WO2021238634A1 - 一种下行预编码方法、装置及基站 - Google Patents

一种下行预编码方法、装置及基站 Download PDF

Info

Publication number
WO2021238634A1
WO2021238634A1 PCT/CN2021/092744 CN2021092744W WO2021238634A1 WO 2021238634 A1 WO2021238634 A1 WO 2021238634A1 CN 2021092744 W CN2021092744 W CN 2021092744W WO 2021238634 A1 WO2021238634 A1 WO 2021238634A1
Authority
WO
WIPO (PCT)
Prior art keywords
downlink
channel
weight
information
iteration
Prior art date
Application number
PCT/CN2021/092744
Other languages
English (en)
French (fr)
Inventor
郭森宝
李桂宝
乐春晖
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
Publication of WO2021238634A1 publication Critical patent/WO2021238634A1/zh

Links

Images

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
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0456—Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • 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
    • H04B7/0426—Power distribution
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • This application relates to the field of digital communication technology, and in particular to a downlink precoding method, device, and base station.
  • Precoding a digital communication signal processing link is to perform a pre-processing on the transmitted signal at the transmitting end when the channel state information is known.
  • the precoding link can be used to transmit multiple data streams in parallel through multiple transmit antennas to increase the peak transmission rate.
  • MIMO multiple input multiple output
  • the transmission power of each channel can be allocated by adjusting the downlink weight in the precoding link.
  • a single user-multiple input multiple output (SU-MIMO) and multi-user multiple input multiple output (multiple user-multiple input multiple output) can be realized by maximizing Shannon’s formula.
  • the weight and power allocation results are solved and output to achieve the optimal spectrum efficiency.
  • the foregoing precoding method requires multiple iterations, which results in a high complexity of solving the global optimal solution, and consumes a lot of computing resources and time.
  • This application provides a downlink precoding method, device, and base station to solve the problem that the traditional precoding method requires multiple iterations.
  • an embodiment of the present application provides a downlink precoding method, and the method includes:
  • the channel information includes the initial weight of the downlink channel and the channel characteristic matrix of the downlink channel;
  • the mutual information maximization (MIMA) algorithm is used to iteratively generate the power allocation weight of the downlink channel
  • the objective function of the MIMA algorithm is the lower bound function of the mutual information of the downlink channel
  • the initialization parameter of the MIMA algorithm is the initial Weights and channel characteristic matrix
  • the base station When the base station transmits downlink data to the terminal, it can first obtain the channel information corresponding to the downlink channel of the terminal, and extract the initial weight and channel characteristic matrix from the downlink channel information, so as to input the mutual information maximization algorithm to iteratively generate the power allocation rights of the downlink channel value.
  • the objective function can be set as the lower bound function of the mutual information of the downlink channel, and the initialization parameters are the initial weight and the channel characteristic matrix.
  • the power allocation weight can be finally obtained, and the downlink data is pre-coded according to the power allocation weight to send the downlink data to be transmitted to the terminal.
  • the power weight can be determined in real time according to the channel information of the downlink channel before the data is transmitted, and the channel power can be allocated, which can reduce power waste and alleviate throughput loss. And, by taking the lower bound function of the mutual information of the downlink channel as the objective function of the MIMA algorithm for optimization, and combining the gradient iteration to simplify the iteration steps, the optimal channel capacity allocation is effectively guaranteed.
  • the channel characteristic matrix used contains information that can characterize the characteristics of the current downlink channel.
  • the pairing information can be extracted from the channel information for a multi-user multiplexing scenario to generate channel characteristics matrix. That is, the step of acquiring the channel information of the downlink channel may include:
  • the pairing information includes one or more of modulation and coding strategy information, modulation scheme information, and constellation point type information of the target terminal and the paired terminal;
  • a channel characteristic matrix is generated, and the elements in the channel characteristic matrix are one or more of modulation and coding strategy information, modulation scheme information, and constellation point type information extracted from the pairing information.
  • the MIMA algorithm is used to iteratively generate the power allocation weight of the downlink channel, including:
  • the iterative gradient value as the initial step factor and the minimum step factor. Then the initial iteration rate is generated according to the initial weight and the channel characteristic matrix.
  • the weight matrix is updated.
