WO2020170267A1 - Procédé et système de compression de signaux reçus au niveau d'une station de base - Google Patents

Procédé et système de compression de signaux reçus au niveau d'une station de base Download PDF

Info

Publication number
WO2020170267A1
WO2020170267A1 PCT/IN2020/050102 IN2020050102W WO2020170267A1 WO 2020170267 A1 WO2020170267 A1 WO 2020170267A1 IN 2020050102 W IN2020050102 W IN 2020050102W WO 2020170267 A1 WO2020170267 A1 WO 2020170267A1
Authority
WO
WIPO (PCT)
Prior art keywords
signals
matrix
basis
time
antenna
Prior art date
Application number
PCT/IN2020/050102
Other languages
English (en)
Inventor
Aswathylakshmi P
RadhaKrishna GANTI
Original Assignee
INDIAN INSTITUTE OF TECHNOLOGY MADRAS (IIT Madras)
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 INDIAN INSTITUTE OF TECHNOLOGY MADRAS (IIT Madras) filed Critical INDIAN INSTITUTE OF TECHNOLOGY MADRAS (IIT Madras)
Publication of WO2020170267A1 publication Critical patent/WO2020170267A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/022—Site diversity; Macro-diversity
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W88/00—Devices specially adapted for wireless communication networks, e.g. terminals, base stations or access point devices
    • H04W88/08—Access point devices
    • H04W88/085—Access point devices with remote components

