公开(公告)号：US20060058983A1

公开(公告)日：2006-03-16

申请号：US10539609

申请日：2004-09-01

申请人： Shoko Araki ,  Hiroshi Sawada ,  Shoji Makino ,  Ryo Mukai

发明人： Shoko Araki ,  Hiroshi Sawada ,  Shoji Makino ,  Ryo Mukai

IPC分类号： G06F15/00 ,  H03F1/26 ,  H04B15/00

CPC分类号： G10L21/0272 ,  G06K9/6245

摘要： This invention achieves high-quality separation of mixed signals in situations where the relationship between the number of signal sources N and the number of sensors M is such that N>M. First, the values of the observed signal observed by M sensors are transformed into frequency domain values, and these frequency domain values are used to calculate the relative values of the observed values between the sensors at each frequency. These relative values are clustered into N clusters, and the representative value of each cluster is calculated. Then, using these representative values, a mask is produced to extract the values of the signals emitted by V (1≦V≦M) signal sources from the frequency-domain signal values, and this mask is used to extract the signal values emitted by V signal sources from these frequency-domain signal values. After that, if V=1 then the limited signal is output directly as a separated signal, while if V≧2 then the separated values are obtained from this limited signal by subjecting it to separation techniques such as ICA.

摘要翻译： 本发明在信号源数目N与传感器数量M之间的关系为N> M的情况下实现混合信号的高质量分离。 首先，由M个传感器观察到的观测信号的值被变换为频域值，并且这些频域值用于计算每个频率下传感器之间的观测值的相对值。 将这些相对值聚类为N个簇，并计算每个簇的代表值。 然后，使用这些代表值，产生掩模以从频域信号值中提取由V（1 <= V <= M）个信号源发射的信号的值，并且该掩码用于提取信号值 由V信号源从这些频域信号值发射。 之后，如果V = 1，则限制信号直接作为分离信号输出，而如果V> = 2，则通过对分离技术（如ICA）进行分离，从该有限信号获得分离值。

公开(公告)号：US20050203981A1

公开(公告)日：2005-09-15

申请号：US10508708

申请日：2004-03-03

申请人： Hiroshi Sawada ,  Ryo Mukai ,  Shoko Araki ,  Shoji Makino

发明人： Hiroshi Sawada ,  Ryo Mukai ,  Shoko Araki ,  Shoji Makino

IPC分类号： G01S3/46 ,  G01S3/74 ,  G01S3/80 ,  G01S3/808 ,  G06F17/10

CPC分类号： G01S3/46 ,  G01S3/74 ,  G01S3/8006 ,  G01S3/808

摘要： Observed signals x1(t) to xj(t) from a plurality of sensors disposed in two dimensions are subject to a short-time Fourier transform, from which signals X1(ω1) to X1(ωN), . . . XJ(ω1) to XJ(ωN) are derived. Using the independent component analysis process, separation matrices W(ω1) to W(ωN) are produced. Their inverse matrices H(ω1) to H(ωN) are calculated, and for each ωn (n=1, . . . , N), using a pair of elements from each column of H(ωn), Hji(ωn) and Hj′i(ωn), an angle {circumflex over (θ)}i,jj′(ωn)=cos−1(arg(|Hji(ωn)/Hj′i(ωn))/(ωnc−1∥dj−dj′∥)) is calculated where arg(α) represents an argument of α, c the propagation velocity of a signal, and ∥dj−dj′∥ represents a spacing between sensors i and j′. Columns are permuted so that {circumflex over (θ)}i,jj′(ωn) obtained from each column of H(ω1) to H(ωN) assume an ascending order. For columns which cannot be permuted, the equation ∥qi−dj′∥/∥qi−dj∥=|Hji(ωn)/Hj′i(ωn)|=DRi,jj′(ωn) is solved for qi to calculate Ri,jj′(ωn)=∥DRi,jj′(ωn)·(dj−dj′)/(DR2i,jj′(ωn)−1)|. Columns of H(ωn) is permuted so that Ri,jj′(ωn) assume an ascending order. H(ωn) is used to solve the permutation problem of W(ω).

公开(公告)号：US07647209B2

公开(公告)日：2010-01-12

申请号：US10594983

申请日：2006-02-07

申请人： Hiroshi Sawada ,  Shoko Araki ,  Ryo Mukai ,  Shoji Makino

发明人： Hiroshi Sawada ,  Shoko Araki ,  Ryo Mukai ,  Shoji Makino

IPC分类号： G06F3/00

CPC分类号： G10L21/0272 ,  G06K9/6245 ,  G10L15/20 ,  G10L2021/02165

摘要： A frequency domain transforming section transforms mixed signals observed by multiple sensors into mixed signals in the frequency domain, a complex vector generating section generates a complex vector by using the frequency-domain mixed signals, a normalizing section generates a normalized vector excluding frequency dependence of the complex vector, and a clustering section clusters the normalized vectors to generate clusters. Then, a separated signal generating section generates separated signals in the frequency domain by using information about the clusters and a time domain transforming section transforms the separated signals in the frequency domain into separated signals in the time domain.

摘要翻译： 频域变换部将多个传感器观测到的混合信号变换为频域的混合信号，复矢量生成部通过使用频域混合信号生成复矢量，归一化部生成除了频率依赖性的归一化矢量 复数向量，聚类部分聚类归一化向量以生成聚类。 然后，分离信号生成部通过使用关于簇的信息来生成频域的分离信号，并且时域变换部将频域中的分离信号变换为时域中的分离信号。

公开(公告)号：US07039546B2

公开(公告)日：2006-05-02

申请号：US10508708

申请日：2004-03-03

申请人： Hiroshi Sawada ,  Ryo Mukai ,  Shoko Araki ,  Shoji Makino

发明人： Hiroshi Sawada ,  Ryo Mukai ,  Shoko Araki ,  Shoji Makino

IPC分类号： G01C9/00

CPC分类号： G01S3/46 ,  G01S3/74 ,  G01S3/8006 ,  G01S3/808

摘要： Signals from a plurality of sources are observed by a plurality of sensors disposed in two dimensions, and the observed signals are subject to a short-time Fourier transform, from which frequency domain signals are derived. Using the independent component analysis process on the frequency domain signals, separation matrices are produced, and an inverse matrix of each separation matrix is calculated. The direction of each source is calculated based on a ratio of a pair of elements in each column of the inverse matrix.

摘要翻译： 来自多个源的信号由设置在二维上的多个传感器观察，并且所观测的信号经受​​短时傅里叶变换，从其导出频域信号。 对频域信号采用独立分量分析过程，产生分离矩阵，并计算每个分离矩阵的逆矩阵。 每个源的方向基于逆矩阵的每列中的一对元素的比率来计算。

公开(公告)号：US07496482B2

公开(公告)日：2009-02-24

申请号：US10539609

申请日：2004-09-01

申请人： Shoko Araki ,  Hiroshi Sawada ,  Shoji Makino ,  Ryo Mukai

发明人： Shoko Araki ,  Hiroshi Sawada ,  Shoji Makino ,  Ryo Mukai

IPC分类号： G06F15/00

CPC分类号： G10L21/0272 ,  G06K9/6245

摘要： A method and a device for signal separation. First, values of signals observed by M sensors are transformed into frequency domain values, and these frequency domain values are used to calculate relative values of the observed values between the sensors at each frequency. These relative values are clustered into N clusters, and the representative value of each cluster is calculated. Then, using these representative values, a mask is produced to extract the values of the signals emitted by V (1≦V≦M) signal sources from the frequency-domain signal values, and this mask is applied to the frequency-domain signal values. After that, if V=1 then the limited signal is output directly as a separated signal, while if V≧2 then the separated values are obtained by separating this limited signal with separation techniques such as ICA.

摘要翻译： 一种用于信号分离的方法和装置。 首先，将M个传感器观察到的信号值变换为频域值，并使用这些频域值来计算每个频率下传感器之间的观测值的相对值。 将这些相对值聚类为N个簇，并计算每个簇的代表值。 然后，使用这些代表值，产生掩模，以从频域信号值中提取由V（1 <= V <= M）个信号源发射的信号的值，并将该掩码应用于频域 信号值。 之后，如果V = 1，则限制信号作为分离信号直接输出，而如果V> = 2，则通过用ICA等分离技术分离该限制信号来获得分离的值。

公开(公告)号：US20080215651A1

公开(公告)日：2008-09-04

申请号：US10594983

申请日：2006-02-07

申请人： Hiroshi Sawada ,  Shoko Araki ,  Ryo Mukai ,  Shoji Makino

发明人： Hiroshi Sawada ,  Shoko Araki ,  Ryo Mukai ,  Shoji Makino

IPC分类号： G06F17/14

CPC分类号： G10L21/0272 ,  G06K9/6245 ,  G10L15/20 ,  G10L2021/02165

摘要： A frequency domain transforming section 2 transforms mixed signals observed by multiple sensors into mixed signals in the frequency domain, a complex vector generating section 3 generates a complex vector by using the frequency-domain mixed signals, a normalizing section 4 generates a normalized vector excluding frequency dependence of the complex vector, and a clustering section 5 clusters the normalized vectors to generate clusters. Then, a separated signal generating section 6 generates separated signals in the frequency domain by using information about the clusters and a time domain transforming section 7 transforms the separated signals in the frequency domain into separated signals in the time domain.

摘要翻译： 频域变换部2将由多个传感器观测到的混合信号变换为频域的混合信号，复矢量生成部3通过使用频域混合信号生成复矢量，归一化部4生成除频率以外的归一化矢量 复合向量的依赖性，并且聚类部分5聚集归一化向量以生成聚类。 然后，分离信号生成部6通过使用关于簇的信息来生成频域的分离信号，时域变换部7将频域中的分离信号变换为时域中的分离信号。