  • the intermediate iteration rate is generated according to the updated weight matrix and channel characteristic matrix. Until the intermediate iteration rate is greater than or equal to the initial iteration rate, update the iteration gradient value; if the iteration gradient value is greater than the minimum step factor, continue to update the weight matrix, and calculate the intermediate iteration rate according to the updated weight matrix, and repeat the iteration ; If the iterative gradient value is less than or equal to the minimum step size factor, the weight matrix is extracted to obtain the power allocation weight.
  • This implementation provides an iterative method based on the MIMA algorithm.
  • the initial weight and channel characteristic matrix can be used as input, and through multiple iterations, the iterative rate is calculated to deduce the objective function to the minimum of the mutual information lower bound function. Value, so as to obtain the optimal power allocation weight after the iteration, and perform the relevant precoding operation.
  • each intermediate iteration rate can be obtained separately, and the obtained intermediate iteration rate can be compared with the initial iteration rate to determine whether the iteration is completed according to the size relationship between the two.
  • the update initial iteration rate is the intermediate iteration rate, that is, the next iteration process can be performed according to the updated initial iteration rate until the intermediate iteration rate is greater than or equal to the initial iteration rate, and the iteration gradient is updated Value, so that when the iterative gradient value is less than or equal to the minimum step size factor, iteratively extracts the weight matrix to obtain the power distribution weight.
  • the initial iteration rate is the target judgment value calculated during each iteration, and the initial iteration rate can be calculated and generated according to the following formula:
  • R n G(H,W n )
  • R n is the initial iteration rate
  • G(H, W) is the objective function, that is, the lower bound function of mutual information:
  • H is the channel characteristic matrix
  • W n is the initial weight matrix
  • e ij is the constellation point code distance
  • i and j are the constellation point numbers respectively
  • M Nt is the total number of constellation points
  • ⁇ 2 is the noise variance
  • n is the iteration index .
  • the iteration rate in each iteration process can be calculated separately. Since the objective function is the lower bound function of mutual information and includes the current channel characteristic matrix and power allocation weight, the objective function gradually tends to Lower bound value, and gradually update the weight matrix with the iterative process.
  • the weight matrix is a matrix used to guide the signal power allocation in the downlink channel, and each element in the matrix can correspond to the specific power allocation situation of each signal transmission process. Therefore, in an implementation manner, the following formula can be used Update the weight matrix:
  • ⁇ is the preset influence coefficient
  • H is the channel characteristic matrix
  • W n is the weight matrix
  • e ij is the constellation point code distance; i and j are the constellation point numbers respectively; M Nt is the total number of constellation points; ⁇ 2 is the noise variance; n is the index of the number of iterations; ⁇ is the iteration gradient value.
  • the weight can be updated according to the above formula.
  • the iterative gradient value can be updated when the intermediate iteration rate is greater than or equal to the initial iteration rate, and the iteration is repeated until the iterative gradient
  • the current weight matrix is extracted to obtain the power allocation weight.
  • calculate the channel capacity according to the output weight matrix so as to set the signal transmission mode according to the channel capacity.
  • the iterative gradient value can be updated according to the following formula:
  • ⁇ is the iterative gradient value
  • k is the convergence coefficient
  • the initial weight obtained from the channel information needs to be input into the MIMA algorithm to start the first iteration.
  • the initial weight is the weight of a single user in the downlink channel; or, the estimated weight of the downlink channel.
  • Single-user weights and estimated weights can be obtained through historical information or from the channel characteristic matrix of the downlink channel, and single-user weights or estimated weights can be used as reference values for the first iteration to finally obtain the output of the MIMA algorithm.
  • performing precoding on downlink data according to the power allocation weight includes: first generating an allocated power value according to the power allocation weight and the channel capacity of the downlink channel; and then applying the allocation Power value to the downlink data to send the downlink data according to the calculated power and weight.
  • the base station can also generate downlink signaling containing pairing information; and send the downlink signaling to the corresponding terminal, so that the pairing information is sent to the target terminal and the paired terminal respectively, so that the terminal can be based on the pairing information Equalize, detect and decode the transmitted data.
  • the downlink signaling is carried in radio resource control signaling and/or downlink control information signaling.
  • an embodiment of the present application also provides a downlink precoding device, which may include an acquisition module, a weight calculation module, and a precoding module for executing the downlink precoding method provided in the first aspect. It may also include modules for executing the method steps in the implementation manners of the first aspect.
  • the acquisition module user acquires channel information of the downlink channel, where the channel information includes the initial weight of the downlink channel and the channel characteristic matrix of the downlink channel;
  • the weight calculation module is used to iteratively generate the power allocation weight of the downlink channel using the MIMA algorithm, the objective function of the MIMA algorithm is the lower bound function of the mutual information of the downlink channel, and the initialization parameters of the MIMA algorithm are the initial weight and the channel characteristic matrix ;
  • the precoding module is configured to perform precoding on downlink data according to the power allocation weight, and the downlink data is data transmitted in the downlink channel.
  • the downlink precoding device may obtain channel information through an acquisition module, and input it to the weight calculation module, so that the weight calculation module uses the MIMA algorithm to iteratively solve the optimal solution. After solving the optimal solution, the precoding module can be used to perform precoding on the downlink data. Since the objective function of the MIMA algorithm running in the weight calculation module is the lower bound function of the mutual information of the downlink channel, and the initialization parameters are the initial weight and the channel characteristic matrix in the channel information, the downlink precoding device can determine according to the channel characteristics Obtain the optimal power distribution plan, simplify the iterative process, and quickly obtain the optimal power distribution weight.
  • the initial weight value is a single user weight value in a downlink channel; or, an estimated weight value of the downlink channel.
  • the initial weight can be used as a reference for the first iteration of the MIMA algorithm for subsequent iterations.
  • the weight calculation module is specifically configured to set the iterative gradient value as an initial step factor and a minimum step factor; and generate the initial iteration rate according to the initial weight and the channel characteristic matrix; if the iterative gradient value Greater than the minimum step size factor, update the weight matrix; generate the intermediate iteration rate according to the updated weight matrix and the channel characteristic matrix; if the intermediate iteration rate is greater than or equal to the initial iteration rate, update the iteration gradient value, and When the iterative gradient value is less than or equal to the minimum step size factor, the weight matrix is extracted to obtain the power distribution weight.
  • the weight calculation module in the foregoing implementation manner can use the gradient iteration algorithm in the MIMA algorithm to simplify the number of iterations and shorten the iteration time of the MIMA algorithm.
  • the weight calculation module is specifically configured to update the initial iteration rate to the intermediate iteration rate if the intermediate iteration rate is less than the initial iteration rate. That is, the weight calculation module can continue the next iteration by updating the initial iteration rate when the intermediate iteration rate is less than the initial iteration rate, until the objective function reaches the lower bound of the mutual information.
  • the initial iteration rate can be calculated and generated according to the following formula:
  • R n G(H,W n )
  • R n is the initial iteration rate
  • G(H, W) is the lower bound function of mutual information
  • H is the channel characteristic matrix
  • W is the initial weight matrix
  • e ij is the constellation point code distance
  • i and j are the constellation point numbers respectively
  • M Nt is the total number of constellation points
  • ⁇ 2 is the noise variance
  • n is the iteration number index.
  • the weight calculation module uses the MIMA algorithm to iteratively solve the optimal solution, it can update the weight matrix according to the following formula:
  • ⁇ is the preset influence coefficient
  • H is the channel characteristic matrix
  • W n is the weight matrix
  • ⁇ is the iterative gradient value
  • e ij is the constellation point code distance; i and j are the constellation point numbers respectively; M Nt is the total number of constellation points; ⁇ 2 is the noise variance; n is the index of the number of iterations.
  • the weight calculation module is specifically configured to update the iteration gradient value if the intermediate iteration rate is greater than or equal to the initial iteration rate, so as to perform the next iteration by adjusting the iteration gradient value. If the iterative gradient value is less than or equal to the minimum step size factor, the weight matrix is output; and the channel capacity is calculated according to the output weight matrix. That is, when the weight calculation module determines that the iterative gradient value is less than or equal to the minimum step size factor, it completes the iteration, and calculates the channel capacity according to the output weight matrix, so that the precoding module performs precoding according to the output result.
  • the weight calculation module can update the iterative gradient value according to the following formula:
  • ⁇ is the iterative gradient value
  • k is the convergence coefficient
  • the acquisition module is also used to acquire pairing information and generate a channel characteristic matrix according to the pairing information.
  • the pairing information includes one or more of the modulation and coding strategy information, modulation scheme information, and constellation point type information of the target terminal and the pairing terminal; the elements in the channel characteristic matrix are the modulation and coding strategy information extracted from the pairing information, One or more of modulation scheme information and constellation point type information.
  • the acquiring module can communicate the information sending mode of the target terminal and the paired terminal, which is convenient for data transmission between the target terminal and the paired terminal.
  • the precoding module is specifically configured to generate the allocated power value according to the power allocation weight and the channel capacity of the downlink channel; and then apply the allocated power value to the downlink data to send according to the calculated power and weight Downlink data.
  • the precoding device further includes a pairing module, the pairing module is used to generate downlink signaling including pairing information, and send the downlink signaling to the target terminal and the paired terminal.
  • the downlink signaling may be carried in radio resource control signaling and/or downlink control information signaling to facilitate the transmission of downlink data.
  • an embodiment of the present application also provides a base station, which may include a signal transceiver station and a controller, wherein the signal transceiver station is connected to the controller; the controller is configured to execute operation instructions to achieve
  • the downlink precoding method provided in the first aspect is used to control the signal transceiver station to perform precoding on downlink data.
  • an embodiment of the present application also provides a communication device, and the communication device may be a terminal or a chip or a system on a chip in the terminal.
  • the communication device can implement the functions performed by the terminal in each of the foregoing aspects or in each possible implementation manner, and the functions can be implemented by hardware.
  • the communication device may include a processor and a communication interface, and the processor may be used to support the communication device to implement the foregoing first aspect or the downlink precoding method involved in any one of the possible implementation manners of the first aspect.
  • the embodiments of the present application also provide a computer-readable storage medium.
  • the computer-readable storage medium may be a readable non-volatile storage medium, and the computer-readable storage medium stores instructions when it is When running on a computer, the computer can execute the downlink precoding method described in the first aspect or any one of the possible implementation manners of the foregoing aspects.
  • the embodiments of the present application also provide a computer program product containing instructions, which when run on a computer, enable the computer to execute the first aspect or any one of the possible implementations of the foregoing aspects.
  • the downlink precoding method is a sixth aspect.
  • an embodiment of the present application also provides a communication device.
  • the communication device may be a terminal or a chip or a system on a chip in the terminal.
  • the communication device includes one or more processors and one or more memories.
  • the one or more memories are coupled with the one or more processors, and the one or more memories are used to store computer program codes, and the computer program codes include computer instructions.
  • the communication device is caused to execute the downlink precoding method described in the foregoing first aspect or any possible implementation manner of the first aspect.
  • the MIMA algorithm is used to iteratively generate the power allocation weight of the downlink channel, where the objective function of the MIMA algorithm is the mutual information of the downlink channel
  • the initialization parameters of the MIMA algorithm are the initial weights in the channel information and the channel characteristic matrix; then the downlink data is pre-coded according to the power allocation weights to realize the transmission of the downlink data.
  • the downlink precoding method takes the lower bound function of mutual information as the objective function, and completes the calculation of power and weight through iterations to determine the optimal constellation combination, and ensures that the channel capacity allocation is maintained in the discrete modulation constellation under a small number of iterations The global optimal solution at time.
  • Figure 1 is a schematic diagram of a MIMO system scenario in an embodiment of the application
  • FIG. 2 is a schematic flowchart of a downlink precoding method in an embodiment of this application.
  • FIG. 3 is a schematic diagram of the constellation combination effect after power and weight adjustment in an embodiment of this application.
  • FIG. 4 is a schematic diagram of a comparison effect of power allocation in an embodiment of the application.
  • FIG. 5 is a schematic flowchart of a MIMA algorithm in an embodiment of this application.
  • FIG. 6 is a schematic diagram of a flow of acquiring channel information in a multi-user multiplexing scenario in an embodiment of the application
  • FIG. 7 is a schematic diagram of a flow of performing precoding in a multi-user multiplexing scenario in an embodiment of the application
  • FIG. 8 is a schematic diagram of a flow of sending downlink signaling in an embodiment of the application.
  • FIG. 9 is a schematic structural diagram of a downlink precoding apparatus in an embodiment of this application.
  • the downlink channel is the channel used by the base station to send data to the terminal, including the physical downlink control channel (PDCCH) and the physical downlink shared channel (PDSCH) (or called Is the downlink data channel).
  • PDCCH physical downlink control channel
  • PDSCH physical downlink shared channel
  • the physical downlink control channel is mainly used to carry downlink control information (DCI), which is usually expressed as DCI signaling.
  • DCI may include public control information (such as system information, etc.) and user-specific information (such as downlink resource allocation indication, uplink scheduling, random access response, uplink power control parameters, etc.).
  • the PDCCH can schedule the data channel through the DCI carried by it.
  • DCI can be used to indicate the transmission parameters of the data channel (such as the time domain resource location of the data channel, etc.).
  • the base station or other network equipment can provide the terminal To send the PDCCH, after receiving the PDCCH, the terminal can first demodulate the DCI in the PDCCH, and then receive or send the data channel at the time domain resource location indicated by the DCI.
  • the physical downlink data channel is used to carry data sent from the base station to the terminal (or called downlink data).
  • the PDCCH may also indicate the time domain resource location of the channel state information reference signal (CSI-RS) through the DCI carried by the PDCCH to trigger the transmission of aperiodic (nonperiodic) CSI-RS.
  • the CSI-RS is used for the terminal to measure the channel state between the terminal and the base station, and the CSI-RS may include one or more channel state measurement resources.
  • the base station may send DCI and CSI-RS for indicating the time domain resource position of the CSI-RS to the terminal.
  • the terminal receives the CSI-RS at the time domain resource position indicated by the DCI, and responds to the channel status included in the CSI-RS.
  • the measurement resource is measured, and the channel state information (channel state information, CSI) is reported to the base station according to the measurement result.
  • the base station can also set the data transmission mode of the downlink channel according to the CSI reported by the terminal, such as setting the optimal constellation combination mode.
  • the base station and the terminal include multiple antennas, which can form multiple data streams when sending downlink data.
  • the base station can send downlink data to the terminal according to the set constellation combination mode, and make full use of the channel capacity.
  • constellation is constellation mapping, which is a digital modulation technology.
  • the process of constellation mapping is to map a finite field "bit" sequence that carries digital information into a "symbol" sequence suitable for transmission.
  • the value space of each symbol can be one-dimensional real number space or two-dimensional real number space (ie, complex number space).
  • Constellation mapping includes two elements, namely constellation and constellation point mapping method (labeling).
  • the constellation diagram represents a set of all values of the output symbols of the constellation mapping, where each point of the constellation diagram corresponds to a value of the output symbols.
  • the constellation point mapping mode represents a specific mapping relationship from input bits (sequence/group) to constellation points, or a specific mapping relationship from constellation points to bits (sequence/group).
  • constellation diagrams mainly include pulse amplitude modulation (PAM) in one-dimensional real space, quadrature amplitude modulation (QAM) in two-dimensional real space, and phase shift keying (PSK) modulation. Wait.
  • PAM
  • the process in which the base station encodes and converts the downlink data to achieve constellation mapping according to the set constellation combination method is called precoding.
  • the base station may perform optimization solutions on the data based on the precoding, such as the power allocation matrix, so as to obtain the corresponding data transmission mode.
  • the following describes the flow of the base station setting the data transmission mode of the downlink channel.
  • the data transmission method of the downlink channel set by the base station can be applied to the communication system of MIMO mode, such as: the fourth generation (4th generation, 4G) system, the long term evolution technology (long term evolution, LTE) system, and the fifth generation (5th generation, 5G) system, new radio (NR) system, NR-vehicle-to-everything (V2X) system, can also be applied to other next-generation communication systems Wait.
  • the fourth generation (4th generation, 4G) system the long term evolution technology (long term evolution, LTE) system
  • 5th generation, 5G) system new radio (NR) system
  • NR-vehicle-to-everything (V2X) system can also be applied to other next-generation communication systems Wait.
  • the MIMO system includes a base station 100 and multiple terminals 200.
  • multiple terminals 200 may establish a communication connection with the base station 100.
  • the base station 100 is mainly used to implement functions such as resource scheduling, wireless resource management, and wireless access control of the terminal 200.
  • the base station 100 may also be other network equipment, such as: a small base station, a wireless access point, a transceiver point (transmission receive point, TRP) transmission point (transmission point, TP), and any one of some other access nodes Wait.
  • the base station 100 has a built-in antenna so that the base station 100 can receive and transmit radio signals during operation and form a radio coverage area.
  • the terminal 200 in the coverage area can connect to the Internet or connect to other terminals through the radio signal transceiving relationship with the base station 100 to implement mobile communication services.
  • the device used to implement the function of the base station may be a base station, or may be a device or functional module capable of supporting the base station to implement the function, such as a chip system.
  • the terminal 200 may be a terminal equipment (terminal equipment) or a user equipment (user equipment, UE), a mobile station (mobile station, MS) or a mobile terminal (mobile terminal, MT), or the like.
  • the terminal can be a mobile phone (mobile phone), a tablet computer, or a computer with wireless transceiver function, it can also be a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control , Wireless terminal in unmanned driving, wireless terminal in telemedicine, wireless terminal in smart grid, wireless terminal in smart city (smart city), smart home, vehicle-mounted terminal, etc.
  • VR virtual reality
  • AR augmented reality
  • Wireless terminal in unmanned driving, wireless terminal in telemedicine, wireless terminal in smart grid, wireless terminal in smart city (smart city), smart home, vehicle-mounted terminal, etc.
  • the base station 100 After the terminal 200 accesses the base station 100, the base station 100 will act as a transmitting end to send data to the terminal 200.
  • the transmitted data is called downlink data, and the corresponding data transmission channel is called a downlink channel.
  • the device used to implement the function of the terminal may be a terminal, or a device capable of supporting the terminal to implement the function, such as a chip system.
  • multiple antennas can be used in both the base station 100 and the terminal 200. Because multiple antennas are used, downlink data can be sent through multiple antennas at the same time to form multiple data streams, thereby improving the transmission efficiency of downlink data. .
  • the base station 100 may also establish a connection relationship with multiple terminals 200 to form a multi-user multiplexing mode, so as to realize communication between the multiple terminals 200.
  • the signal form transmitted in the base station is also different.
  • current mobile communication technologies mostly use digital signals for transmission between the base station 100 and the terminal 200, such as LTE technology and NR technology.
  • the signals defined by LTE technology and NR technology are discrete and uniformly distributed signals, including QPSK, 16QAM, 64QAM, 256QAM and other forms.
  • a single codeword transmission method can be used.
  • the LTE technology uses a single codeword to multi-layer mapping transmission mode when outputting more than 2 streams.
  • the NR technology also uses single-code sub-transmission for transmissions of no more than 4 streams.
  • This single-code word transmission mode has the same number of antennas and the constraints of the sky size, and the single-user or multi-user capacity of the multiple-input multiple-output scenario is limited, which is close to the limit value of the theoretical Shannon formula. If you need to further improve the performance , Need to increase the number of antennas or the size of the sky.
  • the weight matrix (or precoding matrix) on which the downlink data is pre-coded is still the general weight determined according to Shannon’s formula, resulting in power
  • the allocation does not match the transmission characteristics of each signal in the downlink channel. Therefore, the constellation combination configured for each downlink data stream is not an optimal combination method. That is, the demodulation performance of part of the transmission layer is limited by the minimum code distance between constellations, and transmission errors occur; and part of the transmission layer is allocated with higher power and cannot obtain higher throughput.
  • the optimal weight design is related to the characteristics of the input signal, and according to the Shannon formula, when the input signal obeys the Gaussian distribution, the equivalent channel capacity is maximized.
  • an exemplary implementation of the method for solving the optimal weight in this application is as follows:
  • the function corresponding to the channel capacity can be set as the objective function, and through iterative solution, the corresponding power weight value when the channel capacity is maximized can be calculated.
  • the channel capacity can be maximized, namely:
  • h(Y) is the entropy information of the output signal Y
  • X) is the conditional entropy information of the output signal Y when the input signal is X
  • P Y is the variance of the output signal Y
  • P X is the input
  • P S is the input signal power
  • W is the system bandwidth
  • N 0 is the bilateral power spectral density of the white noise
  • N is the number of samples for calculating the mutual information under the bandwidth W.
  • the maximum method based on Shannon's formula can be used to calculate the power distribution weight of the MIMO system.
  • the singular value decomposition (SVD) of the channel autocorrelation matrix is used, and the eigenvectors after SVD decomposition are used for precoding, namely:
  • the optimal value of the weight of a single user can be obtained by taking the feature vector corresponding to the feature value of the maximum number of layers corresponding to the feature vector.
  • SINR signal to loss plus noise ratio
  • [] + is the pseudo-inverse operation
  • (:,k) represents the kth column of the matrix
  • W k is the calculated weight of the k-th user.
  • R k is the k th user channel autocorrelation matrix
  • R i is the i th user autocorrelation matrix
  • W k is the weight calculated in the k-th user.
  • the variance of the input signal is required to obey the Gaussian distribution.
  • the transmitted signal X does not obey the Gaussian distribution, but obeys the discrete uniform distribution. Therefore, the maximum channel capacity obtained by Shannon's formula is only a suboptimal solution, not an optimal solution.
  • the initial iteration parameters can be set according to the channel information in the downlink channel.
  • the algorithm of power control and space rotation can be used to solve the optimal value, namely:
  • K is a quantized granularity parameter, and normalize its (K+1) Nt combinations to obtain the value of each diagonal element:
  • an iteration result can be generated, and the iteration result can be used as the power allocation weight of the downlink data. That is, when the objective function, that is, the channel capacity is maximized, the corresponding power allocation weight is output, and precoding is performed according to the power allocation weight, so that the optimal spectrum efficiency can be achieved.
  • an embodiment of the present application provides a downlink precoding method, which is used to obtain a global optimal solution through a small number of iterations in the process of precoding downlink data, so that the downlink data can be adjusted according to the power allocation weight. Perform precoding operations.
  • the downlink precoding method includes the following steps:
  • S401 Acquire channel information of a downlink channel.
  • the base station can obtain the channel information of the downlink channel.
  • the DCI used to obtain channel information is sent to the terminal through the PDCCH, so that the terminal can measure the channel state measurement resource and feed back the measurement result to the base station.
  • the channel information includes the channel characteristic matrix and the initial weight.
  • a channel characteristic matrix can be formed through feedback results.
  • Each element in the channel characteristic matrix is used to reflect the signal transmission characteristics between the corresponding terminal and the base station.
  • the channel characteristic matrix may include the modulation and coding scheme (MCS) information and modulation scheme of the current downlink channel. Information and constellation point type information, etc.
  • MCS modulation and coding scheme
  • the channel characteristic data can be converted into specific values by using a specific quantization method, so as to be embodied as each element in the above-mentioned channel characteristic matrix.
  • the initial weight can be a single user weight or an estimated weight.
  • the single user weight refers to the weight when the signal is transmitted to the terminal through the downlink channel in a non-MIMO scenario.
  • the estimated weight refers to the weight obtained by estimation based on the access characteristics of the terminal in the scene, such as the number of terminals, the number of antennas, and the size of the sky, combined with the channel capacity.
  • the channel information obtained in step S401 can be used as the initialization parameter of the MIMA algorithm to iteratively solve the optimal solution of the power allocation weights using the MIMA algorithm.
  • S402 Use the channel information as the initialization parameter, and use the MIMA algorithm to iteratively generate the power allocation weight of the downlink channel.
  • the base station can calculate the power distribution weight by calling the MIMA algorithm.
  • the objective function of the called MIMA algorithm is the lower bound function of the mutual information of the downlink channel.
  • the initial weights and the channel characteristic matrix in the channel information are input into the MIMA algorithm to execute the MIMA algorithm Carry out iterative optimization solution.
  • G(H, W) is the lower bound function of mutual information
  • H is the channel characteristic matrix
  • W n is the initial weight matrix
  • e ij is the constellation point code distance
  • i and j are the constellation point numbers
  • M Nt is the constellation point The total number
  • ⁇ 2 is the noise variance
  • n is the index of the number of iterations.