Definitions

  • the embodiments herein relate to wireless communication. More particularly relates to a method and system for compressing signals received at a base station (BS).
  • BS base station
  • Massive multiple-input and multiple-output (MEMO) base station which comprises a large number of antennas supports a plurality of users simultaneously through spatial multiplexing. This improves spectral efficiency and increases the network capacity.
  • C-RAN centralized radio access network
  • the base station is split into two parts: a pooled baseband unit (BBU) at a centralized location common to several cells, and a number of remote radio heads (RRH) distributed geographically over these cells, connected to the central BBU (as shown in FIG. 3).
  • BBU pooled baseband unit
  • RRH remote radio heads
  • the pooling of baseband resources can meet the processing requirements of the massive MEMO systems as well as offer the potential for cooperative radio to reduce interference. Furthermore, network operators can drive down the cost through the concentration of resources at the BBU and the deployment of limited-functionality RRHs in the cells.
  • the twin advantages of reduced cost and interference in C-RAN become conducive to network densification, a key driver for 5G, by allowing a higher density of RRHs to be put in place.
  • massive MIMO combined with C-RAN can potentially support the ultra-high data rates envisioned in 5G.
  • the tight latency constraints and large bandwidths of the 5G require high-speed data transfer in the links between the BBU and the RRH, called the fronthaul.
  • the principal object of the embodiments herein is to provide a method and system for compressing signals received at a base station.
  • Another object of the embodiments herein is to receive a plurality of signals using an array of antenna in the base station over a time span.
  • Another object of the embodiments herein is to compress the plurality of signals by selecting basis signals and representing the plurality of signals in terms of the basis signals.
  • Another object of the embodiments herein is to transmit the basis signal and the plurality of signals represented in terms of the basis signals for each of the received signals of the plurality of signals to a distributed unit (150).
  • the embodiments herein provide a method for compressing signals received at a base station (100).
  • the method includes receiving, by a communicator (126) of the base station (100), a plurality of signals using an array of antenna in the base station (100) over a time span and compressing, by a radio unit (120) of the BS (100), the plurality of signals by selecting basis signals and representing the plurality of signals in terms of the basis signals, where the basis signals are a set of signals of the plurality of signals which are received at a set of antennas of the array of antenna.
  • the method includes transmitting, by the RU (120) of the BS (100), the basis signal and the plurality of signals represented in terms of the basis signals for each of the received signals of the plurality of signals to a distributed unit (140).
  • the compression is obtained by performing a QR decomposing technique.
  • the method of compressing, by the RU (120) of the BS (100), the plurality of signals by selecting the basis signals and representing the plurality of signals in terms of the basis signals includes constructing, by the RU (120) of the BS (100), a time-domain signal matrix based on the plurality of signals received at the array of antenna, wherein the time-domain signal matrix comprises a block of time-domain samples of the signals received at each antenna of the array of antenna and generating, by the RU (120) of the BS(100), a frequency domain signal matrix using the time- domain signal matrix.
  • the method also includes performing, by the RU (120) of the BS (100), resource element de-mapping by dividing the frequency domain signal matrix into sub-matrices corresponding to different users; and compressing, by the RU (120) of the BS (100), each of the sub-matrices of each of the users to obtain component matrixes for each of the sub-matrices of each of the user, wherein the component matrixes comprises the basis signals and the plurality of signals in terms of the basis signals.
  • the method also includes receiving, by the DU (140) of the BS (100), a basis matrix and a projection matrix for each of the sub-matrices of each of the users; and reconstructing, by the DU (140) of the BS (100), each of the sub-matrices by determining a function of the basis matrix and the projection matrix of each of the sub-matrices.
  • the method of compressing, by the RU (120) of the BS (100), the plurality of signals by selecting the basis signals and representing the plurality of signals in terms of the basis signals includes constructing, by the RU (120) of the BS (100), a time-domain signal matrix based on the plurality of signals received at the array of antenna, wherein the time-domain signal matrix comprises a block of time-domain samples of the received signal at each antenna of the array of antenna; and compressing, by the RU (120) of the BS (100), the time-domain signal matrix to obtain component matrixes, wherein the component matrixes comprises the basis signals and the plurality of signals in terms of the basis signals.
  • the method also includes receiving, by the DU (140) of the BS (100), a basis matrix and a projection matrix for the time domain matrix and reconstructing, by the DU (140) of the BS (100), the time domain matrix by determining a function of the basis matrix and the projection matrix for the time domain matrix.
  • the two component matrixes comprise a basis matrix and a projection matrix.
  • the component matrixes represents an approximation of contents of the time-domain signal matrix with fewer number of time-domain samples as compared to the block of the time-domain samples of the time-domain signal matrix which provides the compression.
  • the entries in columns of the time-domain signal matrix represent the signals received at each of the antenna of the array of antenna over a time span of N samples.
  • the basis matrix or a function of the basis matrix; and the projection matrix or a function of the projection matrix is transmitted through a fronthaul link (1000) of the BS (100).
  • the frequency domain matrix is divided into the sub-matrices according to allocated sub-carriers to each of the users.
  • the method of constructing, by the base station (100), the time-domain signal matrix based on the plurality of signals received at the plurality of antenna includes down-converting, by the RU (120) of the base station (100), the plurality of signals received at the plurality of antenna; and generating, by the RU (120) of the base station (100), the time- domain signal matrix based using the plurality of down-converted signals.
  • the basis matrix comprises signals received at a subset of antennas or a function of the signals received at a subset of antennas from the array of antenna of the RU (120) of the BS (100).
  • the subset of antennas comprises antennas with highest received powers or antennas with a function of the highest received powers to represent the plurality of signals received at the plurality of antennas.
  • the projection matrix comprises description of the plurality of signals received at the plurality of antenna in terms of the basis matrix.
  • the sub-carrier allocation of each of the users is known to the BS (100).
  • the dimensions of the time-domain signal matrix is dependent on a total number of sub carriers allocated to all users.
