyo:Kk | 適応ディジタルビームフォーミング装置 |
JP2006093884A (ja) * | 2004-09-21 | 2006-04-06 | Oki Electric Ind Co Ltd | フィルタ装置 |
Also Published As
Publication number | Publication date |
---|---|
JP7076655B2 (ja) | 2022-05-27 |
US20220239283A1 (en) | 2022-07-28 |
EP4044433A1 (en) | 2022-08-17 |
EP4044433A4 (en) | 2022-10-26 |
CA3158143A1 (en) | 2021-06-24 |
EP4044433B1 (en) | 2024-06-26 |
JPWO2021124413A1 (ja) | 2021-06-24 |
US11894822B2 (en) | 2024-02-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP3245171B2 (ja) | 高速アダマール変換を行う方法およびその装置 | |
US11018689B2 (en) | Parallel computing using stochastic circuits and deterministic shuffling networks | |
TW200418264A (en) | Reconfigurable fir filter | |
Premkumar | A formal framework for conversion from binary to residue numbers | |
JP7076655B2 (ja) | フィルタ装置 | |
WO2016181992A1 (ja) | 復号装置、復号方法、およびプログラム | |
US11604852B2 (en) | Signal processing apparatus, method, program, and recording medium | |
WO2023134507A1 (zh) | 随机计算方法、电路、芯片及设备 | |
CN109379191B (zh) | 一种基于椭圆曲线基点的点乘运算电路和方法 | |
US20200104131A1 (en) | Method for Operating a Digital Computer to Reduce the Computational Complexity Associated with Dot Products between Large Vectors | |
Wu et al. | Improving common subexpression elimination algorithm with a new gate-level delay computing method | |
US12019700B2 (en) | Signal processing apparatus, method, program, and recording medium | |
JPWO2007102611A1 (ja) | 補間関数生成回路 | |
KR20160057590A (ko) | 공통 부분식 소거 방법 및 이를 이용한 필터 | |
JP4738408B2 (ja) | 行列演算装置 | |
CN106656189B (zh) | 一种多级折叠内插型模数转换器及其译码方法 | |
Pushpa et al. | Area optimized implementation of Galois field Fourier transform | |
JP2000132539A (ja) | 演算装置 | |
JP4405452B2 (ja) | 逆変換回路 | |
GB2233849A (en) | Digital filters | |
TWI780796B (zh) | 轉換數位影像資料 | |
WO2022168161A1 (ja) | 演算回路、通信装置、制御回路、記憶媒体および演算方法 | |
JP2001177378A (ja) | Firデジタルフィルタ | |
JP2008158855A (ja) | 相関演算器及び相関演算方法 | |
KR100395511B1 (ko) | 유한체 상에서의 병렬 입출력 승산기의 설계 방법 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19956841 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2021558571 Country of ref document: JP Kind code of ref document: A |
|
ENP | Entry into the national phase |
Ref document number: 3158143 Country of ref document: CA |
|
ENP | Entry into the national phase |
Ref document number: 2019956841 Country of ref document: EP Effective date: 20220513 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |