WO2015165118A1 - 一种压缩感知方法及装置 - Google Patents

一种压缩感知方法及装置 Download PDF

Info

Publication number
WO2015165118A1
WO2015165118A1 PCT/CN2014/076688 CN2014076688W WO2015165118A1 WO 2015165118 A1 WO2015165118 A1 WO 2015165118A1 CN 2014076688 W CN2014076688 W CN 2014076688W WO 2015165118 A1 WO2015165118 A1 WO 2015165118A1
Authority
WO
WIPO (PCT)
Prior art keywords
receiver
sparse
support set
receivers
shared
Prior art date
Application number
PCT/CN2014/076688
Other languages
English (en)
French (fr)
Inventor
饶雄斌
刘坚能
孔翔鸣
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP14891091.2A priority Critical patent/EP3133850A4/en
Priority to PCT/CN2014/076688 priority patent/WO2015165118A1/zh
Priority to CN201480078643.6A priority patent/CN106256141A/zh
Publication of WO2015165118A1 publication Critical patent/WO2015165118A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
    • H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
    • H03M7/3059—Digital compression and data reduction techniques where the original information is represented by a subset or similar information, e.g. lossy compression
    • H03M7/3062—Compressive sampling or sensing

Definitions

  • the present invention relates to the field of wireless communications, and in particular, to a compression sensing method and apparatus. Background technique
  • CSIT Channel State Information at The Transmitters
  • FDD Frequency Division Duplex
  • BS Base Station
  • the length of the transmission training sequence increases linearly with the number of antennas at the BS. This results in a very large training sequence overhead at the BS, which will greatly reduce the spatial multiplexing gain and diversity gain that large-scale antennas can bring.
  • the target signal generally exhibits sparsity characteristics.
  • the frequency band utilization of a signal may be sparse due to the burst transmission characteristics of some applications (the user's voice service, the user has a large amount of idle time).
  • the target signals often exhibit strong correlation characteristics within and between the target signals, that is, the target signals tend to exhibit joint sparsity.
  • MIM 0 M u 11 ip 1 e Input Multiple Output
  • the active users that can be observed by the antennas of the same site are consistent, and different sites may Some active users will be shared.
  • CS Comper es ed sen Sensing, Compressive Sampling, Compressive Sampling
  • the theory states that when the target signal exhibits sparsity characteristics, it only needs to have a small amount of sampled values or observation data to effectively recover the target signal.
  • Embodiments of the present invention provide a compressed sensing method and apparatus, which solves the problem of high cost of capturing high-dimensional target signals in existing wireless communication systems.
  • an embodiment of the present invention provides a data processing center, including:
  • a receiving unit configured to receive observation data sent by at least two receivers, where the observation data is a sparse signal of the at least two receivers to the at least two receivers according to an observation matrix of the at least two receivers Sampling data;
  • a determining unit configured to determine, according to the observation data and the joint sparsity characteristic received by the receiving unit, a shared support set, where the sparse signal of each receiver includes at least one support set, and the at least two There is at least one shared support set between the sparse signals of the receiver, wherein the shared support set is a support set shared between the sparse signals of the at least two receivers, and the support set is a sparse signal of each receiver a set of indices of non-zero rows in the matrix, and a set of supports for determining respective sparse signals based on the shared support set and the joint sparse characteristics, and for determining a set of sparse signals according to the shared support set and the respective sparse signals Support sets to determine individual sparse signals.
  • the joint sparsity characteristic includes a sparse signal of each receiver includes at least one support set, and at least one share exists between sparse signals of the at least two receivers a support set, where the shared support set is a common support set between sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including :
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is a sparse letter of the at least two receivers a set of indices of non-zero rows shared in the two-dimensional matrix;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the at least two receivers obtain the observation data by sampling the sparse signals of the at least two receivers according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set
  • ⁇ the number of observed signals for the first receiver
  • Ni C NxT is a matrix of N rows and columns.
  • an embodiment of the present invention provides a receiver, including:
  • an acquiring unit configured to acquire observation data of the first receiver, where the observation data is that the first receiver samples data of the sparse signal of the first receiver according to an observation matrix of the first receiver;
  • a determining unit configured to determine, according to the joint sparsity characteristic and the observation data acquired by the acquiring unit, the shared support set, where the joint sparsity characteristic is that the sparse signal of each receiver in the at least two receivers includes at least one support set, And at least one shared support set exists between the sparse signals of the at least two receivers, where the shared support set is a support set shared between sparse signals of the at least two receivers, where the support set is An index set of non-zero rows in a matrix of sparse signals of respective receivers, a backhaul connection is supported between the first receiver and the second receiver, and for the shared support set and the joint sparse characteristics Determining a support set of the sparse signal of the first receiver, and determining a sparse signal of the first receiver according to the support set of the shared support set and the sparse signal of the first receiver.
  • the joint sparsity characteristic is that the sparse signal of each receiver in the at least two receivers includes at least one support set, and There is at least one shared support set between sparse signals of at least two receivers, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is received by each The index set of non-zero rows in the matrix of the sparse signal of the machine, including:
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix a collection of non-zero rows of indexes;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the determining unit is configured to determine, according to the joint sparsity characteristic and the observation data, An estimated support set of the sparse signal of the first receiver, and specifically for exchanging the estimated support set with an estimated support set of the second receiver, and determining the shared support set according to the joint sparsity characteristic
  • the two receivers are receivers of the at least two receivers other than the first receiver that support a backhaul connection with the first receiver.
  • the first receiver obtains observation data by sampling the sparse signal of the first receiver according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observation data, ⁇ is the observation matrix of the first receiver
  • the observation noise of the second receiver e
  • C MxiV is a matrix of N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set
  • ⁇ is the number of observation signals of the first receiver
  • ⁇ ( ⁇ is the matrix of the ⁇ row
  • NC NxT is the matrix of N rows ⁇ column.
  • an embodiment of the present invention provides a data processing center, including: a receiver, configured to receive observation data sent by at least two receivers, where the observation data is that the at least two receivers are according to the at least two Data obtained by sampling the sparse signals of the at least two receivers by the observation matrix of the receivers;
  • a processor configured to determine, according to the observation data and the joint sparsity characteristic received by the receiver, a shared support set, where the sparse signal of each receiver includes at least one support set, and the at least two There is at least one shared support set between the sparse signals of the receiver, wherein the shared support set is a support set shared between the sparse signals of the at least two receivers, and the support set is a sparse signal of each receiver a set of indices of non-zero rows in the matrix, and a set of supports for determining respective sparse signals based on the shared support set and the joint sparse characteristics, and for determining a set of sparse signals according to the shared support set and the respective sparse signals Support sets to determine individual sparse signals.
  • the joint sparsity characteristic includes a sparse signal of each receiver including at least one support set, and at least one sharing between the sparse signals of the at least two receivers a support set, where the shared support set is a common support set between sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including :
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix a collection of non-zero rows of indexes;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the at least two receivers obtain the observation data by sampling the sparse signals of the at least two receivers according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observation data ⁇ , ⁇ is the observation matrix of the first receiver
  • the observation noise of the second receiver e
  • C MxiV is a matrix of N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set
  • N columns The matrix
  • ⁇ is the number of observation signals of the first receiver
  • Ni C NxT is the matrix of N rows and columns.
  • the data processing center includes the central processor.
  • an embodiment of the present invention provides a receiver, including:
  • a processor configured to acquire observation data of the first receiver, where the observation data is that the first receiver samples data of the sparse signal of the first receiver according to an observation matrix of the first receiver, and And determining, according to the joint sparse characteristic and the acquired observation data, the shared support set, wherein the joint sparsity characteristic is that the sparse signal of each of the at least two receivers includes at least one support set, and the at least two receiving There is at least one shared support set between the sparse signals of the machine, wherein the shared support set is a support set shared between the sparse signals of the at least two receivers, and the support set is a sparse signal of each receiver An index set of non-zero rows in the matrix, the backhaul connection is supported between the first receiver and the second receiver, and the first receiving is determined according to the shared support set and the joint sparsity characteristic a support set of sparse signals of the machine, and a support set for sparse signals according to the shared support set and the first receiver, Sparse signal of said first receiver.
  • the joint sparsity characteristic is that a sparse signal of each receiver in the at least two receivers includes at least one support set, and the sparse signals of the at least two receivers There is at least one shared support set between, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is non-zero in a matrix of sparse signals of each receiver
  • the index set of rows including:
  • Each receiver includes at least one antenna, and each antenna of at least two receivers is thin
  • the signal is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix a collection of non-zero rows of indexes;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the processor is configured to determine, according to the joint sparsity characteristic and the observation data, An estimated support set of the sparse signal of the first receiver, and specifically for exchanging the estimated support set with an estimated support set of the second receiver, and determining the shared support set according to the joint sparsity characteristic,
  • the two receivers are receivers of the at least two receivers other than the first receiver that support a backhaul connection with the first receiver.
  • the first receiver obtains observation data by sampling the sparse signal of the first receiver according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set, and is N columns
  • the matrix, ⁇ is the number of observation signals of the first receiver, and the matrix of M columns
  • Ni C NxT is the matrix of N rows and columns.
  • the receiver includes a base station or base station controller.
  • an embodiment of the present invention provides a compressed sensing method, which includes: Receiving observation data transmitted by at least two receivers, the observation data being obtained by sampling the sparse signals of the at least two receivers according to an observation matrix of the at least two receivers by the at least two receivers data;
  • the shared support set Determining, according to the received observation data and the joint sparse characteristic, the shared support set, wherein the sparse signal of each receiver includes at least one support set, and at least one of the sparse signals of the at least two receivers exists at least a shared support set, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver ;
  • Each of the sparse signals is determined based on the shared support set and the support set of the respective sparse signals.
  • the joint sparsity characteristic includes a sparse signal of each receiver includes at least one support set, and at least one share exists between sparse signals of the at least two receivers a support set, where the shared support set is a common support set between sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including :
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix a collection of non-zero rows of indexes;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observation data, ⁇ , ⁇ is the observation matrix of the first receiver
  • the observation noise of the second receiver e C MxiV is a matrix of N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set, and is N columns
  • ⁇ is the number of observation signals of the first receiver
  • Ni C NxT is the matrix of N rows and columns.
  • an embodiment of the present invention provides a compressed sensing method, including:
  • observation data of the first receiver is obtained by the first receiver sampling the sparse signal of the first receiver according to an observation matrix of the first receiver;
  • the joint sparsity characteristic is that the sparse signal of each of the at least two receivers includes at least one support set, and the at least two receivers are sparse
  • the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is a matrix of sparse signals of each receiver a zero row index set, and a backhaul connection is supported between the first receiver and the second receiver;
  • a sparse signal of the first receiver is determined based on a support set of the shared support set and the sparse signal of the first receiver.
  • the joint sparsity characteristic is that a sparse signal of each receiver of the at least two receivers includes at least one support set, and the sparse signals of the at least two receivers There is at least one shared support set between, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is non-zero in a matrix of sparse signals of each receiver
  • the index set of rows including:
  • Each receiver includes at least one antenna, and each antenna of at least two receivers is thin
  • the signal is composed of antenna numbers, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows;
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix a collection of non-zero rows of indexes;
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the determining, according to the joint sparsity characteristic and the observation data, the shared support set specifically:
  • the second receiver is the at least two receivers A receiver other than a receiver that supports a backhaul connection with the first receiver.
  • the first receiver obtains observation data by sampling the sparse signal of the first receiver according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observation data, ⁇ is the observation matrix of the first receiver
  • the observation noise of the second receiver e
  • C MxiV is a matrix of N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set, and is N columns
  • ⁇ is the number of observed signals of the first receiver
  • ⁇ ( ⁇ is the matrix of the ⁇ row
  • Ni C NxT is the matrix of the N rows ⁇ column.
  • Embodiments of the present invention provide a compression sensing method. After receiving data of observations sent by at least one receiver, the data processing center determines a shared support set according to all observation data and joint sparsity characteristics, and determines shared support. After the collection, the data office The management center determines the support set of each sparse signal according to the shared support set and the joint sparse characteristic. Finally, the data processing center determines each sparse signal according to the shared support set and the support set of each sparse signal, wherein the observation data is determined by each receiver according to each receiver.
  • the observation matrix of the receiver samples the sparse signal of the receiver, and the joint sparsity characteristic has a shared support set between all sparse signals, and at least one support set exists in each sparse signal, and the support set corresponds to each sparse signal.
  • FIG. 1 is a schematic structural diagram 1 of a data processing center according to an embodiment of the present invention.
  • FIG. 2 is a schematic structural diagram 1 of a receiver according to an embodiment of the present invention.
  • FIG. 3 is a second schematic structural diagram of a data processing center according to an embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram 2 of a receiver according to an embodiment of the present invention.
  • FIG. 5 is a first application diagram of a compressed sensing method according to an embodiment of the present invention.
  • FIG. 6 is a schematic flowchart 1 of a compressed sensing method according to an embodiment of the present invention.
  • FIG. 7 is a schematic diagram 1 of a joint sparse characteristic structure according to an embodiment of the present invention.
  • FIG. 8 is a schematic diagram 2 of a joint sparse characteristic structure according to an embodiment of the present invention.
  • FIG. 9 is a second application scenario of a compressed sensing method according to an embodiment of the present invention.
  • FIG. 10 is a schematic flowchart 2 of a compressed sensing method according to an embodiment of the present invention.
  • GSM global system for mobile communications
  • CDMA code division multiple access
  • TDMA Time division multiple access
  • WCDMA Time division multiple access
  • WCDMA wideband code division multiple access
  • FDMA frequency division multiple access
  • OFDMA orthogonal frequency division multiple access
  • SC-FDMA single carrier FDMA
  • GPRS general packet radio service
  • genera 1 packet radio service system
  • LTE long term evolution
  • the technical solution of the present invention considers a framework for signal acquisition in a wireless communication system, and the technical framework can cover many specific application problems.
  • the technical framework can be applied to a distributed RF (Radio Frequency) terminal to simulate information. Conversion issues, CSIT acquisition issues in multi-user large-scale antenna networks, etc.
  • the embodiment of the present invention provides a data processing center 1, as shown in FIG. 1, comprising: a receiving unit 10, configured to receive observation data sent by at least two receivers, where the observation data is based on the at least two receivers
  • the observation matrix of the at least two receivers samples the resulting data of the sparse signals of the at least two receivers.
  • a determining unit 11 configured to determine, according to the observation data and the joint sparsity characteristic received by the receiving unit 10 , a shared support set, where the sparse signal of each receiver includes at least one support set, and the at least one support set There is at least one shared support set between the sparse signals of the two receivers, wherein the shared support set is a support set shared between the sparse signals of the at least two receivers, and the support set is for each receiver An index set of non-zero rows in a matrix of sparse signals, and a set of supports for determining respective sparse signals according to the shared support set and the joint sparse characteristic, and for determining a sparse signal according to the shared support set and the respective sparse A set of support for the signal to determine each sparse signal.
  • the joint sparse characteristic includes a sparse signal of each receiver including at least one support set, and at least one shared support set exists between sparse signals of the at least two receivers, where the shared support set is a common support set between the sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including:
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna at the same sparse domain value is a two-dimensional matrix of rows.
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix A collection of non-zero rows of indexes.
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the at least two receivers obtain the observation data by sampling the sparse signals of the at least two receivers according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set, and is N columns
  • the matrix, ⁇ is the number of observation signals of the first receiver, and the matrix of M columns
  • Ni C NxT is the matrix of N rows and columns.
  • the embodiment of the invention provides a data processing center, which mainly comprises a receiving unit and a processing unit. After receiving the observation data sent by the at least one receiver, the data processing center determines the shared support set according to all the observation data and the joint sparsity characteristic, and after determining the shared support set, the data processing center according to the shared support set and the joint The sparse characteristic determines the support set of each sparse signal.
  • the data processing center determines each sparse signal according to the shared support set and the support set of each sparse signal, wherein the observed data is
  • Each receiver samples the sparse signal of the receiver according to the observation matrix of the receiver, and the joint sparsity characteristic has a shared support set between all sparse signals, and at least one support set exists in each sparse signal, and the support set is An index set of non-zero rows in a matrix corresponding to each sparse signal, and a row in the non-zero behavior matrix where the row vector is not zero.
  • the invention utilizes the joint sparsity inside and between the sparse signals, solves the problem of high cost of capturing sparse signals in the existing wireless communication system, and reduces the cost of capturing sparse signals.
  • An embodiment of the present invention provides a receiver 1, as shown in FIG. 2, including:
  • the acquiring unit 10 is configured to acquire observation data of the first receiver, where the observation data is that the first receiver samples the sparse signal of the first receiver according to the observation matrix of the first receiver to obtain data. .
  • a determining unit 1 1 configured to determine, according to the joint sparsity characteristic and the observation data acquired by the acquiring unit 10, a shared support set, where the joint sparsity characteristic is that at least two receivers of each receiver have at least a sparse signal a support set, and at least one shared support set exists between sparse signals of the at least two receivers, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, where The support set is a set of indices of non-zero rows in a matrix of sparse signals of respective receivers, a backhaul connection is supported between the first receiver and the second receiver, and is used according to the shared support set and Determining a support set of the sparse signal of the first receiver, and determining a support set of the sparse signal according to the shared support set and the first receiver, determining the first receiver Sparse signal.
  • the joint sparsity characteristic is that the sparse signal of each receiver of the at least two receivers includes at least one support set, and at least one shared support set exists between the sparse signals of the at least two receivers, where
  • the shared support set is a support set shared between sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including:
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and each antenna has a signal at the same sparse domain value.
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix A collection of non-zero rows of indexes.
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the determining unit 1 1 is specifically configured to determine, according to the joint sparsity characteristic and the observation data, an estimated support set of the sparse signal of the first receiver, and specifically, to use the estimated support set Exchanging with an estimated support set of the second receiver, and determining the shared support set according to the joint sparse characteristic, the second receiver being other than the first receiver of the at least two receivers A receiver that supports a backhaul connection with the first receiver.
  • the first receiver samples the sparse signal of the first receiver according to the following formula to obtain observation data:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set
  • ⁇ the number of observed signals for the first receiver
  • Ni C NxT is a matrix of N rows and columns.
  • the embodiment of the invention provides a receiver, which mainly comprises an obtaining unit and a determining unit. After acquiring the observation data, the receiver determines the shared support set according to the joint sparsity characteristic and its observation data, and then determines the support set of the sparse signal according to the shared support set and the joint sparsity characteristic, and finally, according to the shared support set and a sparse signal support set, wherein the sparse signal of the first receiver is determined, wherein the observation data is data obtained by sampling the sparse signal according to the observation matrix of the receiver, and the joint sparsity characteristic is a shared support set between all sparse signals, each There is at least one support set inside the sparse signal, and the support set is The index set of non-zero rows in the matrix corresponding to the sparse signal, the row in the non-zero behavior matrix where the row vector is not zero.
  • the invention utilizes the joint sparsity inside and between the sparse signals, solves the problem of high cost of capturing high-dimensional sparse signals
  • the embodiment of the present invention provides a data processing center.
  • the data processing center may include a receiver 10, a processor 11, a memory 12, and a system bus 13, wherein
  • the receiver 10, the processor 11 and the memory 12 are connected via the system bus 13 and communicate with each other.
  • Processor 11 may be a single core or multi-core central processing unit, or a particular integrated circuit, or one or more integrated circuits that are configured to be inventive embodiments.
  • the memory 12 may be a high speed RAM (random access memory) memory or a non-volatile memory such as at least one disk memory.
  • the memory 12 is used to store execution instructions of the data processing center. Specifically, the software processing program and the software code may be included in the execution instruction of the data processing center.
  • the receiver 10 is configured to receive observation data sent by at least two receivers, where the observation data is that the at least two receivers are based on the observation matrix of the at least two receivers The data obtained by sampling the sparse signal of the receiver.
  • the processor 11 is configured to determine, according to the observation data and the joint sparsity characteristic received by the receiver 10, a shared support set, where the sparse signal of each receiver includes at least one support set, and the at least one There is at least one shared support set between the sparse signals of the two receivers, wherein the shared support set is a support set shared between the sparse signals of the at least two receivers, and the support set is for each receiver An index set of non-zero rows in a matrix of sparse signals, and a set of supports for determining respective sparse signals according to the shared support set and the joint sparse characteristic, and for determining a sparse signal according to the shared support set and the respective sparse A set of support for the signal to determine each sparse signal.
  • the joint sparsity characteristic includes a sparse signal for each receiver One less support set, and at least one shared support set exists between the sparse signals of the at least two receivers, wherein the shared support set is a common support set between sparse signals of the at least two receivers,
  • the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including:
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna at the same sparse domain value is a two-dimensional matrix of rows.
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix A collection of non-zero rows of indexes.
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the at least two receivers obtain the observation data by sampling the sparse signals of the at least two receivers according to the following formula:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C is a complex set
  • ⁇ the number of observed signals for the first receiver
  • Ni C NxT is a matrix of N rows and columns.
  • the data processing center may be a central processing unit, which is not limited in the embodiment of the present invention.
  • Embodiments of the present invention provide a data processing center. After receiving data of observations sent by at least one receiver, the data processing center determines a shared support set according to all observation data and joint sparsity characteristics, and determines a shared support set. Afterwards, the data processing center determines the support set of each sparse signal according to the shared support set and the joint sparse feature. Finally, the data processing center determines according to the shared support set and the support set of each sparse signal.
  • Each of the sparse signals wherein the observation data is data obtained by sampling the sparse signal of the receiver according to the observation matrix of the receiver, and the joint sparsity characteristic is a shared support set between all sparse signals, and each sparse signal exists internally At least one support set, the support set is an index set of non-zero rows in the matrix corresponding to each sparse signal, and the rows in the non-zero behavior matrix are not zero.
  • the invention utilizes the joint sparsity inside and between the sparse signals, solves the problem of high cost of capturing sparse signals in the existing wireless communication system, and reduces the cost of capturing sparse signals.
  • the embodiment of the present invention provides a receiver.
  • the receiver may include a processor 10, a memory 11, and a system bus 12, where
  • Processor 10 may be a single core or multi-core central processing unit, or a particular integrated circuit, or one or more integrated circuits that are configured to invent an embodiment of the invention.
  • the memory 11 may be a high speed RAM (random access memory) memory or a non-volatile memory such as at least one disk memory.
  • RAM random access memory
  • non-volatile memory such as at least one disk memory.
  • the memory 11 is used to store execution instructions of the data processing center. Specifically, the software processing program and the software code may be included in the execution instruction of the data processing center.
  • the processor 10 is configured to acquire observation data of the first receiver, where the observation data is that the first receiver samples the sparse signal of the first receiver according to the observation matrix of the first receiver. Obtaining data, and determining a shared support set according to the joint sparse characteristic and the acquired observation data, wherein the joint sparsity characteristic is that the sparse signal of each of the at least two receivers includes at least one support set, and There is at least one shared support set between sparse signals of at least two receivers, wherein the shared support set is a support set shared between sparse signals of the at least two receivers, and the support set is each receiver An index set of non-zero rows in a matrix of sparse signals, a backhaul connection supported between the first receiver and the second receiver, and for determining a location based on the shared support set and the joint sparsity characteristic a support set of sparse signals of the first receiver, and a branch for sparse signals according to the shared support set and the first receiver Supporting, determining a spars
  • the joint sparsity characteristic is that the sparse signal of each receiver of the at least two receivers includes at least one support set, and at least one shared support set exists between the sparse signals of the at least two receivers, where
  • the shared support set is a support set shared between sparse signals of the at least two receivers, where the support set is an index set of non-zero rows in a matrix of sparse signals of each receiver, including:
  • Each receiver includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and the signal of each antenna at the same sparse domain value is a two-dimensional matrix of rows.
  • the sparse signals of the at least two receivers have at least one of the shared support sets in the two-dimensional matrix, and the shared support set is that the sparse signals of the at least two receivers are shared in the two-dimensional matrix A collection of non-zero rows of indexes.
  • the sparse signals of the respective receivers have at least one of the support sets in the two-dimensional matrix, and the support set is that the sparse signals of the respective receivers are not zero on the same line in the two-dimensional matrix. Index collection.
  • the processor 10 is specifically configured to determine, according to the joint sparsity characteristic and the observation data, an estimated support set of the sparse signal of the first receiver, and specifically, to use the estimated support set Exchanging with an estimated support set of the second receiver, and determining the shared support set according to the joint sparse characteristic, the second receiver being other than the first receiver of the at least two receivers A receiver that supports a backhaul connection with the first receiver.
  • the first receiver samples the sparse signal of the first receiver according to the following formula to obtain observation data:
  • the number of receivers, and ⁇ 2 is the sparse signal of the first receiver
  • the observed data ⁇ , ⁇ is the observation matrix of the first receiver, and is the observed noise of the z'th receiver.
  • e C MxiV is a matrix of M rows and N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows of the two-dimensional matrix
  • C represents a complex set
  • YC TxN is a ⁇ N
  • ⁇ is the number of observation signals of the first receiver
  • ⁇ is the matrix of the queue
  • N ⁇ C NxT is the matrix of the N rows and columns.
  • the receiver may be a base station or a base station controller, which is not limited in the embodiment of the present invention.
  • the embodiment of the invention provides a receiver, after obtaining the observation data, the receiver determines the shared support set according to the joint sparsity characteristic and the observation data, and then determines the support of the sparse signal according to the shared support set and the joint sparsity characteristic. And finally, according to the shared support set and the support set of the sparse signal, the sparse signal is determined, wherein the observation data is obtained by sampling the sparse signal according to the observation matrix of the first receiver, and the joint sparseness is generated] There is a shared support set between all sparse signals. There is at least one support set in each sparse signal.
  • the support set is an index set of non-zero rows in the matrix corresponding to each sparse signal, and the rows in the non-zero behavior matrix are not zero.
  • An embodiment of the present invention provides a compression sensing method, and an application scenario thereof is shown in FIG. 5.
  • the entire wireless communication system includes a data processing center and ( ⁇ 2) receivers, where ⁇ is the first (1 ⁇ ⁇ )
  • the sparse signal observed by the receiver obtains observation data about the sparse signal ⁇ for the first receiver, and ⁇ ⁇ + ⁇ , ⁇ ; is the observation matrix of the first receiver, which is the receiver of the second receiver Observing the noise, the receivers will send the obtained observation data ⁇ ⁇ back to the data processing center.
  • the data processing center centrally recovers the sparse signal.
  • the method includes:
  • the data processing center receives the observation data sent by the at least two receivers, wherein the observation data is data obtained by sampling the sparse signals of the at least two receivers according to the observation matrix of the at least two receivers by the at least two receivers. .
  • the data processing center may be a central processing unit.
  • the application scenario of the embodiment of the present invention is that the data processing center centrally recovers the sparse signal, and therefore, the data processing center first receives the observation data sent by at least two receivers.
  • the target signal is generally Sparse characteristics will occur.
  • the frequency band utilization of the signal may be sparse; or, in a large-scale antenna
  • the CSI Channel State Information
  • the BS Base Station
  • MS Mobile Users
  • the compressed sensing method provided by the embodiment of the present invention solves the problem of high cost of high-dimensional target signal acquisition in a wireless communication system, and is mainly directed to a high-dimensional target signal. Therefore, the target signal in the embodiment of the present invention is a sparse signal.
  • a sparse signal refers to a signal that can be expressed by fewer coefficients while allowing a small loss.
  • the length of the signal is N. If the signal can be represented by a linear combination of a set of bases ⁇ ,..., ⁇ ... ⁇ : ⁇ where ⁇ ⁇ represents the transposition of ⁇ , then:
  • Equation — ⁇ and JC are matrices of N rows and 1 column, and ⁇ is a matrix of N rows and N columns.
  • the signal JC has only «N non-zero coefficients on a certain basis, the signal is sparse and has compressibility, which is called the sparse domain of the signal X.
  • the sparse domain of the sparse signal in the embodiment of the present invention may be a signal frequency band or a channel information, which is not limited in the embodiment of the present invention.
  • each receiver in the embodiment of the present invention includes at least one antenna, and the sparse signal of each antenna of at least two receivers is composed of antenna numbers, and each antenna has a signal on the same sparse domain value.
  • the two-dimensional matrix Each receiver samples the sparse signal of the receiver according to the observation matrix of the receiver using the following formula to obtain its observation data:
  • the observed data, ⁇ is the observation matrix of the first receiver
  • ⁇ ⁇ is the observed noise of the second receiver
  • .eC MxiV is a matrix of N columns
  • N is the number of antennas included in the first receiver
  • M is the number of rows in the two-dimensional matrix (it depends on the characteristics of the sparse signal (eg For example, if the channel time domain information is represented, the channel delay and the accuracy of the measurement determine the size of M), it can be expressed by >> 1 (far greater than 1).
  • N ⁇ l, Y t C TxN ⁇ is a matrix of N columns, ⁇ is the number of observed signals or the number of samples collected (The value of ⁇ is related to sparsity and M, and S is sparsity, then T is generally on the order of S * l og (M/S)), and O ⁇ e C ⁇ M is a matrix of M columns.
  • N t & C NxT is a matrix of N rows of T.
  • the observation matrix used by each receiver may be the same or different, and the embodiment of the present invention is not limited. Specifically, the observation matrix is channel information of each receiver according to its location. Other factors determine.
  • observation matrix of each receiver in the embodiment of the present invention may be the same, or may be different, that is, the location of each receiver may be distributed or may be the same area, and thus, the application of the embodiment of the present invention.
  • the scenario is more in line with the actual application, and the application scenario is not too restrictive.
  • the data processing center determines the shared support set according to the received observation data and the joint sparsity characteristic.
  • the joint sparse characteristic includes a sparse signal of each receiver including at least one support set, and at least one shared support set exists between sparse signals of at least two receivers, wherein the shared support set is sparse of at least two receivers A set of supports shared between signals, the set of supports being a set of indices of non-zero rows in the matrix of sparse signals for each receiver.
  • each receiver in the embodiment of the present invention includes at least one antenna, and a sparse signal of each antenna of each receiver is a sparse signal, and a sparse signal between the sparse signals of at least two receivers and each receiver The interior exhibits certain joint sparsity characteristics.
  • the wireless communication system includes a receiver, and each receiver includes two antennas.
  • the sparse signal of each antenna of each receiver is composed of antenna numbers, and each The signal of the antenna on the same sparse domain value is a two-dimensional matrix of rows, wherein the sparse signal representing the first receiver is a matrix of M rows and 2 columns.
  • the sparse signal representing the first receiver is a matrix of M rows and 2 columns.
  • the joint sparse characteristic mentioned in the embodiment of the present invention includes a shared support set existing between at least two sparse signals and a support set existing inside each sparse signal.
  • the joint sparseness characteristic in the embodiment of the present invention is different from the joint sparseness characteristic in the distributed theory of distributed sensing in the prior art.
  • the joint sparseness feature used in the existing distributed compressed sensing theory is only applicable to a single-point wireless communication system, and only a shared support set exists between sparse signals of at least two receivers; the joint sparse characteristic of the embodiment of the present invention is not only considered There is a shared support set between the sparse signals of at least two receivers, and a support set existing inside each of the sparse signals is also considered.
  • the joint sparsity feature in the embodiment of the present invention not only considers the joint sparsity characteristics between different sparse signals, but also considers the joint sparsity characteristics inside the sparse signal.
  • the joint sparse feature in the embodiment of the present invention is a new type of feature that is more suitable for common scenarios in a wireless communication system.
  • the data processing center determines the shared support set according to the received observation data and the joint sparsity characteristic.
  • the data processing center may use any existing feasible signal recovery algorithm to determine the shared support set, which is not limited in the embodiment of the present invention.
  • the signal recovery algorithm may be an OMP (Orthogonal Matching Pursuit) or an L-1 norm-based optimization algorithm.
  • the wireless communication system includes three receivers, each of which includes two antennas, representing a sparse signal of the first receiver, and is a matrix of 7 rows and 2 columns, ⁇ , X 2 , 3 ⁇
  • the composition of the signal matrix exists in the embodiment of the present invention
  • the joint sparse property where the shared support set in the signal matrix consisting of ⁇ A , x 2 , 3 ⁇ is the fourth row in the matrix.
  • the 0MP algorithm is an iterative algorithm, in which data processing can find the position of one of the support sets of the sparse signals corresponding to one antenna from the observation data of one antenna in each iteration, that is, in each iteration,
  • the data processing center finds an element of the support set (ie, a row number) from the observation data of each antenna.
  • the data processing center will get 6 line numbers according to the 0MP algorithm, assuming that the 6 line numbers obtained by the data processing center are ⁇ 4, 4, 4, 2, 4, 6 ⁇ , due to the sparse signal of 6 antennas.
  • the shared support set is included, that is, the six sparse signals contain the same row number, so the data center selects the sequence number 4 with the most repetitions from the 6 row numbers according to the joint sparse characteristic as the shared support set in the sparse signal.
  • the data processing center determines the support set of each sparse signal according to the shared support set and the joint sparsity characteristic.
  • the data processing center After determining the shared support set, the data processing center determines the support set of each sparse signal by using any feasible signal recovery algorithm according to the shared support set and the joint sparsity characteristic.
  • the sparse signal components of at least two receivers are grouped by antenna numbers, and the signals of each antenna in the same sparse domain value are two-dimensional matrices of rows, and the shared support set determined by ' ⁇ in the data processing is also an antenna.
  • the serial number is a column, and the signal of each antenna in the same sparse domain value is a matrix of rows, and the data processing center can obtain the signals corresponding to the sparse signals of the respective receivers in the shared support set by the shared support set, so the data According to the shared support set and the joint sparsity feature, the processing center can determine the support set of each sparse signal.
  • the data processing center determines that the element of the shared support set A of ⁇ A , X 2 , 3 ⁇ is ⁇ 4 ⁇
  • the data processing center utilizes sparseness on all antennas of each receiver.
  • the support set of the signals has the same characteristics, and the observation data of all the antennas of each receiver are processed together.
  • the data processing center puts the element ⁇ 4 ⁇ in the shared support set ⁇ ⁇ into the set of support sets of 2 , equivalent to the number of observations already made to the second receiver.
  • the data processing center determines each sparse signal according to the shared support set and the support set of each sparse signal.
  • At least one support set exists in the sparse signal of each receiver, and there is a shared support set between the sparse signals of at least two receivers, and after the data processing center determines the support set of each sparse signal, according to the determined each The sparse signal support set and the shared support set can determine each sparse signal.
  • the data processing center determines that the shared support set of ⁇ A , X 2 , 3 ⁇ is ⁇ ⁇ , and the support set of sparse signal 2 is and , the data processing center obtains the second receiver by column.
  • the sparse signal is in the shared support set and the corresponding signal in 2 , since the sparse signal is mainly composed of these two parts, the data processing center can obtain the sparse signal 2 by using the signal recovery algorithm in the existing compressed sensing method.
  • the sparse signal recovery process in the embodiment of the present invention utilizes the joint sparseness between the sparse signals and the internal, which can greatly reduce the sampling rate or the amount of observed data of each receiver.
  • the observation data of each receiver is the data sampled by the sparse signal, and the data amount of the observation data acquired by each receiver is small, so that each receiver transmits its observation data to the data processing center, reducing the receiver from The backhaul load of the data processing center also reduces the processing complexity of the calculation of each receiver, and the processing delay of each receiver is also reduced accordingly.
  • Embodiments of the present invention provide a compression sensing method. After receiving data of observations sent by at least one receiver, the data processing center determines a shared support set according to all observation data and joint sparsity characteristics, and determines shared support. After the collection, the data processing center determines the support set of each sparse signal according to the shared support set and the joint sparse feature. Finally, the data processing center determines each sparse signal according to the shared support set and the support set of each sparse signal, wherein the observed data is Each receiver samples the sparse signal of the receiver according to the observation matrix of the receiver, and the joint sparsity characteristic has a shared support set between all sparse signals, and at least one support exists in each sparse signal.
  • the support set is an index set of non-zero rows in a matrix corresponding to each sparse signal, and a row in a non-zero behavior matrix whose row vector is not zero.
  • the embodiment of the present invention provides a compression sensing method, and an application scenario thereof is shown in FIG. 9.
  • the entire wireless communication system includes KK ⁇ 2 receivers, and different receivers support backhaul connections, so that each There can be information exchange between receivers.
  • ⁇ is the sparse signal observed by the receiver (l ⁇ i ⁇ K), and the observation data about the sparse signal ⁇ is obtained for the first receiver, and ⁇ ⁇ + ⁇ , ⁇ ;
  • the observation matrix of the receiver is the observation noise of the second receiver.
  • each receiver obtains its own observation data and separately recovers the respective sparse signals through information interaction.
  • the method includes:
  • the first receiver acquires its observation data, wherein the observation data is that the first receiver samples the sparse signal of the first receiver according to its observation matrix to obtain data.
  • the executor of the embodiment of the present invention is a first receiver, wherein the first receiver is any receiver in the wireless communication system, which is not limited in the embodiment of the present invention.
  • the first receiver may be a base station or a base station controller.
  • the compressed sensing method provided by the embodiment of the present invention is applicable to each receiver to obtain respective observation data, and separately recovers the respective sparse signals through information interaction. Therefore, the first receiver first acquires the observation data thereof.
  • the first receiver acquires its observation data and at least two receivers in the S 1 0 1 of the first embodiment sample the sparse signals of the at least two receivers according to the observation matrix of the at least two receivers to obtain observation data.
  • the process is exactly the same and will not be repeated.
  • the first receiver determines the shared support set based on the joint sparsity characteristics and the observed data.
  • the first receiver first determines an estimated support set of the sparse signal of the first receiver according to the joint sparsity characteristic and the observation data thereof, and then the first receiver divides the estimated support set from the at least two receivers Other than a receiver and the first receiver The estimated support set of the receivers with backhaul connections is exchanged through the backhaul connection. Finally, the shared support set is determined according to the joint sparse characteristics.
  • the first receiver exchanges its estimated support set with the estimated support set of the receiver connected to the first receiver supporting the backhaul other than the first receiver through the backhaul connection, which may be the first receiver to estimate its support
  • the set is performed with an estimated support set of each receiver other than the first receiver, or the first receiver sets its estimated support set and the estimated support of each receiver other than the first receiver in the information center.
  • the embodiment of the present invention is not limited.
  • the information center mentioned here may be a simple data exchange center, has no data processing capability, and the data of the information center may be only an estimated support set of each receiver, and only one used for each estimated support set.
  • the bit (0 or 1) indicates; the data processing center in the first embodiment can process all the observation data, and each observation data is a real number, and generally needs 12 bits or more to be accurately represented.
  • the first receiver may use any feasible signal recovery algorithm to determine the shared support set, which is not limited in the embodiment of the present invention.
  • the signal recovery algorithm may be 0MP (Orthogonal Matching Pursuit) or an optimization algorithm based on the Bu-1 norm.
  • the wireless communication system includes three receivers, each of which includes two antennas, representing a sparse signal of the first receiver, and is a matrix of 7 rows and 2 columns, ⁇ , X 2, ⁇
  • the signal matrix of the composition has the joint sparse property mentioned in the embodiment of the present invention.
  • the first receiver uses the same characteristics of the support set of the sparse signals of all its antennas, and the observation data of all the antennas are processed together by the 0MP algorithm.
  • the 0MP algorithm is an iterative algorithm, in which it is possible to find the position of one of the support sets of the corresponding sparse signals from the observation data of one of the antennas of the first receiver, that is, at each time In the iteration, the first receiver finds an element of a support set (ie, a row number) from the observation data of each of its antennas. After one iteration, the first receiver exchanges the elements of the estimated support set with the estimated support set line number of the second receiver and the estimated support set line number of the third receiver through a backhaul connection.
  • the sparse signals of the three receivers contain a shared support set, ie 3
  • the sparse signals of the receivers contain the same row number, so the first receiver is based on the joint sparse characteristic after obtaining the estimated support set of other receivers that support the backhaul connection with the first receiver.
  • the feature with the shared support set contains the most repeated elements (in this example, ⁇ 4 ⁇ ) as an element in the shared support set.
  • the first receiver repeats the above steps in accordance with the number of elements T 1 of the preset shared support set to obtain a shared support set.
  • the first receiver determines, according to the shared support set and the joint sparse characteristic, a support set of the sparse signal of the first receiver.
  • the first receiver After determining the shared support set, the first receiver determines the support set of the sparse signal of the first receiver using any feasible signal recovery algorithm based on the shared support set and the joint sparsity characteristic.
  • all the sparse signal components are arranged by the antenna serial number, and the signal of each antenna in the same sparse domain value is a two-dimensional matrix of rows, and the shared support set determined by the first receiver is also listed by the antenna serial number, and each The signal of the antenna on the same domain value is a matrix of rows.
  • the first receiver can determine the support set of its sparse signal according to the shared support set and the joint sparsity characteristic. The method is the same as the method for determining the support set of each sparse signal by the data processing center in S 1 03 of the first embodiment, and details are not described herein again.
  • the first receiver determines the sparse signal according to the support set of the shared support set and the sparse signal of the first receiver.
  • At least one support set exists in the sparse signal of each receiver, and there is a shared support set between all sparse signals, and after the first receiver determines the support set of the sparse signal, the support set according to the determined sparse signal is determined. With the shared support set, determine its sparse signal.
  • the sparse signal recovery process in the embodiment of the present invention utilizes the joint sparseness between the sparse signals and the internals, which can greatly reduce the sampling rate or the amount of observed data of each receiver.
  • Each receiver distributes the recovered sparse signals in a distributed manner with high flexibility and robustness, and only needs to exchange the estimated support set information of the sparse signals between the receivers, and there is no other information, so the backhaul between the receivers is The load is small.
  • Embodiments of the present invention provide a compression sensing method, in which a first receiver acquires its view After measuring the data, the shared support set is determined according to the joint sparsity characteristic and the observation data thereof, and then the first receiver determines the support set of the sparse signal according to the shared support set and the joint sparse characteristic, and finally, the first receiver is based on the shared support And a set of support of the sparse signal, the sparse signal of the first receiver is determined, wherein the observation data is that the first receiver samples the sparse signal according to the observation matrix of the first receiver, and the joint sparsity characteristic is all sparse signals There is a shared support set between each of the sparse signals, and at least one support set exists in the sparse signal.
  • the support set is an index set of non-zero rows in the matrix corresponding to each sparse signal, and the rows in the non-zero behavior matrix are not zero.
  • the compressed sensing method provided by the present invention relates to a joint sparse feature widely applicable to a wireless communication scenario, and the method can recover all the observed data together to recover the sparse signal, or recover the respective data from each receiver.
  • the sparse signal that is, the application scenario of the first embodiment or the application scenario of the second embodiment
  • the sparse signal recovery process the shared support set of all sparse signals is first recovered by using the joint sparse feature of the sparse signal. Then, the support set of each sparse signal is recovered, and finally each individual sparse signal is recovered.
  • the recovery process of the entire sparse signal is applied with the joint sparse characteristic and the compressed sensing theory, which greatly reduces the acquisition cost of the sparse signal.
  • the units may or may not be physically separated, and the components displayed as units may be one physical unit or multiple physical units, that is, may be located in one place. Or it can be distributed to many different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • the integrated unit if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a readable storage medium.
  • the technical solution of the present invention may be embodied in the form of a software product in the form of a software product, or a part of the technical solution, which is stored in a storage medium.
  • a number of instructions are included to cause a device (which may be a microcontroller, chip, etc.) or a processor to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access memory, a magnetic disk, or an optical disk, and the like, which can store program codes.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Radio Transmission System (AREA)

Abstract

本发明实施例提供一种压缩感知方法及装置，涉及无线通信领域，能够降低捕获稀疏信号的代价。该方法包括：接收至少两个接收机发送的观测数据，观测数据为至少两个接收机根据至少两个接收机的观测矩阵对至少两个接收机的稀疏信号采样得到的数据；根据观测数据和联合稀疏特性，确定共享支撑集，联合稀疏特性为各个接收机的稀疏信号包含至少一个支撑集，且至少两个接收机的稀疏信号之间存在至少一个共享支撑集，共享支撑集为至少两个接收机的稀疏信号之间共享的支撑集；根据共享支撑集和联合稀疏特性，确定各个稀疏信号的支撑集；根据共享支撑集和各个稀疏信号的支撑集，确定各个稀疏信号。

Description

一种压缩感知方法及装置
技术领域
本发明涉及无线通信领域， 尤其涉及一种压缩感知方法及装置。 背景技术
随着无线通信系统中维度的增加，获取通信系统中高维度的目标 信号（无线网络中需要捕获和恢复的信号）变得异常困难。 例如， 无线 通信系统中超宽频段的信号若采用传统的 ADC ( Analog-to-Digital Converter, 模拟信号到信息的转换） 采样， 则 ADC 的采样速率需随 信号频段宽度线性增长， 才能无损采样。 由于制作工艺的限制， 高速 率的 ADC硬件成本会很高， 采用传统 ADC采样技术会使得超宽频无线 通信的硬件成本急剧上升。 另外， 大规模天线的 FDD ( Frequency Division Duplex, 频分双工）无线通信系统中获取 CSIT ( Channel State Information at The Transmitters, 发射端的信道信息状态 ) 的传统的方法要求 BS ( Base Station, 基站） 端发送训练序列长度 随 BS端的天线数目线性增长，这样导致 BS端非常大的训练序列开销， 将会极大的降低大规模天线所能够带来的空域复用增益和分集增益 等。
此外， 随着无线通信系统中目标信号维度不断上升， 目标信号一 般会出现稀疏特性。 例如， 在超宽频带通信系统中， 由于某些应用的 突发传输特性 （移动用户的语音业务， 用户会有大量的空闲时间 ）， 信号的频段利用情况可能是稀疏的。 另外， 由于无线传输环境的共享 特性， 目标信号内部以及之间往往还表现很强的关联特性， 即目标信 号往往会呈现联合稀疏性。例如，在一个分布式天线的 M I M 0( M u 11 i p 1 e Input Multiple Output , 多天线输入输出 ) 通信系统中， 同一个站 点的天线所能观测到的活跃用户是一致的， 不同的站点可能会共享一 部分活跃用户 。 针对目标信号出现的稀疏特性， 现在已经提出 了 CS (Compr es s ed Sensing, 压缩感知； 或 Compressive Samp ling, 压 缩采样）理论。 该理论指出， 当 目标信号呈现稀疏特性的时候， 只需 要拥有少量采样值或观测数据， 就可以将目标信号有效地恢复出来。 目前， 已经有把 cs 理论引入到无线通信系统中的一些技术方法。 但 是， 这些技术方法往往只是考虑单点对单点的系统并采取最直接的引 入方法， 而对高维目标信号（稀疏信号）的获取代价依旧较大。
发明内容
本发明的实施例提供一种压缩感知方法及装置，解决了现有无线 通信系统中捕获高维目标信号代价较高的问题。
为达到上述目的， 本发明的实施例采用如下技术方案：
第一方面， 本发明实施例提供数据处理中心， 包括：
接收单元， 用于接收至少两个接收机发送的观测数据， 所述观测 数据为所述至少两个接收机根据所述至少两个接收机的观测矩阵对 所述至少两个接收机的稀疏信号采样得到的数据；
确定单元，用于根据所述接收单元接收到的观测数据和联合稀疏 特性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀疏信号 包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在 至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机 的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信号 的矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和所述联 合稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述共享支 撑集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。
在第一方面的第一种可能的实现方式中，所述联合稀疏特性为各 个接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机 的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为 所述至少两个接收机的稀疏信号之间公共的支撑集， 所述支撑集为各 个接收机的稀疏信号的矩阵中非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
在第一方面的第二种可能的实现方式中，所述至少两个接收机根 据以下公式对至少两个接收机的稀疏信号采样得到观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量，
Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
第二方面， 本发明实施例提供一种接收机， 包括：
获取单元， 用于获取第一接收机的观测数据， 所述观测数据为所 述第一接收机根据所述第一接收机的观测矩阵对所述第一接收机的 稀疏信号采样得到数据；
确定单元，用于根据联合稀疏特性和所述获取单元获取到的观测 数据， 确定共享支撑集， 所述联合稀疏特性为至少两个接收机中各个 接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的 稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所 述至少两个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个 接收机的稀疏信号的矩阵中非零行的索引集合， 所述第一接收机与所 述第二接收机之间支持回程连接， 以及用于根据所述共享支撑集和所 述联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑集， 以及用 于根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集， 确定 所述第一接收机的稀疏信号。
在第二方面的第一种可能的实现方式中，所述联合稀疏特性为至 少两个接收机中各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共享的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
结合第二方面的第一种可能的实现方式，在第一方面的第二种可 能的实现方式中， 所述确定单元， 具体用于根据所述联合稀疏特性和 所述观测数据， 确定所述第一接收机的稀疏信号的估计支撑集， 以及 具体用于将所述估计支撑集与第二接收机的估计支撑集交换， 并根据 所述联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至少 两个接收机中除所述第一接收机以外的其他与所述第一接收机支持 回程连接的接收机。
在第一方面的第三种可能的实现方式中，所述第一接收机根据以 下公式对所述第一接收机的稀疏信号采样得到观测数据：
= Φ, + N i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,.为第 个接收机的观测矩阵， 为第 ζ'个接收机 的观测噪声， . e CMxiV为 Μ行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ^ ( ΤχΜ为 Γ行 Μ列的矩阵， N CNxT为 N行 Γ列的矩阵。 第三方面， 本发明实施例提供一种数据处理中心， 包括： 接收器， 用于接收至少两个接收机发送的观测数据， 所述观测数 据为所述至少两个接收机根据所述至少两个接收机的观测矩阵对所 述至少两个接收机的稀疏信号采样得到的数据；
处理器， 用于根据所述接收器接收到的观测数据和联合稀疏特 性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀疏信号包 含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在至 少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的 稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信号的 矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和所述联合 稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述共享支撑 集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。
在第三方面的第一种可能的实现方式中，所述联合稀疏特性为各 个接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机 的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为 所述至少两个接收机的稀疏信号之间公共的支撑集， 所述支撑集为各 个接收机的稀疏信号的矩阵中非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
在第三方面的第二种可能的实现方式中，所述至少两个接收机根 据以下公式对至少两个接收机的稀疏信号采样得到观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ 其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 ζ'个接收机 的观测噪声， . e CMxiV为 Μ行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
结合前述的第三方面或者第三方面的第一种可能的实现方式至 第二种可能的实现方式中的任意一种， 在第三方面的第三种可能的实 现方式中所述数据处理中心包括中央处理器。
第四方面， 本发明实施例提供一种接收机， 包括：
处理器， 用于获取第一接收机的观测数据， 所述观测数据为所述 第一接收机根据所述第一接收机的观测矩阵对所述第一接收机的稀 疏信号采样得到数据， 以及用于根据联合稀疏特性和获取到的观测数 据， 确定共享支撑集， 所述联合稀疏特性为至少两个接收机中各个接 收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的稀 疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述 至少两个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个接 收机的稀疏信号的矩阵中非零行的索引集合， 所述第一接收机与所述 第二接收机之间支持回程连接， 以及用于根据所述共享支撑集和所述 联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑集， 以及用于 根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集， 确定所 述第一接收机的稀疏信号。
在第四方面的第一种可能的实现方式中，所述联合稀疏特性为至 少两个接收机中各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共享的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
结合第四方面的第一种可能的实现方式，在第四方面的第二种可 能的实现方式中， 所述处理器， 具体用于根据所述联合稀疏特性和所 述观测数据， 确定所述第一接收机的稀疏信号的估计支撑集， 以及具 体用于将所述估计支撑集与第二接收机的估计支撑集交换， 并根据所 述联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至少两 个接收机中除所述第一接收机以外的其他与所述第一接收机支持回 程连接的接收机。
在第四方面的第三种可能的实现方式中，所述第一接收机根据以 下公式对所述第一接收机的稀疏信号采样得到观测数据：
Υ = ΦίΧί + Νί i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
结合前述的第四方面或第四方面的第一种可能的实现方式至第 三种可能的实现方式中的任意一种， 在第四方面的第四种可能的实现 方式中， 所述接收机包括基站或者基站控制器。
第五方面， 本发明实施例提供一种压缩感知方法， 其特征在于， 包括： 接收至少两个接收机发送的观测数据，所述观测数据为所述至少 两个接收机才艮据所述至少两个接收机的观测矩阵对所述至少两个接 收机的稀疏信号采样得到的数据；
根据接收到的观测数据和联合稀疏特性， 确定共享支撑集， 所述 联合稀疏特性为各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共享的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合；
根据所述共享支撑集和所述联合稀疏特性，确定各个稀疏信号的 支撑集；
根据所述共享支撑集和所述各个稀疏信号的支撑集，确定各个稀 疏信号。
在第五方面的第一种可能的实现方式中，所述联合稀疏特性为各 个接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机 的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为 所述至少两个接收机的稀疏信号之间公共的支撑集， 所述支撑集为各 个接收机的稀疏信号的矩阵中非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
在第五方面的第二种可能的实现方式中，所述至少两个接收机根 据以下公式对至少两个接收机的稀疏信号采样得到观测数据： γ. = Φ.χ. + N. ί = ι，.··
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 ζ'个接收机 的观测噪声， . e CMxiV为 Μ行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
第六方面， 本发明实施例提供压缩感知方法， 包括：
获取第一接收机的观测数据，所述观测数据为所述第一接收机根 据所述第一接收机的观测矩阵对所述第一接收机的稀疏信号采样得 到数据；
根据联合稀疏特性和所述观测数据， 确定共享支撑集， 所述联合 稀疏特性为至少两个接收机中各个接收机的稀疏信号包含有至少一 个支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个共享 支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之 间共享的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零 行的索引集合， 所述第一接收机与所述第二接收机之间支持回程连 接；
根据所述共享支撑集和所述联合稀疏特性，确定所述第一接收机 的稀疏信号的支撑集；
根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集，确 定所述第一接收机的稀疏信号。
在第六方面的第一种可能的实现方式中，所述联合稀疏特性为至 少两个接收机中各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共享的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
结合第六方面的第一种可能的实现方式，在第六方面的第二种可 能的实现方式中， 所述根据联合稀疏特性和所述观测数据， 确定共享 支撑集， 具体包括：
根据所述联合稀疏特性和所述观测数据，确定所述第一接收机的 稀疏信号的估计支撑集；
将所述估计支撑集与第二接收机的估计支撑集交换，并根据所述 联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至少两个 接收机中除所述第一接收机以外的其他与所述第一接收机支持回程 连接的接收机。
在第六方面的第三种可能的实现方式中，所述第一接收机根据以 下公式对所述第一接收机的稀疏信号采样得到观测数据：
= Φ, + N i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,.为第 个接收机的观测矩阵， 为第 ζ'个接收机 的观测噪声， . e CMxiV为 Μ行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ^ ( ΤχΜ为 Γ行 Μ列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
本发明的实施例提供一种压缩感知方法，数据处理中心在接收到 至少一个接收机发送的观测数据后， 数据处理中心根据所有观测数据 和联合稀疏特性， 确定共享支撑集， 并在确定共享支撑集后， 数据处 理中心根据共享支撑集和联合稀疏特性， 确定各个稀疏信号的支撑 集， 最后， 数据处理中心根据共享支撑集和各个稀疏信号的支撑集， 确定各个稀疏信号， 其中， 观测数据为各个接收机根据该接收机的观 测矩阵对该接收机的稀疏信号采样得到的数据， 联合稀疏特性为所有 稀疏信号之间存在共享支撑集， 各个稀疏信号内部存在至少一个支撑 集， 支撑集为与各个稀疏信号对应的矩阵中非零行的索引集合， 非零 行为矩阵中行向量不为零的行。 通过该方案， 本发明利用稀疏信号内 部和之间的联合稀疏性， 解决了现有无线通信系统中捕获稀疏信号代 价较高的问题， 降低捕获稀疏信号的代价。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面 将对实施例或现有技术描述中所需要使用的附图作简单地介绍， 显而 易见地， 下面描述中的附图仅仅是本发明的一些实施例， 对于本领域 普通技术人员来讲， 在不付出创造性劳动的前提下， 还可以根据这些 附图获得其他的附图。
图 1为本发明实施例的数据处理中心结构示意图一；
图 2为本发明实施例的接收机结构示意图一；
图 3为本发明实施例的数据处理中心结构示意图二；
图 4为本发明实施例的接收机结构示意图二；
图 5为本发明实施例的压缩感知方法的应用场景图一；
图 6为本发明实施例的压缩感知方法流程示意图一；
图 7为本发明实施例的联合稀疏特性结构示意图一；
图 8为本发明实施例的联合稀疏特性结构示意图二；
图 9为本发明实施例的压缩感知方法的应用场景图二；
图 1 0为本发明实施例的压缩感知方法流程示意图二。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方 案进行清楚、 完整地描述， 显然， 所描述的实施例仅仅是本发明一部 分实施例， 而不是全部的实施例。 基于本发明中的实施例， 本领域普 通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例， 都属于本发明保护的范围。
本文中描述的各种技术可用于各种无线通信系统中，例如全球移 动通信系统 ( GSM, global system for mobile communications ), 码分多址 ( CDMA, code division multiple access ) 系统, 时分多 址 ( TDMA, time division multiple access ) 系统， 宽带码分多址 ( WCDMA , wideband code division multiple access wireless ), 频分多址 ( FDMA, frequency division multiple addressing ) 系统 正交频分多址 ( OFDMA , orthogonal frequency-division multiple access )系统，单载波 FDMA( SC-FDMA )系统，通用分组无线业务（ GPRS, genera 1 packet radio service ) 系统, 长期演进 ( LTE, long term evolution ) 系统， 以及其他此类通信系统。
本发明技术方案考虑的是无线通信系统中信号捕获的框架，该技 术框架可以覆盖很多具体的应用问题， 比如， 该技术框架可以应用到 分布式的 RF ( Radio Frequency, 射频） 端模拟到信息的转换问题， 多用户大规模天线网络中 CSIT的获取问题等。
实施例一
本发明实施例提供一种数据处理中心 1, 如图 1所示， 包括： 接收单元 10, 用于接收至少两个接收机发送的观测数据， 所述 观测数据为所述至少两个接收机根据所述至少两个接收机的观测矩 阵对所述至少两个接收机的稀疏信号采样得到的数据。
确定单元 11, 用于根据所述接收单元 10接收到的观测数据和联 合稀疏特性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀 疏信号包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之 间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个 接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀 疏信号的矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和 所述联合稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述 共享支撑集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。 进一步地，所述联合稀疏特性为各个接收机的稀疏信号包含有至 少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个 共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信 号之间公共的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中 非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵。
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合。
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
进一步地，所述至少两个接收机才艮据以下公式对至少两个接收机 的稀疏信号采样得到观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
本发明实施例提供一种数据处理中心，主要包括接收单元和处理 单元。 数据处理中心在接收到至少一个接收机发送的观测数据后， 数 据处理中心根据所有观测数据和联合稀疏特性， 确定共享支撑集， 并 在确定共享支撑集后， 数据处理中心根据共享支撑集和联合稀疏特 性， 确定各个稀疏信号的支撑集， 最后， 数据处理中心根据共享支撑 集和各个稀疏信号的支撑集， 确定各个稀疏信号， 其中， 观测数据为 各个接收机根据该接收机的观测矩阵对该接收机的稀疏信号采样得 到的数据， 联合稀疏特性为所有稀疏信号之间存在共享支撑集， 各个 稀疏信号内部存在至少一个支撑集， 支撑集为与各个稀疏信号对应的 矩阵中非零行的索引集合， 非零行为矩阵中行向量不为零的行。 通过 该方案， 本发明利用稀疏信号内部和之间的联合稀疏性， 解决了现有 无线通信系统中捕获稀疏信号代价较高的问题， 降低捕获稀疏信号的 代价。
实施例二
本发明实施例提供一种接收机 1 , 如图 2所示， 包括：
获取单元 1 0 , 用于获取第一接收机的观测数据， 所述观测数据 为所述第一接收机根据所述第一接收机的观测矩阵对所述第一接收 机的稀疏信号采样得到数据。
确定单元 1 1 , 用于根据联合稀疏特性和所述获取单元 1 0获取到 的观测数据， 确定共享支撑集， 所述联合稀疏特性为至少两个接收机 中各个接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接 收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑 集为所述至少两个接收机的稀疏信号之间共享的支撑集， 所述支撑集 为各个接收机的稀疏信号的矩阵中非零行的索引集合， 所述第一接收 机与所述第二接收机之间支持回程连接， 以及用于根据所述共享支撑 集和所述联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑集， 以及用于根据所述共享支撑集和所述第一接收机的稀疏信号的支撑 集， 确定所述第一接收机的稀疏信号。
进一步地，所述联合稀疏特性为至少两个接收机中各个接收机的 稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号 之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两 个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的 稀疏信号的矩阵中非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵。
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合。
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
进一步地， 所述确定单元 1 1 , 具体用于根据所述联合稀疏特性 和所述观测数据， 确定所述第一接收机的稀疏信号的估计支撑集， 以 及具体用于将所述估计支撑集与第二接收机的估计支撑集交换， 并根 据所述联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至 少两个接收机中除所述第一接收机以外的其他与所述第一接收机支 持回程连接的接收机。
进一步地，所述第一接收机根据以下公式对所述第一接收机的稀 疏信号采样得到观测数据：
= Φ, + N i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量，
Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
本发明实施例提供一种接收机， 主要包括获取单元和确定单元。 接收机在获取其观测数据后， 根据联合稀疏特性和其观测数据， 确定 共享支撑集， 然后， 根据共享支撑集和联合稀疏特性， 确定其稀疏信 号的支撑集， 最后， 根据共享支撑集和其稀疏信号的支撑集， 确定第 一接收机的稀疏信号， 其中， 观测数据为根据该接收机的观测矩阵对 其稀疏信号采样得到数据， 联合稀疏特性为所有稀疏信号之间存在共 享支撑集， 各个稀疏信号内部存在至少一个支撑集， 支撑集为与各个 稀疏信号对应的矩阵中非零行的索引集合， 非零行为矩阵中行向量不 为零的行。 通过该方案， 本发明利用稀疏信号内部和之间的联合稀疏 性， 解决了现有无线通信系统中捕获高维稀疏信号代价较高的问题， 降低捕获稀疏信号的代价。
实施例三
本发明实施例提供一种数据处理中心， 如图 3所示， 该数据处理 中心可以包括接收器 10、 处理器 11、 存储器 12 和系统总线 13, 其 中，
接收器 10、 处理器 11和存储器 12通过系统总线 13连接并完成 相互间通信。
处理器 11 可能为单核或多核中央处理单元， 或者为特定集成电 路， 或者为被配置成本发明实施例的一个或多个集成电路。
存储器 12 可以为高速 RAM ( Random Access Memory, 随机存储 器） 存储器， 也可以为非易失性存储器 （ non-volatile memory ), 例 如至少一个磁盘存储器。
存储器 12用于存储数据处理中心的执行指令。 具体的， 数据处 理中心的执行指令中可以包括软件程序和软件代码。
具体的， 接收器 10, 用于接收至少两个接收机发送的观测数据， 所述观测数据为所述至少两个接收机才艮据所述至少两个接收机的观 测矩阵对所述至少两个接收机的稀疏信号采样得到的数据。
处理器 11, 用于根据所述接收器 10接收到的观测数据和联合稀 疏特性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀疏信 号包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之间存 在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收 机的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信 号的矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和所述 联合稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述共享 支撑集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。
进一步地，所述联合稀疏特性为各个接收机的稀疏信号包含有至 少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个 共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信 号之间公共的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中 非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵。
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合。
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
进一步地，所述至少两个接收机才艮据以下公式对至少两个接收机 的稀疏信号采样得到观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量，
Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
具体的， 所述数据处理中心可以为中央处理器， 本发明实施例不 做限定。
本发明实施例提供一种数据处理中心，数据处理中心在接收到至 少一个接收机发送的观测数据后， 数据处理中心根据所有观测数据和 联合稀疏特性， 确定共享支撑集， 并在确定共享支撑集后， 数据处理 中心根据共享支撑集和联合稀疏特性， 确定各个稀疏信号的支撑集， 最后， 数据处理中心根据共享支撑集和各个稀疏信号的支撑集， 确定 各个稀疏信号， 其中， 观测数据为各个接收机根据该接收机的观测矩 阵对该接收机的稀疏信号采样得到的数据， 联合稀疏特性为所有稀疏 信号之间存在共享支撑集， 各个稀疏信号内部存在至少一个支撑集， 支撑集为与各个稀疏信号对应的矩阵中非零行的索引集合， 非零行为 矩阵中行向量不为零的行。 通过该方案， 本发明利用稀疏信号内部和 之间的联合稀疏性， 解决了现有无线通信系统中捕获稀疏信号代价较 高的问题， 降低捕获稀疏信号的代价。
实施例四
本发明实施例提供一种接收机， 如图 4所示， 该接收机可以包括 处理器 10、 存储器 11和系统总线 12, 其中，
处理器 10和存储器 11通过系统总线 12连接并完成相互间通信。 处理器 10可能为单核或多核中央处理单元， 或者为特定集成电 路， 或者为被配置成本发明实施例的一个或多个集成电路。
存储器 11 可以为高速 RAM ( Random Access Memory, 随机存储 器） 存储器， 也可以为非易失性存储器 （ non-volatile memory ), 例 如至少一个磁盘存储器。
存储器 11 用于存储数据处理中心的执行指令。 具体的， 数据处 理中心的执行指令中可以包括软件程序和软件代码。
具体的， 处理器 10, 用于获取第一接收机的观测数据， 所述观 测数据为所述第一接收机根据所述第一接收机的观测矩阵对所述第 一接收机的稀疏信号采样得到数据， 以及用于根据联合稀疏特性和获 取到的观测数据， 确定共享支撑集， 所述联合稀疏特性为至少两个接 收机中各个接收机的稀疏信号包含有至少一个支撑集， 且所述至少两 个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享 支撑集为所述至少两个接收机的稀疏信号之间共享的支撑集， 所述支 撑集为各个接收机的稀疏信号的矩阵中非零行的索引集合， 所述第一 接收机与所述第二接收机之间支持回程连接， 以及用于根据所述共享 支撑集和所述联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑 集， 以及用于根据所述共享支撑集和所述第一接收机的稀疏信号的支 撑集， 确定所述第一接收机的稀疏信号。
进一步地，所述联合稀疏特性为至少两个接收机中各个接收机的 稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号 之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两 个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的 稀疏信号的矩阵中非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵。
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合。
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
进一步地， 所述处理器 1 0 , 具体用于根据所述联合稀疏特性和 所述观测数据， 确定所述第一接收机的稀疏信号的估计支撑集， 以及 具体用于将所述估计支撑集与第二接收机的估计支撑集交换， 并根据 所述联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至少 两个接收机中除所述第一接收机以外的其他与所述第一接收机支持 回程连接的接收机。
进一步地，所述第一接收机根据以下公式对所述第一接收机的稀 疏信号采样得到观测数据：
= Φ, + N i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Y CTxN为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ 。ΤχΜ为 Γ行 Μ列的矩阵， N^ CNxT为 N行 Γ列的矩阵。
进一步地， 所述接收机可以为基站， 也可以为基站控制器， 本发 明实施例不做限定。
本发明实施例提供一种接收机， 接收机在获取其观测数据后， 根 据联合稀疏特性和其观测数据， 确定共享支撑集， 然后， 根据共享支 撑集和联合稀疏特性， 确定其稀疏信号的支撑集， 最后， 根据共享支 撑集和其稀疏信号的支撑集， 确定其稀疏信号， 其中， 观测数据为根 据第一接收机的观测矩阵对其稀疏信号采样得到数据， 联合稀疏特 ' )·生 为所有稀疏信号之间存在共享支撑集， 各个稀疏信号内部存在至少一 个支撑集， 支撑集为与各个稀疏信号对应的矩阵中非零行的索引集 合， 非零行为矩阵中行向量不为零的行。 通过该方案， 本发明利用稀 疏信号内部和之间的联合稀疏性， 解决了现有无线通信系统中捕获高 维稀疏信号代价较高的问题， 降低捕获稀疏信号的代价。
实施例五
本发明实施例提供一种压缩感知方法， 其应用场景如图 5所示， 整个无线通信系统中包含有一个数据处理中心和 ( ≥2)个接收机， 其中， ^为第 ( 1≤ ≤ ) 个接收机观测到的稀疏信号， 为第 个接 收机获得关于其稀疏信号^的观测数据， 且 ζ ΦιΑ + Λ^ , Φ;为第 个接 收机的观测矩阵， 为第 ζ'个接收机的观测噪声， 个接收机将获得 的观测数据 { }分别送回数据处理中心， 在该无线系统中， 由数据处 理中心集中式的恢复出稀疏信号。 如图 6所示， 该方法包括：
S 1 01、数据处理中心接收至少两个接收机发送的观测数据，其中， 观测数据为至少两个接收机根据至少两个接收机的观测矩阵对至少 两个接收机的稀疏信号采样得到的数据。
本发明实施例中， 数据处理中心可以为中央处理器。
本发明实施例的应用场景为由数据处理中心集中式的恢复出稀 疏信号， 因此， 数据处理中心首先接收至少两个接收机发送的观测数 据。
在无线通信系统中， 随着目标信号维度不断上升， 目标信号一般 会出现稀疏特性。 比如， 超宽频带通信系统中， 由于某些应用的突发 传输特性 （移动用户的语音业务， 用户会有大量的空闲时间 ）， 信号 的频段利用情况可能是稀疏的； 或者， 在大规模天线网络中， 由于空 间的有限散射性等，从 BS( Base Station,基站）到 MS ( Mobile Users, 移动用户 ） 的 CSI ( Channel State Information, 信道状态信息） 可能会出现稀疏特性； 或者， 在随机接入通信网络中， 真正活跃的用 户往往是稀疏的。 本发明实施例提供的压缩感知方法解决无线通信系 统中高维目标信号捕获代价大的问题， 主要针对高维目标信号， 因此 本发明实施例中的目标信号为稀疏信号。
具体的，稀疏信号是指在允许较小损失的情况下可以通过较少的 系数表达出来的信号。 示例性的， 信号 的长度为 N, 如果信号 能 够用一组基^^二^^ ，…， ^… ^^：^其中， ψτ代表 Ψ的转置） 的线 性组合表示， 则：
x =∑^kak =^a ,
式 — α与 JC是 N行 1列的矩阵， Ψ是 N行 N列的矩阵。 当信号 JC 在某个基 Ψ上仅有 «N个非零系数 时， 信号 为稀疏信号， 具有 可压缩性， Ψ称为信号 X的稀疏域。
需要说明的是，本发明实施例中的稀疏信号的稀疏域可以为信号 频段， 也可以是信道信息， 本发明实施例不做限定。
具体的， 本发明实施例中的各个接收机包含至少一个天线， 至少 两个接收机的每个天线的稀疏信号组成以天线序号为列， 每个天线在 同一稀疏域域值上的信号为行的二维矩阵。 各个接收机根据该接收机 的观测矩阵采用以下公式对该接收机的稀疏信号采样获得其观测数 据：
γ. = Φ.χ. + N. ί = ι，.··
其中， 为第 个接收机的稀疏信号， 为 的观测数据， Φ,.为第 个接收机的观测矩阵， Λ ^为第 ζ'个接收机的观测噪声， 为接收机的 数量， 且 ≥2, .eCMxiV为 Μ行 N列的矩阵， N为第 个接收机包含 的天线数量， M为二维矩阵中行的数量（它取决于稀疏信号的特性（例 如， 若 代表信道时域信息时， 信道时延和测量的精度决定 M 的大 小）， 可以用 > > 1 (远大于 1 ) 表示， 在稀疏度足够高而要求不高的时 候， M可能为几十，要求高的时候 M以大于 1 00甚至更大为好）， N≥l , Yt CTxN ^ Γ行 N列的矩阵， Γ为观测信号的数量或称为采集的样点的 数目 （Γ的值与稀疏度和 M都有关， 以 S 表示稀疏度， 则 T 一般在 S * l og (M/ S)的量级）， O^ e C^M为 Γ行 M列的矩阵， Nt & CNxT为 N行 T 的矩阵。
需要说明的是，本发明实施例中各个接收机采用的观测矩阵可以 相同， 也可以不同， 本发明实施例不做限定， 具体的， 观测矩阵是由 各个接收机根据其所在位置的信道信息与其他因素决定的。
另， 由于本发明实施例中各个接收机的观测矩阵可以相同， 也可 以不同， 也就是说各个接收机的位置可以是分布的， 也可以是同处一 个地区， 这样， 本发明实施例的应用场景更加符合实际应用， 应用场 景没有太大限制。
S 1 02、 数据处理中心根据接收到的观测数据和联合稀疏特性， 确 定共享支撑集。
其中，联合稀疏特性为各个接收机的稀疏信号包含有至少一个支 撑集， 且至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 共享支撑集为至少两个接收机的稀疏信号之间共享的支撑集， 支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集合。
具体的， 本发明实施例中的各个接收机包含至少一个天线， 各个 接收机的每个天线的稀疏信号均为稀疏信号， 且至少两个接收机的稀 疏信号之间与各个接收机的稀疏信号内部表现出一定的联合稀疏特 性。
示例性的， 如图 7所示， 以无线通信系统中包含有 个接收机， 各个接收机包含两个天线为例， 个接收机的每个天线的稀疏信号组 成以天线序号为列， 每个天线在同一稀疏域域值上的信号为行的二维 矩阵， 其中， 代表第 个接收机的稀疏信号， 是 M行 2 列的矩阵。 同一个矩阵 内部， 存在同时全不为零的行向量， 即各个接收机的稀 疏信号内部存在至少一个索引集合^^为其支撑集， 也就是其中一个 接收机的所有天线的稀疏信号的支撑集相同； 个矩阵 A, χ2, ......，
^^存在一个共享支撑集，即存在一个索引集合 Ωε，使得 A, Χ2, … …， ^^在索引属于 的行上面都是非零的。 在图 3 中， 对于第二个接收 机的稀疏信号 2而言，除了包括属于第二个接收机与其他接收机之间 存在的共享支撑集 Ωε的一部分外， 还包括稀疏信号 2内部存在的两 个支撑集 和 。
由上可知，本发明实施例中所提到的联合稀疏特性包含至少两个 稀疏信号之间存在的共享支撑集和各个稀疏信号内部存在的支撑集。
需要说明的是，本发明实施例中的联合稀疏特性与现有理论界的 分布式压缩感知理论中的联合稀疏特性是不同的。 现有的分布式压缩 感知理论使用的联合稀疏特性仅适用于单点无线通信系统， 只考虑了 至少两个接收机的稀疏信号之间存在共享支撑集； 本发明实施例的联 合稀疏特性不仅考虑了至少两个接收机的稀疏信号之间存在共享支 撑集， 还考虑了各个稀疏信号内部存在的支撑集。
综上所述，本发明实施例中的联合稀疏特性不仅仅考虑了不同稀 疏信号之间的联合稀疏特性， 而且还考虑了稀疏信号内部的联合稀疏 特性。 与现有的联合稀疏特性相比， 本发明实施例中的联合稀疏特性 是一种新型的、 更适用于无线通信系统中常见场景的特性。
具体的，数据处理中心在接收到至少两个接收机发送的观测数据 后， 该数据处理中心根据接收到的观测数据和联合稀疏特性， 确定共 享支撑集。
其中，数据处理中心可以采用任何现有可行的信号恢复算法确定 共享支撑集， 本发明实施例不做限定。 示例性的， 信号恢复算法可以 为 OMP ( Orthogonal Matching Pursuit, 正交匹 西己追踪算法）， 也可 以为基于 L-1 范数的优化算法。
示例性的， 如图 8所示， 无线通信系统中包含有 3个接收机， 每 个接收机包含两个天线， 代表第 个接收机的稀疏信号， 是 7 行 2 列的矩阵， { ， X2, 3}组成的信号矩阵中存在本发明实施例中提 到的联合稀疏特性， 其中， {A , x2 , 3 }组成的信号矩阵中共享支 撑集所在位置为矩阵中第 4行。 数据处理中心在接收到 3个接收机发 送的观测数据后， 利用 0MP算法对每个接收机的每个天线的观测数据 进行处理。 0MP算法是迭代算法， 在每次迭代中数据处理中能够从其 中一个天线的观测数据中找出与其相对应的稀疏信号的其中一个支 撑集的所在位置， 也就是说， 在每一次迭代中， 数据处理中心从每个 天线的观测数据中找出一个支撑集的元素 （即一个行序号）。 在本示 例中， 数据处理中心根据 0MP算法会得到 6个行序号， 假设数据处理 中心得到的 6 个行序号为 {4， 4， 4， 2， 4， 6} , 由于 6 个天线的稀疏信号 包含有共享支撑集， 即这 6个稀疏信号中包含有相同的行序号， 所以 数据中心根据联合稀疏特性从这 6 个行序号中选出其中重复次数最 多的序号 4 作为共享支撑集在稀疏信号的矩阵中的行序号， 进一步 地， 数据处理中心根据预先设定的共享支撑集中元素的个数 进行 T1 = l次迭代， 得到共享支撑集的完整集合。
S 1 0 3、 数据处理中心根据共享支撑集和联合稀疏特性， 确定各个 稀疏信号的支撑集。
数据处理中心在确定共享支撑集后，根据该共享支撑集和联合稀 疏特性， 利用任何可行的信号恢复算法确定各个稀疏信号的支撑集。
具体的， 至少两个接收机的稀疏信号组成以天线序号为列， 每个 天线在同一稀疏域域值上的信号为行的二维矩阵， 数据处理中' ^确定 的共享支撑集也是以天线序号为列， 每个天线在同一稀疏域域值上的 信号为行的矩阵， 数据处理中心通过共享支撑集可以按列获得各个接 收机的稀疏信号在共享支撑集中相对应的信号， 因此， 数据处理中心 根据共享支撑集与联合稀疏特性， 可以确定各个稀疏信号的支撑集。
示例性的， 如图 8 所示， 数据处理中心在确定 {A , X2 , 3 }的 共享支撑集 A的元素为 {4}之后， 数据处理中心利用每个接收机的所 有天线上的稀疏信号的支撑集相同的特性， 对每个接收机的所有天线 的观测数据进行共同处理。 例如， 数据处理中心把共享支撑集 Ωε中的 元素 {4}放入 2的支撑集集合， 等效于已经对第 2个接收机的观测数 据利用 OMP算法进行了 次迭代，数据处理中心根据预先设定的 2支 撑集的元素总个数 Τ2 ί= 3继续进行 Τ2 ί - Τ1 = 2次迭代， 得到 2支撑集 { Ω1 ? 2} 。
S 1 04、 数据处理中心根据共享支撑集和各个稀疏信号的支撑集， 确定各个稀疏信号。
本发明实施例中各个接收机的稀疏信号内部存在至少一个支撑 集， 至少两个接收机的稀疏信号之间存在共享支撑集， 数据处理中心 确定各个稀疏信号的支撑集后， 根据确定好的各个稀疏信号的支撑集 与共享支撑集， 即可确定各个稀疏信号。
例如， 如图 8 所示， 数据处理中心在确定 {A , X2 , 3 }的共享 支撑集为 Ωε , 稀疏信号 2的支撑集为 和 后， 数据处理中心按列 获得第 2 个接收机的稀疏信号在共享支撑集以及 和 2中相对应的 信号， 由于稀疏信号主要由这两部分组成， 数据处理中心采用现有的 压缩感知方法中信号恢复的算法可以得到稀疏信号 2。
综上所述，本发明实施例中的稀疏信号恢复过程利用了稀疏信号 之间与内部的联合稀疏特性， 可以大大降低各个接收机的采样率或观 测数据量。 各个接收机的观测数据为对稀疏信号采样得到的数据， 则 各个接收机获取的观测数据的数据量较小， 这样， 各个接收机将其观 测数据发送至数据处理中心， 降低了从接收机到数据处理中心的回程 负荷， 同时， 也降低了各个接收机的计算的处理复杂度， 各个接收机 的处理时延也相应降低。
本发明的实施例提供一种压缩感知方法，数据处理中心在接收到 至少一个接收机发送的观测数据后， 数据处理中心根据所有观测数据 和联合稀疏特性， 确定共享支撑集， 并在确定共享支撑集后， 数据处 理中心根据共享支撑集和联合稀疏特性， 确定各个稀疏信号的支撑 集， 最后， 数据处理中心根据共享支撑集和各个稀疏信号的支撑集， 确定各个稀疏信号， 其中， 观测数据为各个接收机根据该接收机的观 测矩阵对该接收机的稀疏信号采样得到的数据， 联合稀疏特性为所有 稀疏信号之间存在共享支撑集， 各个稀疏信号内部存在至少一个支撑 集， 支撑集为与各个稀疏信号对应的矩阵中非零行的索引集合， 非零 行为矩阵中行向量不为零的行。 通过该方案， 本发明利用稀疏信号内 部和之间的联合稀疏性， 解决了现有无线通信系统中捕获稀疏信号代 价较高的问题， 降低捕获稀疏信号的代价。
实施例六
本发明实施例提供一种压缩感知方法， 其应用场景如图 9所示， 整个无线通信系统中包含有 K K≥2、个接收机， 且 个不同的接收机 之间支持有回程连接， 使得各个接收机之间可以有信息交换。 其中， ^为第 ( l≤i≤K ) 个接收机观测到的稀疏信号， 为第 个接收机获 得关于其稀疏信号^的观测数据， 且 ζ ΦιΑ + Λ^ , Φ;为第 ζ·个接收机的 观测矩阵， 为第 ζ·个接收机的观测噪声， 在该无线系统中， 各个接 收机获得各自的观测数据， 并通过信息交互单独恢复各自的稀疏信 号。 如图 1 0所示， 该方法包括：
S 2 0 第一接收机获取其观测数据， 其中， 观测数据为第一接收 机根据其观测矩阵对第一接收机的稀疏信号采样得到数据。
本发明实施例中的执行主体为第一接收机， 其中， 第一接收机为 无线通信系统中的任意一个接收机， 本发明实施例不做限定。
本发明实施例中第一接收机可以为基站， 也可以为基站控制器。 由于本发明实施例提供的压缩感知方法适用于各个接收机获得 各自的观测数据， 并通过信息交互单独恢复各自的稀疏信号， 因此， 第一接收机先获取其观测数据。
具体的， 第一接收机获取其观测数据的过程与实施例一的 S 1 0 1 中至少两个接收机根据至少两个接收机的观测矩阵对至少两个接收 机的稀疏信号采样得到观测数据的过程完全相同， 不再赘述。
S 2 0 2、 第一接收机根据联合稀疏特性和观测数据， 确定共享支撑 集。
具体的， 第一接收机根据联合稀疏特性和其观测数据， 首先确定 第一接收机的稀疏信号的估计支撑集， 然后， 第一接收机将其估计支 撑集与至少两个接收机中除第一接收机以外的其他与第一接收机支 持回程连接的接收机的估计支撑集通过回程连接进行交换， 最后， 根 据联合稀疏特性确定共享支撑集。
其中，第一接收机将其估计支撑集与除第一接收机以外的其他与 第一接收机支撑回程连接的接收机的估计支撑集通过回程连接进行 交换可以是第一接收机将其估计支撑集与除第一接收机以外的各个 接收机的估计支撑集进行——交换， 也可以是第一接收机将其估计支 撑集与第一接收机以外的各个接收机的估计支撑集在信息中心处进 行交换， 本发明实施例并不做限定。
需要说明的是， 这里提到的信息中心可以是简单的数据交换中 心 ， 不具有数据处理能力， 且该信息中心的数据可以仅为各个接收机 的估计支撑集， 每个估计支撑集仅用一个比特 （ 0 或 1 ) 表示； 实施 例一中的数据处理中心可以是对全部观测数据进行处理， 每个观测数 据均为实数， 一般需要 12比特以上才能精确表示。
其中，第一接收机可以采用任何可行的信号恢复算法确定共享支 撑集， 本发明实施例不做限定。 示例性的， 信号恢复算法可以为 0MP ( Orthogonal Matching Pursuit, 正交匹 西己追踪算法）， 也可以为基 于 卜 1 范数的优化算法。
示例性的， 如图 8所示， 无线通信系统中包含有 3个接收机， 每 个接收机包含两个天线， 代表第 个接收机的稀疏信号， 是 7 行 2 列的矩阵， { ， X2 , }组成的信号矩阵中存在本发明实施例中提 到的联合稀疏特性。 其中， 第 1个接收机利用其所有天线的稀疏信号 的支撑集相同的特性， 通过 0MP算法对其所有天线的观测数据进行共 同处理。 0MP算法是迭代算法， 在每次迭代中可以从第 1个接收机的 其中一个天线的观测数据中找出与其相对应的稀疏信号的其中一个 支撑集的所在位置， 也就是说， 在每一次迭代中， 第 1个接收机从其 每个天线的观测数据中找出一个支撑集的元素 （即一个行序号）。 第 1个接收机经过一次迭代后把估计的支撑集的元素与第 2个接收机的 估计支撑集行序号和第 3 个接收机的估计支撑集行序号通过回程连 接进行交换。 由于这 3个接收机的稀疏信号包含有共享支撑集， 即 3 个接收机的稀疏信号中包含有相同的行序号， 所以第 1个接收机在得 到其他与第 1个接收机支持回程连接的接收机的估计支撑集后， 第 1 个接收机根据联合稀疏特性中包含有共享支撑集的特性选取重复次 数最多的元素 （在本示例中为 {4} ) 作为共享支撑集中的一个元素。 第 1个接收机根据预先设定的共享支撑集的元素的个数 T1 重复上述 步骤 次， 得到共享支撑集。
S 203、 第一接收机根据共享支撑集和联合稀疏特性， 确定第一接 收机的稀疏信号的支撑集。
第一接收机在确定共享支撑集后，根据该共享支撑集和联合稀疏 特性， 利用任何可行的信号恢复算法确定第一接收机的稀疏信号的支 撑集。
具体的， 所有稀疏信号组成以天线序号为列， 每个天线在同一稀 疏域域值上的信号为行的二维矩阵， 第一接收机确定的共享支撑集也 是以天线序号为列， 每个天线在同一域值上的信号为行的矩阵。 第一 接收机根据共享支撑集与联合稀疏特性， 可以确定其稀疏信号的支撑 集。 其中， 该方法与实施例一的 S 1 03 中数据处理中心确定各个稀疏 信号的支撑集的方法相同， 在此不再赘述。
S 204、第一接收机根据共享支撑集和第一接收机的稀疏信号的支 撑集， 确定其稀疏信号。
本发明实施例中各个接收机的稀疏信号内部存在至少一个支撑 集， 所有稀疏信号之间存在共享支撑集， 第一接收机确定其稀疏信号 的支撑集后， 根据确定好的稀疏信号的支撑集与共享支撑集， 确定其 稀疏信号。
综上所述，本发明实施例中的稀疏信号恢复过程利用了稀疏信号 之间与内部的联合稀疏特性， 可以大大降低了各个接收机的采样率或 观测数据量。 各个接收机分布式的恢复各个稀疏信号， 灵活度高和鲁 棒性好， 且各个接收机之间只需交换稀疏信号的估计支撑集信息， 而 没有其他信息， 所以各个接收机之间的回程负载较小。
本发明的实施例提供一种压缩感知方法，第一接收机在获取其观 测数据后， 根据联合稀疏特性和其观测数据， 确定共享支撑集， 然后， 第一接收机根据共享支撑集和联合稀疏特性， 确定其稀疏信号的支撑 集， 最后， 第一接收机根据共享支撑集和其稀疏信号的支撑集， 确定 第一接收机的稀疏信号， 其中， 观测数据为第一接收机根据第一接收 机的观测矩阵对其稀疏信号采样得到数据， 联合稀疏特性为所有稀疏 信号之间存在共享支撑集， 各个稀疏信号内部存在至少一个支撑集， 支撑集为与各个稀疏信号对应的矩阵中非零行的索引集合， 非零行为 矩阵中行向量不为零的行。 通过该方案， 本发明利用稀疏信号内部和 之间的联合稀疏性， 解决了现有无线通信系统中捕获高维稀疏信号代 价较高的问题， 降低捕获稀疏信号的代价。
本发明提供的压缩感知方法，涉及到的联合稀疏特性广泛适用于 无线通信场景， 该方法能够不论是将所有观测数据合在一起联合恢复 稀疏信号， 还是从每个接收机的观测数据单独恢复各自的稀疏信号， 即不管应用场景为实施例一的应用场景， 还是实施例二的应用场景， 在稀疏信号恢复过程中都要利用稀疏信号的联合稀疏特性先恢复出 所有稀疏信号的共享支撑集， 然后再恢复出各个稀疏信号的支撑集， 最后恢复出各个单独的稀疏信号， 整个稀疏信号的恢复过程应用联合 稀疏特性和压缩感知理论， 大大降低了稀疏信号的获取代价。
所属领域的技术人员可以清楚地了解到， 为描述的方便和简洁， 仅以上述各功能模块的划分进行举例说明， 实际应用中， 可以根据需 要而将上述功能分配由不同的功能模块完成， 即将装置的内部结构划 分成不同的功能模块， 以完成以上描述的全部或者部分功能。 上述描 述的装置的具体工作过程， 可以参考前述方法实施例中的对应过程， 在此不再赘述。
在本申请所提供的几个实施例中， 应该理解到， 所揭露的装置和 方法， 可以通过其它的方式实现。 例如， 以上所描述的装置实施例仅 仅是示意性的。
所述的单元可以是或者也可以不是物理上分开的，作为单元显示 的部件可以是一个物理单元或多个物理单元， 即可以位于一个地方， 或者也可以分布到多个不同地方。 可以根据实际的需要选择其中的部 分或者全部单元来实现本实施例方案的目的。
另外，在本发明各个实施例中的各功能单元可以集成在一个处理 单元中， 也可以是各个单元单独物理存在， 也可以两个或两个以上单 元集成在一个单元中。 上述集成的单元既可以采用硬件的形式实现， 也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的 产品销售或使用时， 可以存储在一个可读取存储介质中。 基于这样的 理解， 本发明的技术方案本质上或者说对现有技术做出贡献的部分或 者该技术方案的全部或部分可以以软件产品的形式体现出来， 该软件 产品存储在一个存储介质中， 包括若干指令用以使得一个设备（可以 是单片机， 芯片等） 或处理器 （ processor ) 执行本发明各个实施例 所述方法的全部或部分步骤。 而前述的存储介质包括： U盘、 移动硬 盘、 只读存储器 （ ROM, Read-Only Memory ), 随机存取存储器、 磁碟 或者光盘等各种可以存储程序代码的介质。
以上所述， 仅为本发明的具体实施方式， 但本发明的保护范围并 不局限于此， 任何熟悉本技术领域的技术人员在本发明揭露的技术范 围内， 可轻易想到变化或替换， 都应涵盖在本发明的保护范围之内。 因此， 本发明的保护范围应所述以权利要求的保护范围为准。

Claims

权 利 要 求 书
1、 一种数据处理中心， 其特征在于， 包括：
接收单元， 用于接收至少两个接收机发送的观测数据， 所述观测 数据为所述至少两个接收机根据所述至少两个接收机的观测矩阵对 所述至少两个接收机的稀疏信号采样得到的数据；
确定单元，用于根据所述接收单元接收到的观测数据和联合稀疏 特性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀疏信号 包含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在 至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机 的稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信号 的矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和所述联 合稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述共享支 撑集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。
2、 根据权利要求 1 所述的数据处理中心， 其特征在于， 所述联 合稀疏特性为各个接收机的稀疏信号包含有至少一个支撑集， 且所述 至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所 述共享支撑集为所述至少两个接收机的稀疏信号之间公共的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集合， 包 括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
3、 根据权利要求 1 所述的数据处理中心， 其特征在于， 所述至 少两个接收机根据以下公式对至少两个接收机的稀疏信号采样得到 观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,.为第 个接收机的观测矩阵， 为第 ζ'个接收机 的观测噪声， . e CMxiV为 Μ行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ^ ( ΤχΜ为 Γ行 Μ列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
4、 一种接收机， 其特征在于， 包括：
获取单元， 用于获取第一接收机的观测数据， 所述观测数据为所 述第一接收机根据所述第一接收机的观测矩阵对所述第一接收机的 稀疏信号采样得到数据；
确定单元，用于根据联合稀疏特性和所述获取单元获取到的观测 数据， 确定共享支撑集， 所述联合稀疏特性为至少两个接收机中各个 接收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的 稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所 述至少两个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个 接收机的稀疏信号的矩阵中非零行的索引集合， 所述第一接收机与所 述第二接收机之间支持回程连接， 以及用于根据所述共享支撑集和所 述联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑集， 以及用 于根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集， 确定 所述第一接收机的稀疏信号。
5、 根据权利要求 4 所述的接收机， 其特征在于， 所述联合稀疏 特性为至少两个接收机中各个接收机的稀疏信号包含有至少一个支 撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个共享支撑 集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共 享的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的 索引集合， 包括： 各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
6、 根据权利要求 5所述的接收机， 其特征在于，
所述确定单元， 具体用于根据所述联合稀疏特性和所述观测数 据， 确定所述第一接收机的稀疏信号的估计支撑集， 以及具体用于将 所述估计支撑集与第二接收机的估计支撑集交换， 并根据所述联合稀 疏特性确定所述共享支撑集， 所述第二接收机为所述至少两个接收机 中除所述第一接收机以外的其他与所述第一接收机支持回程连接的 接收机。
7、 根据权利要求 5 所述的接收机， 其特征在于， 所述第一接收 机根据以下公式对所述第一接收机的稀疏信号采样得到观测数据：
Υ = ΦίΧί + Νί i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
8、 一种数据处理中心， 其特征在于， 包括：
接收器， 用于接收至少两个接收机发送的观测数据， 所述观测数 据为所述至少两个接收机根据所述至少两个接收机的观测矩阵对所 述至少两个接收机的稀疏信号采样得到的数据； 处理器， 用于根据所述接收器接收到的观测数据和联合稀疏特 性， 确定共享支撑集， 所述联合稀疏特性为各个接收机的稀疏信号包 含有至少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在至 少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的 稀疏信号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信号的 矩阵中非零行的索引集合， 以及用于根据所述共享支撑集和所述联合 稀疏特性， 确定各个稀疏信号的支撑集， 以及用于根据所述共享支撑 集和所述各个稀疏信号的支撑集， 确定各个稀疏信号。
9、 根据权利要求 8 所述的数据处理中心， 其特征在于， 所述联 合稀疏特性为各个接收机的稀疏信号包含有至少一个支撑集， 且所述 至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所 述共享支撑集为所述至少两个接收机的稀疏信号之间公共的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集合， 包 括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
1 0、 根据权利要求 8所述的数据处理中心， 其特征在于， 所述至 少两个接收机根据以下公式对至少两个接收机的稀疏信号采样得到 观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ^ ( ΤχΜ为 Γ行 Μ列的矩阵， N^ CNxT为 N行 Γ列的矩阵。
1 1、 根据权利要求 8- 1 0任一项所述的数据处理中心， 其特征在 于，
所述数据处理中心包括中央处理器。
1 2、 一种接收机， 其特征在于， 包括：
处理器， 用于获取第一接收机的观测数据， 所述观测数据为所述 第一接收机根据所述第一接收机的观测矩阵对所述第一接收机的稀 疏信号采样得到数据， 以及用于根据联合稀疏特性和获取到的观测数 据， 确定共享支撑集， 所述联合稀疏特性为至少两个接收机中各个接 收机的稀疏信号包含有至少一个支撑集， 且所述至少两个接收机的稀 疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述 至少两个接收机的稀疏信号之间共享的支撑集， 所述支撑集为各个接 收机的稀疏信号的矩阵中非零行的索引集合， 所述第一接收机与所述 第二接收机之间支持回程连接， 以及用于根据所述共享支撑集和所述 联合稀疏特性， 确定所述第一接收机的稀疏信号的支撑集， 以及用于 根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集， 确定所 述第一接收机的稀疏信号。
1 3、 根据权利要求 1 2 所述的接收机， 其特征在于， 所述联合稀 疏特性为至少两个接收机中各个接收机的稀疏信号包含有至少一个 支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个共享支 撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间 共享的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行 的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵； 所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
1 4、 根据权利要求 1 3所述的接收机， 其特征在于，
所述处理器， 具体用于根据所述联合稀疏特性和所述观测数据， 确定所述第一接收机的稀疏信号的估计支撑集， 以及具体用于将所述 估计支撑集与第二接收机的估计支撑集交换， 并根据所述联合稀疏特 性确定所述共享支撑集， 所述第二接收机为所述至少两个接收机中除 所述第一接收机以外的其他与所述第一接收机支持回程连接的接收 机。
1 5、 根据权利要求 1 3所述的接收机， 其特征在于， 所述第一接 收机根据以下公式对所述第一接收机的稀疏信号采样得到观测数据：
= Φ, + N i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量，
Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
1 6、 根据权利要求 1 2 - 1 5任一项所述的接收机， 其特征在于， 所述接收机包括基站或者基站控制器。
1 7、 一种压缩感知方法， 其特征在于， 包括：
接收至少两个接收机发送的观测数据，所述观测数据为所述至少 两个接收机才艮据所述至少两个接收机的观测矩阵对所述至少两个接 收机的稀疏信号采样得到的数据；
根据接收到的观测数据和联合稀疏特性， 确定共享支撑集， 所述 联合稀疏特性为各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间共享的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合；
根据所述共享支撑集和所述联合稀疏特性，确定各个稀疏信号的 支撑集；
根据所述共享支撑集和所述各个稀疏信号的支撑集，确定各个稀 疏信号。
1 8、 根据权利要求 1 7 所述的压缩感知方法， 其特征在于， 所述 联合稀疏特性为各个接收机的稀疏信号包含有至少一个支撑集， 且所 述至少两个接收机的稀疏信号之间存在至少一个共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之间公共的支撑 集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零行的索引集 合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵；
所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
1 9、 根据权利要求 1 7 所述的压缩感知方法， 其特征在于， 所述 至少两个接收机根据以下公式对至少两个接收机的稀疏信号采样得 到观测数据：
γ. = Φ.χ. + N. ί = ι,- - -κ
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， Yt CTxN ^ Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量，
Γ行 M列的矩阵， Ni CNxT为 N行 Γ列的矩阵。
20、 一种压缩感知方法， 其特征在于， 包括：
获取第一接收机的观测数据，所述观测数据为所述第一接收机根 据所述第一接收机的观测矩阵对所述第一接收机的稀疏信号采样得 到数据；
根据联合稀疏特性和所述观测数据， 确定共享支撑集， 所述联合 稀疏特性为至少两个接收机中各个接收机的稀疏信号包含有至少一 个支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个共享 支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信号之 间共享的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中非零 行的索引集合， 所述第一接收机与所述第二接收机之间支持回程连 接；
根据所述共享支撑集和所述联合稀疏特性，确定所述第一接收机 的稀疏信号的支撑集；
根据所述共享支撑集和所述第一接收机的稀疏信号的支撑集，确 定所述第一接收机的稀疏信号。
21、 根据权利要求 20所述的压缩感知方法， 其特征在于， 所述 联合稀疏特性为至少两个接收机中各个接收机的稀疏信号包含有至 少一个支撑集， 且所述至少两个接收机的稀疏信号之间存在至少一个 共享支撑集， 其中， 所述共享支撑集为所述至少两个接收机的稀疏信 号之间共享的支撑集， 所述支撑集为各个接收机的稀疏信号的矩阵中 非零行的索引集合， 包括：
各个接收机包含至少一个天线，至少两个接收机的每个天线的稀 疏信号组成以天线序号为列， 每个天线在同一稀疏域域值上的信号为 行的二维矩阵； 所述至少两个接收机的稀疏信号在所述二维矩阵中存在至少一 个所述共享支撑集， 所述共享支撑集为所述至少两个接收机的稀疏信 号在所述二维矩阵中共享的非零行的索引集合；
所述各个接收机的稀疏信号在所述二维矩阵中存在至少一个所 述支撑集， 所述支撑集为所述各个接收机的稀疏信号在所述二维矩阵 中同一行上同时不为零的索引集合。
22、 根据权利要求 21 所述的压缩感知方法， 其特征在于， 所述 根据联合稀疏特性和所述观测数据， 确定共享支撑集， 具体包括： 根据所述联合稀疏特性和所述观测数据，确定所述第一接收机的 稀疏信号的估计支撑集；
将所述估计支撑集与第二接收机的估计支撑集交换，并根据所述 联合稀疏特性确定所述共享支撑集， 所述第二接收机为所述至少两个 接收机中除所述第一接收机以外的其他与所述第一接收机支持回程 连接的接收机。
2 3、 根据权利要求 20所述的压缩感知方法， 其特征在于， 所述 第一接收机根据以下公式对所述第一接收机的稀疏信号采样得到观 测数据：
Υ = ΦίΧί + Νί i = l,- - -K
其中， 为接收机的数量，且 ≥2 , 为第 个接收机的稀疏信号， 为 的观测数据， Φ,·为第 个接收机的观测矩阵， 为第 z'个接收机 的观测噪声， . e CMxiV为 M行 N列的矩阵， 且 N为第 个接收机包含的 天线数量， N≥l , M为所述二维矩阵的行的数量， C代表复数集， 为 Γ行 N列的矩阵， Γ为第 个接收机的观测信号的数量， Φ^ ( ΤχΜ为 Γ行 Μ列的矩阵， N^ CNxT为 N行 Γ列的矩阵。
PCT/CN2014/076688 2014-04-30 2014-04-30 一种压缩感知方法及装置 WO2015165118A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP14891091.2A EP3133850A4 (en) 2014-04-30 2014-04-30 Compressed sensing method and device
PCT/CN2014/076688 WO2015165118A1 (zh) 2014-04-30 2014-04-30 一种压缩感知方法及装置
CN201480078643.6A CN106256141A (zh) 2014-04-30 2014-04-30 一种压缩感知方法及装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2014/076688 WO2015165118A1 (zh) 2014-04-30 2014-04-30 一种压缩感知方法及装置

Publications (1)

Publication Number Publication Date
WO2015165118A1 true WO2015165118A1 (zh) 2015-11-05

Family

ID=54358069

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/076688 WO2015165118A1 (zh) 2014-04-30 2014-04-30 一种压缩感知方法及装置

Country Status (3)

Country Link
EP (1) EP3133850A4 (zh)
CN (1) CN106256141A (zh)
WO (1) WO2015165118A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102045118A (zh) * 2010-10-22 2011-05-04 清华大学 基于伪逆相乘的压缩感知重建算法
CN102970044A (zh) * 2012-11-23 2013-03-13 南开大学 一种基于回溯的迭代重加权压缩传感重构算法
CN103476040A (zh) * 2013-09-24 2013-12-25 重庆邮电大学 带有隐私保护的分布式压缩感知数据融合方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120259590A1 (en) * 2011-04-11 2012-10-11 Jong Chul Ye Method and apparatus for compressed sensing with joint sparsity
CN102882530B (zh) * 2012-09-17 2015-04-08 南京邮电大学 一种压缩感知信号重构方法
CN103178853B (zh) * 2013-03-21 2016-04-27 哈尔滨工业大学 基于压缩感知的稀疏信号欠采样方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102045118A (zh) * 2010-10-22 2011-05-04 清华大学 基于伪逆相乘的压缩感知重建算法
CN102970044A (zh) * 2012-11-23 2013-03-13 南开大学 一种基于回溯的迭代重加权压缩传感重构算法
CN103476040A (zh) * 2013-09-24 2013-12-25 重庆邮电大学 带有隐私保护的分布式压缩感知数据融合方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3133850A4 *

Also Published As

Publication number Publication date
CN106256141A (zh) 2016-12-21
EP3133850A1 (en) 2017-02-22
EP3133850A4 (en) 2017-06-28

Similar Documents

Publication Publication Date Title
US9654264B2 (en) Beam forming using a dual polarized antenna arrangement
JP6108250B2 (ja) チャネル状態情報を報告および受信する方法およびデバイス
WO2015101109A1 (zh) 一种信道状态信息测量、参考信号的发送方法和装置
RU2007103348A (ru) Эффективное вычисление матриц пространственного фильтра для управления разнесением на передаче в системе связи mimo
WO2018171604A1 (zh) 信息的传输方法和设备
JP6473207B2 (ja) 通信方法及びデバイス
CN107852387B (zh) 降低大规模多输入多输出系统中预编码矩阵计算和用户设备分组复杂度的方法
CN110034804B (zh) 对无线通信系统估计角度信息的方法和装置
US10466345B1 (en) Time-of-arrival estimation with subspace methods
US20180013590A1 (en) Transforming and combining signals from antenna array
WO2018107664A1 (zh) 一种干扰抑制方法、装置及计算机存储介质
WO2017076371A1 (zh) 一种波束赋形的方法及装置
CN105024957B (zh) 一种直扩码分多址上行链路信道估计方法及装置
WO2022001240A1 (zh) 一种检测方法、装置和通信设备
CN106233783A (zh) 信道测量方法、信道测量装置、用户设备及系统
WO2017107697A1 (zh) 下行信道重构方法以及装置
WO2017080359A1 (zh) 一种干扰消除方法、装置及基站
WO2015165118A1 (zh) 一种压缩感知方法及装置
WO2012152020A1 (zh) 一种信道均衡方法、基站和系统
WO2017166200A1 (zh) 一种数据传输方法和装置
WO2024000179A1 (zh) 一种上行mimo传输的天线全相干传输码字的确定方法及其装置
WO2023240654A1 (zh) 一种部分天线相干传输码字的确定方法及其装置
WO2024021129A1 (zh) 上行mimo传输8天线端口多天线面板的码字确定方法及其装置
WO2014153681A1 (zh) Mimo系统中的多天线信道码本反馈方法及装置
WO2023273609A1 (zh) 一种无源互调源数目确定方法及相关设备

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: 14891091

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

REEP Request for entry into the european phase

Ref document number: 2014891091

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2014891091

Country of ref document: EP