  • the method of transmitting, by the RU (120) of the BS (100), the basis signal and the plurality of signals represented in terms of the basis signal for each of the received signals of the plurality of signals to the DU (140) includes performing, by the RU (120) of the BS (100), a quantization mechanism on the basis signal and the plurality of signals represented in terms of the basis signal for each of the received signals of the plurality of signals; and transmitting, by the RU (120) of the BS (100), the quantized basis signals and the quantized signals of the plurality of signals represented in terms of the basis signal for each of the received signals of the plurality of signals to the DU (140).
  • the embodiments herein provide a base station (100) for compressing signals received at the base station (100).
  • the base station (100) includes a Radio unit (RU) (120) connected to a distributed unit (DU) (140) through a fronthaul link (1000).
  • the RU (120) includes a memory (122), a processor (124), and a communicator (126).
  • the RU (120) is configured to receive a plurality of signals using an array of antenna in the base station (100) over a time span and compress the plurality of signals by selecting basis signals and representing the plurality of signals in terms of the basis signals, wherein the basis signals are a set of signals of the plurality of signals which are received at a set of antennas of the array of antenna.
  • the RU (120) is also configured to transmit the basis signal and the plurality of signals represented in terms of the basis signals for each of the received signals of the plurality of signals to the distributed unit (DU) (140) through the fronthaul link (1000).
  • the base station (100) includes a distributed unit (DU) (140) connected to a Radio unit (RU) (120) through a fronthaul link (1000).
  • the DU (140) includes a memory (142), a processor (144) and a communicator (146).
  • the DU (140) is configured to receive a basis matrix and a projection matrix for a time domain matrix from the RU (120) through the fronthaul link (1000) and reconstruct the time domain matrix by determining a function of the basis matrix and the projection matrix for the time domain matrix.
  • the base station (100) includes a distributed unit (DU) (140) connected to a Radio unit (RU) (120) through a fronthaul link (1000).
  • the DU (140) includes a memory (142), a processor (144) and a communicator (146).
  • the DU (140) is configured to receive a basis matrix and a projection matrix for each sub-matrix corresponding to each users, from the Radio unit (RU) (120) through the fronthaul link (1000); and reconstruct each of the sub- matrices corresponding to each of the users by determining a function of the basis matrix and the projection matrix for the time domain matrix.
  • FIG. 1 is a block diagram of a system for compressing signals received at a base station (100), according to the embodiments as disclosed herein;
  • FIG. 2A is a flow chart 200a illustrating the method for compressing the signals received at the BS (100), according to the embodiments as disclosed herein;
  • FIG. 2B is another flow chart 200b illustrating the method for compressing the signals received at the BS (100), according to the embodiments as disclosed herein;
  • FIG. 3 is an example Massive MIMO C-RAN architecture which provides compression of signals for transmission in a fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein;
  • FIG. 4 illustrates proposed functional splits between a remote radio head (RRH) (120) and a baseband unit (BBU) (140) in an uplink in a massive MIMO system, according to the embodiments as disclosed herein;
  • RRH remote radio head
  • BBU baseband unit
  • FIG. 5 illustrates the compression of the signals received the BS (100) at various stages before transmission in the fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein;
  • FIG. 6 illustrates a method of approximating each of user sub- matrixes as a function of component matrices before transmission in the fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein;
  • FIG. 7 is a graph plot illustrating compression ratios (CRs) for proposed method of compression as a function of the actual number of antennas (Lu) for signal subspace of UEs u, according to the embodiments as disclosed herein;
  • FIG. 8A is a graph illustrating Bit Error Rate (BER) performance for QR compression for 8 users, according to the embodiments as disclosed herein;
  • FIG. 8B is a graph illustrating Bit Error Rate (BER) performance for QR compression for 12 users, according to the embodiments as disclosed herein;
  • FIGS. 9A-9B are graph plots illustrating un-coded Bit Error Rate (BER) performance for the proposed method in comparison with the SVD compression, according to the embodiments as disclosed herein.
  • BER Bit Error Rate
  • circuits constituting a block may be implemented by dedicated hardware, or by a processor (e.g., one or more programmed microprocessors and associated circuitry), or by a combination of dedicated hardware to perform some functions of the block and a processor to perform other functions of the block.
  • a processor e.g., one or more programmed microprocessors and associated circuitry
  • Each block of the embodiments may be physically separated into two or more interacting and discrete blocks without departing from the scope of the disclosure.
  • the blocks of the embodiments may be physically combined into more complex blocks without departing from the scope of the disclosure.
  • radio base station (100) and base station (100) refers to the same and may be used inter-changeably throughout the specification.
  • the terms radio head (RRH) (120) and Radio Unit (RU) (120) refers to the same and may be used inter-changeably throughout the specification, depending on the technology referred in the specification.
  • the terms baseband unit (BBU) (140) and Distributed Unit (DU) (140) refers to the same and may be used inter- changeably throughout the specification, depending on the technology referred in the specification.
  • the embodiments herein provide a method for compressing signals received at a base station (100).
  • the method includes receiving, by a communicator (120) of the base station (100), a plurality of signals using an array of antenna in the base station (100) over a time span and compressing, by a radio unit (120) of the BS (100), the plurality of signals by selecting basis signals and representing the plurality of signals in terms of the basis signals, where the basis signals are a set of signals of the plurality of signals which are received at a set of antennas of the array of antenna.
  • the method includes transmitting, by the RU (120) of the BS (100), the basis signal and the plurality of signals represented in terms of the basis signals for each of the received signals of the plurality of signals to a distributed unit (150).
  • FIG. 1 is a block diagram of system for compressing signals received at a base station (100), according to the embodiments as disclosed herein.
  • a radio base station (100) or a base station (100) has a remote radio head (RRH) (120) that performs radio frequency and low-PHY layer processing, and a baseband unit (BBU) (140) that performs high-PHY, MAC and RLC layer processing.
  • Fronthaul (1000) is a data transport link that connects the RRH (120) and BBU (140).
  • the RRH (120) is called Radio Unit (RU) (120)
  • the BBU (140) is called Distributed Unit (DU) (140).
  • a base station (100) that supports massive Multiple Input Multiple Output (MIMO) has the RU (120) equipped with antenna arrays that have a large number of antennas for transmission/reception so that many users can be supported simultaneously. Having large number of antennas increases the amount of data that has to be transported in the fronthaul link (1000), whose capacity is limited. Therefore, compression techniques are required to reduce the data load in the fronthaul (1000).
  • MIMO Multiple Input Multiple Output
  • the base station (100) includes a radio unit (RU) (120) and a distributed unit (DU) (140) connected through a fronthaul link (1000).
  • RU radio unit
  • DU distributed unit
  • the RU (120) includes a memory (122), a processor (124), a communicator (124), a down conversion management engine (128), a component matrix generation engine (130) and a resource element (RE) demapping engine (132).
  • the RU (120) performs radio frequency and low- PHY layer processing.
  • the memory (122) can include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.
  • the memory (122) may, in some examples, be considered a non-transitory storage medium.
  • the term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term“non-transitory” should not be interpreted that the memory (122) is non-movable.
  • the memory (122) is configured to store larger amounts of information than the memory.
  • a non-transitory storage medium may store data that can, over time, change (e.g ., in Random Access Memory (RAM) or cache).
  • RAM Random Access Memory
  • the processor (124) is configured to coordinate the functions of the hardware elements of the base station (100).
  • the communicator (126) is configured to receive a plurality of signals using an array of antenna in the base station (100) over a time span.
  • the fronthaul link (1000) is configured to transmit the basis signal and the plurality of signals represented in terms of the basis signals to the DU (140).
  • the down conversion management engine (128) is configured to down-convert the plurality of signals received at the plurality of antennas of the base station (100).
  • the component matrix generation engine (130) is configured to generate the time-domain signal matrix using the plurality of down-converted signals.
  • the time-domain signal matrix includes a block of time-domain samples of the signals received at each antenna of the array of antenna.
  • the component matrix generation engine (130) is configured to generate a frequency domain signal matrix using the time-domain signal matrix and compress each of the sub-matrices of each of the users to obtain component matrixes for each of the sub-matrices of each of the user.
  • the component matrixes include the basis signals and the plurality of signals in terms of the basis signals.
  • the component matrix generation engine (130) is configured to compress each of the sub-matrices of each of the users to obtain component matrixes for each of the sub-matrices of each of the user.
  • the component matrixes comprise the basis signals and the plurality of signals in terms of the basis signals.
  • the frequency domain matrix is divided into the sub- matrices according to allocated sub-carriers to each of the users.
  • the component matrix generation engine (130) is configured to generate the time-domain signal matrix based on the plurality of signals received at the array of antenna and compress the time- domain signal matrix to obtain the component matrixes.
  • the time-domain signal matrix includes a block of time-domain samples of the received signal at each antenna of the array of antenna and the dimensions of the time-domain signal matrix is dependent on a total number of sub carriers allocated to all users.
  • the component matrix generation engine (130) is configured to compress the time-domain signal matrix to obtain component matrixes, wherein the component matrixes comprise the basis signals and the plurality of signals in terms of the basis signals.
  • the compression is obtained by performing a decomposition technique such as for example but not limited to: a QR decomposing technique, singular value decomposition (SVD) compression technique and the like.
  • the two component matrix includes a basis matrix and a projection matrix.
  • the component matrixes represents an approximation of the contents of the time-domain signal matrix with fewer number of time-domain samples as compared to the block of the time-domain samples of the time- domain signal matrix which provides the compression.
  • the entries in the columns of the time-domain signal matrix represent the signals received at each of the antenna of the array of antenna over a time span of N samples.
  • the basis matrix or a function of the basis matrix and the projection matrix or a function of the projection matrix of each of the user is transmitted through the fronthaul link (1000) of the BS (100).
  • the basis matrix includes the signals received at a subset of antennas or a function of the signals received at a subset of antennas from the array of antenna of the RU (120) of the BS (100).
  • the subset of antennas includes the antennas with the highest received powers or antennas with a function of the highest received powers to represent the plurality of signals received at the plurality of antennas.
  • the projection matrix encompasses the description of the plurality of signals received at the plurality of antenna in terms of the basis matrix.
  • the RE demapping engine (132) is configured to perform the resource element de-mapping by dividing the frequency domain signal matrix into the sub-matrices corresponding to different users.
  • the distributed unit (DU) (150) includes a memory (142), a processor (144), a communicator (146) and a reconstruction management engine (148).
  • the DU (140) performs high-PHY, MAC and RLC layer processing.
  • the memory (142) can include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories.
  • the memory (142) may, in some examples, be considered a non-transitory storage medium.
  • the term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term“non-transitory” should not be interpreted that the memory (142) is non-movable.
  • the memory (142) is configured to store larger amounts of information than the memory.
  • a non-transitory storage medium may store data that can, over time, change (e.g ., in Random Access Memory (RAM) or cache).
  • the processor (144) is configured to coordinate the functions of the hardware elements of the base station (100).
  • the communicator (146) is configured to receive the basis matrix and the projection matrix from the DU (140) through the fronthaul link (1000).
  • the reconstruction management engine (148) is configured to receive a basis matrix and a projection matrix for the time domain matrix and reconstruct the time domain matrix by determining a function of the basis matrix and the projection matrix for the time domain matrix.
  • the reconstruction management engine (148) is configured to receive a basis matrix and a projection matrix for each of the sub-matrices of each of the users and reconstruct each of the sub-matrices by determining a function of the basis matrix and the projection matrix of each of the sub-matrices.
  • FIG. 1 shows the hardware elements of the base station (100) but it is to be understood that other embodiments are not limited thereon.
  • the base station (100) may include less or more number of elements.
  • the labels or names of the elements are used only for illustrative purpose and does not limit the scope of the invention.
  • One or more components can be combined together to perform same or substantially similar function.
  • FIG. 2A is a flow chart 200a illustrating the method for compressing the signals received at the BS (100), according to the embodiments as disclosed herein.
  • the BS (100) receive the plurality of signals using the array of antenna in the base station (100) over the time span.
  • the communicator (126) of the RU (120) is configured to receive the plurality of signals using the array of antenna in the base station (100) over the time span.
  • the BS (100) constructs the time-domain signal matrix based on the plurality of signals received at the array of antenna.
  • the component matrix generation engine (130) is configured to constructs the time-domain signal matrix based on the plurality of signals received at the array of antenna.
  • the BS (100) generates the frequency domain signal matrix using the time-domain signal matrix.
  • the component matrix generation engine (130) is configured to generate the frequency domain signal matrix using the time-domain signal matrix.
  • the BS (100) performs the resource element de- mapping by dividing the frequency domain signal matrix into the sub-matrices corresponding to different users.
  • the RE demapping engine (132) is configured to perform the resource element de-mapping by dividing the frequency domain signal matrix into the sub-matrices corresponding to different users.
  • the BS (100) compresses each of the sub-matrices of each of the users to obtain the component matrixes for each of the sub- matrices of each of the user.
  • the component matrix generation engine (130) is configured to compress each of the sub-matrices of each of the users to obtain the component matrixes for each of the sub-matrices of each of the user.
  • the BS (100) transmits the component matrixes for each of the sub-matrices of each of the user.
  • the fronthaul link (1000) is configured to transmit the component matrixes for each of the sub-matrices of each of the user.
  • the BS (100) receives the basis matrix and the projection matrix for each of the sub-matrices of each of the users.
  • the communicator (146) at the DU (140) is configured to receive the basis matrix and the projection matrix for each of the sub-matrices of each of the users.
  • the BS (100) reconstructs each of the sub-matrices by determining the function of the basis matrix and the projection matrix of each of the sub-matrices.
  • the reconstruction management engine (148) is configured to reconstruct each of the sub-matrices by determining the function of the basis matrix and the projection matrix of each of the sub-matrices.
  • FIG. 2B is another flow chart 200b illustrating the method for compressing the signals received at the BS (100), according to the embodiments as disclosed herein.
  • the BS (100) receive the plurality of signals using the array of antenna in the base station (100) over the time span.
  • the communicator (126) of the RU (120) is configured to receive the plurality of signals using the array of antenna in the base station (100) over the time span.
  • the BS (100) constructs the time-domain signal matrix based on the plurality of signals received at the array of antenna.
  • the component matrix generation engine (130) is configured to construct the time-domain signal matrix based on the plurality of signals received at the array of antenna.
  • the BS (100) compresses the time-domain signal matrix to obtain the component matrixes.
  • the component matrix generation engine (130) is configured to compress the time-domain signal matrix to obtain the component matrixes.
  • the BS (100) transmits the component matrixes which comprise the basis matrix and the projection matrix for the time-domain signal matrix.
  • the fronthaul link (1000) is configured to transmit the component matrixes which comprise the basis matrix and the projection matrix for the time-domain signal matrix.
  • the BS (100) receives the basis matrix and the projection matrix for the time domain matrix from the RU (140) through the fronthaul link (1000).
  • the communicator (146) at the DU (140) is configured to receive the basis matrix and the projection matrix for the time domain matrix from the RU (140) through the fronthaul link (1000).
  • the BS (100) reconstructs the time domain matrix by determining function of the basis matrix and the projection matrix for the time domain matrix.
  • the reconstruction management engine (148) is configured to reconstruct the time domain matrix by determining function of the basis matrix and the projection matrix for the time domain matrix.
  • FIG. 3 is an example Massive MIMO C-RAN architecture which provides the compression of the signals for transmission in the fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein.
  • a massive MIMO base station comprises a large number of antennas which has the ability to support many users simultaneously through spatial multiplexing, which improves spectral efficiency and increases the network capacity.
  • the processing complexity that such a system requires makes centralized radio access network (C-RAN) a more suitable architecture for implementation.
  • C-RAN the base station (100) is split into two parts: a pooled baseband unit (BBU) (140) at a centralized location common to several cells; and a number of remote radio heads (RRH) (120) distributed geographically over the cells and connected to the central BBU (140).
  • BBU pooled baseband unit
  • RRH remote radio heads
  • the pooling of the baseband resources can meet the processing requirements of the massive MIMO systems as well as offer the potential for cooperative radio to reduce interference.
  • network operators reduce the cost of implementation due to the concentration of the resources at the BBU (140) and deployment of limited-functionality RRHs in the cells.
  • the massive MIMO combined with the C-RAN can potentially support the ultra-high data rates envisioned in the 5G.
  • the tight latency constraints and large bandwidths of the 5G require high-speed data transfer in the links between the BBU (140) and the RRH (120) called the fronthaul link (1000).
  • the fronthaul capacity demand scales with the number of antennas at the RRH (120) and laying such high capacity optical fibers for each of the antenna stream in the BBU-RRH link would drive up the cost for the network operators.
  • the low-PHY functional split between the BBU (140) and RRH (120) requires a data rate up to 236 Gbps for a 100MHz bandwidth.
  • mmWave millimeter-wave
  • THz terahertz
  • the conventional methods used for uplink fronthaul compression are for example: point-to-point (P2P) compression, distributed source coding, compressed sensing (CS) and spatial filtering.
  • P2P point-to-point
  • CS compressed sensing
  • spatial filtering a key drawback of the approaches is a need to divide the antenna array into many groups and then apply the processing separately.
  • the set of antennas that are not selected either remain inactive or the data received at the antennas are discarded.
  • the Principal Component Analysis (PCA) compression algorithm uses the inherent sparsity of the MIMO channels to reduce the number of links required in the fronthaul.
  • the PCA compression performs a low-rank approximation of the matrix consisting of the received signals by leveraging the signal correlation across space and time.
  • the PCA compression requires computing the singular value decomposition (SVD) of the matrix wherein the complexity is prohibitively high for large matrix dimensions as in the massive MIMO case, since the RRH (120) has limited processing resources. Therefore, the above mentioned issues are addressed by the proposed method which uses QR decomposition technique which sends the received signal matrix as the component matrix of the basis matrix and the projection matrix.
  • the projection matrix comprises description of the plurality of signals received at the plurality of antenna in terms of the basis matrix.
  • FIG. 4 illustrates the proposed functional splits between the BBU (140) and the RRH (120) in the uplink in the massive MIMO system, according to the embodiments as disclosed herein.
  • the data rate in the fronthaul link (1000) is impacted by the functional split between the BBU (140) and the RRH (120).
  • the data rate is almost halved when moving from a split A to a split B, as the cyclic prefix (CP) and the guard bands are removed.
  • the resource elements (RE) are demapped and the users are separated which results in the reduction in the data rate which is also dependent on the resource block utilization.
  • the split D depends on the modulation order where for large modulation orders, the split D can increase the data rate, as more bits are required to represent each of the samples.
  • the uplink fronthaul data rate is reduced in two stages with a split at C as shown in the FIG. 4 where the CP and the guard band are removed and the RE demapping is completed at the RRH (120). Also, at the split C, the users are separated and a low complexity algorithm based on the QR decomposition is used to approximate the matrix composed of the complex baseband received signals.
  • the advantages of the proposed compression method includes a user-specific compression technique; provides a good approximation of the received signal matrix; information from the antennas that are not selected are not lost while achieving high compression ratios; and provides denoising gain leading to better error performance compared to conventional uncompressed system.
  • FIG. 5 illustrates the compression of the signals received the BS (100) at various stages before transmission in the fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein.
  • FIG. 6 illustrates the method of approximating each of the user sub-matrixes as the function of the component matrices before transmission in the fronthaul link (1000) of the BS (100), according to the embodiments as disclosed herein.
  • the uplink multiple access scheme is Orthogonal frequency-division multiple access (OFDMA). Therefore, the bit-stream from each user undergoes the M- QAM symbol mapping followed by the OFDM modulation.
  • the OFDM modulation consists of the sub-carrier mapping according to the resources allocated to the user, IFFT, and addition of a cyclic prefix (CP).
  • the OFDM symbols reach the RRH (120) of the base station (100) through multi-path channels.
  • the received signal at antenna r at sampling instant n is
  • xu is the OFDM symbol from the user u
  • hr,u is the multipath channel response from the user u to antenna represents the
  • wr is the additive white Gaussian noise (AWGN) with variance s 2 at antenna r.
  • AWGN additive white Gaussian noise
  • Each column of the received signal matrix Y represents the signal received at each antenna over a time span of N samples. Assuming a maximum of L multi- paths for each user, equation (1) is expanded to get:
  • W is the N r X N matrix of the complex AWGN at the RRH (120).
  • the received signal matrix Y needs to be sent from the RRH (120) to the BBU (140) via the fronthaul link (1000). Since the dimension of the Y, N XN r , is large in the massive MIMO system the dimension of the Y needs to be reduced to achieve compression.
  • low rank approximation is a tool used in signal processing to represent a matrix of large dimension using a lower dimensional subspace for analysis while retaining the essential information contained in the original matrix. From (2), we observe that the received signal matrix in the absence of noise would simply be Therefore, the true rank of the
  • Y is the rank of the noiseless matrix and is atmost N U L.
  • the simplest way to reduce the dimension of Y is to select the N u L columns of Y that have the largest vector norms, which correspond to the antennas with the highest received powers. However, if this selection criterion is applied in the time-domain, where the users are not separated, the set of antennas chosen will be common to all users. Since we choose the antennas based on their total received powers, the users nearer to the RRH (120) that contribute more power, will be favored over the users farther from the RRH (120).
  • the proposed method avoids the users nearer to the RRH (120) that contribute more power from being favored over the users farther from the RRH (120) by separating the users as shown in FIG. 4 before reducing the dimension of the matrix Y.
  • the matrix Y is converted to the frequency domain by applying FFT so that the users can be separated according to the sub-carriers allocated to the users.
  • the sub-carrier allocation of each user is known to the base station (100).
  • the matrix Y can be decomposed in the frequency domain into the sub-matrices corresponding to each of the users and then the proposed method selects the best antennas in each of the sub-matrices.
  • Y f is the frequency domain received signal matrix obtained from Y, then the true rank (Y f ) £ N U L.
  • rank of Y f is the rank of Now, the DFT matrix, F is unitary
  • the proposed method can choose up-to L largest norm columns from each of these sub-matrices to represent each user’s signal subspace. Since there is correlation between the antennas, the signal subspace chosen must have dimension L which is got by applying Gram-Schmidt orthogonalisation to each successive antenna chosen until up-to L orthogonal components are obtained for each user.
  • the data from the ( N r - L) antennas are projected onto the previously obtained signal subspace.
  • the advantage of projecting the signals include: since the signal subspace is of dimension atmost L, the rest of the (N r - L) components represent noise directions and by projecting the signals onto the signal subspace, only the desired signal components are extracted from the antennas while eliminating the noise components. Due to this the post- processing SNR increased by contributing to the overall signal power and simultaneously decreasing the overall noise power. Therefore, a denoising gain is provided that leads to better error performance with the compressed data compared to using the uncompressed data which is similar to the denoising gain of the low rank approximation observed in image processing applications.
  • each user sub-matrix is approximated as the function of two matrices: an orthogonal matrix Q (called the basis matrix) with upto L columns representing the basis for the signal subspace, and a projection matrix R that is upper triangular upto the column L.
  • Q orthogonal matrix
  • R projection matrix
  • the process of compression of the signals at the RRH (120) and the decompression at the BBU (140) in the MIMO C-RAN architecture is provided.
  • the RF down-conversion is performed on the signals received at the RRH (120).
  • the RF down-converted signals are used to construct the baseband signal matrix Y using the signals received at the N r antennas over a time span of N symbols. Therefore, the baseband signal matrix has a dimension of N X N r . Without the loss of generality, N is chosen to be the duration of one OFDM symbol.
  • the RRH (120) removes the CP and applies the fast Fourier transform (FFT) to the Y to convert the baseband signal matrix to the frequency domain signal matrix Y f . If the total number of sub-carriers allocated to all the users is N f , then the Y f is of dimension N f X N r . The removal of the CP and the guard-bands (which do not need to be sent to the BBU (140)) almost halves the amount of data to be sent. Further, at step 4, the RRH (120) performs the resource element (RE) demapping for separating the signals from the different users.
  • FFT fast Fourier transform
  • the RE demapping is performed by dividing the Y f into the plurality of sub-matrices corresponding to the different users according to the sub-carriers allocated to the individual users.
  • the sub-matrix of the user u is represented as Y u . If N fu is the number of sub-carriers allotted to the user u, then the Y u has a dimension of N fu X N r . Further, the RRH (120) assumes that the users are independent and L ⁇ Nr ⁇ N fu . Therefore, the true rank of the Y u will be less than the N r and equal to the number of independent multi-paths in the channel for user u.
  • the RRH (120) applies the QR compression algorithm to each of the Y u , for example as described below:
  • the low-rank approximated matrix Y uO Q u R u .
  • the RRH (120) chooses the L u antennas having the highest received powers from the Y u to form the columns of the Q u .
  • the value of the L u is chosen based on the channel state information and the required error performance for the user u.
  • the L is the maximum number of multi-paths that are assumed for each of the users but the L u is the actual number of antennas chosen for the signal subspace of the user u depending on the channel and requirements.
  • qi and n denote column vector i of Q u and R u , respectively.
  • e k i denotes column vector of length k with first component 1 and rest 0s.
  • the Y is converted to the frequency domain and divided into the plurality of the sub-matrices Y u corresponding to each user u.
  • the Y u is of dimension N fu X N r , where the N fu is the number of sub-carriers allotted to the user u.
  • Each Y u is then approximated to the product of the matrices Q u and R u by the QR approximation.
  • the Q u is of the dimension N fu X L u and the R u of dimension L u X N r , where L u ⁇ N r .
  • Uniform scalar quantization of b Q bits is applied to the samples of each Q u and R u . Therefore, the number of bits after compression, B cmp is given by:
  • the BBU (140) receives the Q u and the R u for the plurality of sub-matrices. Further, the BBU (140) reconstructs all the Y u o by taking the product of the corresponding Q u and R u . Further, at step 7, the decoding is initiated with the assumption that the channel is known at the BBU (140) and a zero-forcing equalization is performed on each of the reconstructed Y u o. At step 8, the BBU (140) performs the joint decoding where each user’s symbols from all the N r antennas are combined and the M-QAM symbols are demodulated.
  • FIG. 7 is a graph plot illustrating the compression ratios (CRs) for the proposed method of compression as a function of the actual number of antennas (Lu) for the signal subspace of UEs u, according to the embodiments as disclosed herein.
  • Compression Ratios for the proposed method as a function of L u , for 256 RRH antennas, 8 users and 12 users.
  • the CR is inversely proportional to Lu, as observed from the FIG. 7.
  • S VD singular value decomposition
  • the users are also not separated in the SVD compression scheme which does not have the flexibility of choosing user- specific trade-off between the compression ratio and the error performance present in the proposed method. Further, in another example even if the SVD compression is applied in the frequency-domain after the removal of the guard band and the user separation, calculating the SVD compression of each of the user matrix is more expensive in terms of complexity than the proposed method which uses the QR decomposition.
  • FIG. 8A is a graph illustrating Bit Error Rate (BER) performance for QR compression for 8 users, according to the embodiments as disclosed herein.
  • BER Bit Error Rate
  • FIG. 8B is a graph illustrating Bit Error Rate (BER) performance for QR compression for 12 users, according to the embodiments as disclosed herein.
  • BER Bit Error Rate
  • the Bit Error Rate (BER) performance of the proposed method is determined through Monte Carlo simulations using a massive MIMO uplink link level simulator in the baseband.
  • a 3 GPP tapped delay line (TDL) Rayleigh fading channel model for 5G is used for the BER performance evaluation. Further, a 100MHz bandwidth and 30kHz sub-carrier spacing is considered for which the FFT length is estimated to be 4096 and the CP length is 288. Therefore, the length of one OFDM symbol is 4384, which is the number of time samples N that is considered for one compression block. Further, 64-QAM is used with 256 receive antennas at the RRH (120). Thus, the dimension of the received signal matrix Y that is to be compressed is 4384 X 256.
  • the simulation parameters are provided in Table.1.
  • the BER performance is compared for an un-coded system using the proposed compression method against the SVD compression. Further, the BER of the baseline uncompressed system is also plotted for reference. Assuming uniform linear array, an antenna correlation is generated at the RRH (120) according to an exponential correlation model, with correlation coefficient of 0.7.
  • the UEs are allocated resource blocks (RBs) according to the received SNRs at the RRH.
  • the UEs with higher SNRs are allocated more RBs than those with lower SNRs.
  • the number of RBs allocated is 26, 28, 30, 32, 34, 36, 38, 40, respectively.
  • the RB allocation is 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, respectively. In both the cases, the total number of RBs allocated should not exceed 273.
  • the true rank of each user sub-matrix Yu is found to be 12, corresponding to the 12 non-zero taps in the multipath channel model used.
  • FIG. 7 shows the CRs achieved for different values of the Lu. Further observations reveal that lower the value of Lu, higher is the CR achieved.
  • the BER plots for two values of Lu i.e., 12 and 24 are provided in the FIGS. 8A- 8B to evaluate the impact of Lu on the performance of the proposed method. The observations reveal that the proposed method performs better for the higher value of the Lu for both the 8 UEs and the 12 UEs. Thus, the choice of the Lu in the proposed method is a trade-off between a desired CR and a required error performance.
  • Table. 2 shows the CRs achieved for the different values of the Lu and the Nu.
  • FIGS. 9A-9B are graph plots illustrating un-coded Bit Error
  • Rate (BER) performance for the proposed method in comparison with the S VD compression, according to the embodiments as disclosed herein.
  • the BER for the uncompressed system for both user cases are also plotted as reference.
  • the BERs are plotted for both the proposed method and the SVD compression with the constant value of the CR.
  • the samples are not converted into the frequency domain from the time domain.
  • the guard bands are not removed and the UEs are not separated resulting in a greater number of samples to be sent to the BBU than in the proposed method.
  • the plots in FIGS. 9A-9B also include the BER for the uncompressed system for both the number of UEs is 8 and 12 respectively.
  • the BER improvement for the proposed method compared to no compression is obtained due to the denoising gain of the low rank approximation applied in the proposed method. Therefore, the link level simulations show that the proposed method achieves up-to 17.4 X compression for 8 users and 14.5 X compression for 12 users in the wireless communication system.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Radio Transmission System (AREA)

Abstract

L'invention concerne selon certains modes de réalisation un procédé de compression de signaux reçus au niveau d'une station de base. Le procédé comprend la réception, par un dispositif de communication de la station de base, d'une pluralité de signaux à l'aide d'un réseau d'antennes dans la station de base sur un intervalle de temps et la compression, par une unité radio de la station de base, de la pluralité de signaux par sélection de signaux de base et représentation de la pluralité de signaux en termes de signaux de base, les signaux de base étant un ensemble de signaux de la pluralité de signaux qui sont reçus au niveau d'un ensemble d'antennes du réseau d'antennes. En outre, le procédé comprend la transmission, par l'unité radio de la station de base, du signal de base et de la pluralité de signaux représentés en termes de signaux de base pour chacun des signaux reçus de la pluralité de signaux à une unité distribuée.
PCT/IN2020/050102 2019-02-19 2020-01-31 Procédé et système de compression de signaux reçus au niveau d'une station de base WO2020170267A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IN201941006504 2019-02-19
IN201941006504 2019-02-19

Publications (1)

Publication Number Publication Date
WO2020170267A1 true WO2020170267A1 (fr) 2020-08-27

Family

ID=72143617

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IN2020/050102 WO2020170267A1 (fr) 2019-02-19 2020-01-31 Procédé et système de compression de signaux reçus au niveau d'une station de base

Country Status (1)

Country Link
WO (1) WO2020170267A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116470927A (zh) * 2023-04-17 2023-07-21 上海毫微太科技有限公司 一种数据处理方法、装置、设备和存储介质

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014076004A2 (fr) * 2012-11-16 2014-05-22 Telefonica, S.A. Procédé et système de compression et décompression sans perte de signaux numériques en bande de base dans des réseaux d'accès radio lte avancés distribués
US9998187B2 (en) * 2014-10-13 2018-06-12 Nxgen Partners Ip, Llc System and method for combining MIMO and mode-division multiplexing

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014076004A2 (fr) * 2012-11-16 2014-05-22 Telefonica, S.A. Procédé et système de compression et décompression sans perte de signaux numériques en bande de base dans des réseaux d'accès radio lte avancés distribués
US9998187B2 (en) * 2014-10-13 2018-06-12 Nxgen Partners Ip, Llc System and method for combining MIMO and mode-division multiplexing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HERBERT TAUB ET AL., PRINCIPLES OF COMMUNICATION SYSTEM, 3RD EDITION, CHAPTER 1, 7 September 2008 (2008-09-07) *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116470927A (zh) * 2023-04-17 2023-07-21 上海毫微太科技有限公司 一种数据处理方法、装置、设备和存储介质
CN116470927B (zh) * 2023-04-17 2024-06-04 上海毫微太科技有限公司 一种数据处理方法、装置、设备和存储介质

Similar Documents

Publication Publication Date Title
US11646929B1 (en) Spreading and precoding in OFDM
US9628231B2 (en) Spreading and precoding in OFDM
US8811145B2 (en) Detection method and apparatus for multiple-input multiple-output single carrier block transmission system
US11128322B2 (en) Methods and devices for determination of beamforming information
CN111133726B (zh) Mimo系统的帧结构感知压缩
KR20100097161A (ko) 다중입력 다중출력 시스템에서 프리코딩 방법 및 장치
KR102318817B1 (ko) 무선통신 시스템에서 필터 뱅크 다중 반송파 기법을 위한 전처리 방법 및 장치
JP6019298B2 (ja) 無線通信システム、無線送信装置および無線通信方法
US10256885B2 (en) Terminal device, feedback information generation method, and base station device
WO2021049986A1 (fr) Réduction de dimension de signal à l'aide d'une transformation non linéaire
WO2013168792A1 (fr) Dispositif de réception sans fil, dispositif d'émission sans fil, système de communication sans fil, programme et circuit intégré
WO2020170267A1 (fr) Procédé et système de compression de signaux reçus au niveau d'une station de base
CN112997459A (zh) Mimo-ofdm的papr降低
CN113411108B (zh) 信号调制与解调的方法、装置以及存储介质
JP4590604B2 (ja) 通信装置
CN108631829B (zh) 联合功率分配、预编码与解码方法及其基站
US9031169B2 (en) Sparse channel detection, estimation, and feedback
CN104868942A (zh) 通信设备和通信系统
CN113315551B (zh) 分层空时码系统的信号检测方法、装置和计算机设备
Aswathylakshmi et al. QR approximation for fronthaul compression in uplink massive MIMO
JP2015056690A (ja) 端末装置および受信装置
CN108418615B (zh) 一种基于mu-mimo有限反馈系统的用户调度方法
Ma et al. Distributed signal compressive quantization and parallel interference cancellation for cloud radio access network
CN101719816A (zh) 实现自适应mimo-scfde系统低反馈速率的方法
Aafreen et al. Low complexity csi feedback technique for fdd massive mimo systems

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: 20760085

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20760085

Country of ref document: EP

Kind code of ref document: A1