  • the mathematical model of the mutual information maximization problem can be equivalent to the problem of minimizing the lower bound function of mutual information G(H,W), that is, the objective function of the MIMA algorithm is set as follows:
  • the mutual information maximization problem can be transformed into the mutual information lower bound problem for power allocation and weight calculation. Since the iterative calculation complexity of the lower bound function of mutual information G(H,W) is lower than that of the mutual information maximization function, the calculation complexity in the iterative process can be simplified, that is, the complexity in engineering implementation can be reduced , It is helpful to quickly determine the global optimal solution of the channel capacity in the discrete modulation constellation.
  • the weight matrix W n can also be updated by the gradient iteration method, namely:
  • ⁇ is the preset influence coefficient
  • H is the channel characteristic matrix
  • W n is the weight matrix
  • ⁇ is the iterative gradient
  • e ij is the constellation point code distance; i and j are the constellation point numbers respectively; M Nt is the total number of constellation points; ⁇ 2 is the noise variance; n is the iteration number index.
  • the lower bound function of mutual information G(H,W) can be used as the objective function
  • the channel characteristic matrix H and the initial weight W 1 can be used as the iterative initialization parameters to perform gradient iteration to obtain the minimum objective function G(H,W)
  • the corresponding weight matrix W in the case of transformation.
  • S403 Perform precoding on the downlink data according to the power allocation weight.
  • the output result of step S402 can be used as the precoding input to perform precoding on the downlink data. Since the base station and the terminal can use the multi-layer constellation combination for data transmission, the iterative solution can be used to obtain the combination of power allocation weights and data to achieve the optimal multi-layer constellation combination, and to ensure that the combined constellation can be obtained Optimal throughput.
  • a combination of multi-layer constellations can be used for data transmission.
  • the MIMA algorithm provided in the above-mentioned embodiment is used for solution calculation, which can ensure that the combined constellation method can obtain the optimal throughput.
  • two QPSK constellations are combined into a discrete Gaussian constellation through constellation convergence, which is more in line with Gaussian distribution.
  • the traditional method is used to allocate power P0+P1 to layer 0 and P2 to layer 1. Due to the data discrete characteristics, allocating higher power to layer 0 cannot achieve higher throughput. , Resulting in power waste and throughput loss. However, after using the MIMA algorithm provided in the foregoing embodiment, since it is assumed that the data is discrete data, power P0 can be allocated to layer 0, and power P0+P1 can be allocated to L1, thereby ensuring optimal throughput performance.
  • the step of using the MIMA algorithm to iteratively generate the power allocation weight of the downlink channel shown in step 2 may specifically include the following steps:
  • S4021 Set the iterative gradient value as the initial step size factor and set the minimum step size factor.
  • the initial value of the subsequent iterative process can be determined.
  • the iteration can be started according to the initial step size factor, and the time to complete the iteration can be determined by comparing with the minimum step size factor.
  • S4022 Generate an initial iteration rate according to the initial weight and the channel characteristic matrix.
  • R n G(H,W n )
  • R n is the initial iteration rate
  • G(H, W) is the lower bound function of mutual information, namely:
  • H is the channel characteristic matrix
  • W n is the initial weight
  • e ij is the constellation point code distance
  • i and j are the constellation point numbers respectively
  • M Nt is the total number of constellation points
  • ⁇ 2 is the noise variance
  • n is the number of iterations index.
  • the iteration gradient value can be compared with the minimum step size factor to determine whether the iteration has been completed. If the iterative gradient value is greater than the minimum step size factor, this iteration process has not been completed. Therefore, the weight matrix W n can be updated and the next iteration can be continued. If the iteration gradient value is less than or equal to the minimum step size factor, it is determined that this iteration has been completed, so the weight matrix at this time can be output, so as to calculate the channel capacity according to the output weight matrix.
  • the set initial step size factor must be greater than the set minimum step size factor ⁇ min , so in the first iteration process, the iteration gradient value is greater than the minimum step size factor, That is to update the weight matrix.
  • the weight matrix can be updated according to the following formula:
  • ⁇ is the iterative gradient value
  • ⁇ is the preset influence coefficient
  • I the first derivative of the lower bound function of mutual information G(H,W) with respect to the weight W, namely:
  • H is the channel characteristic matrix
  • W is the initial weight
  • e ij is the constellation point code distance
  • i and j are the constellation point numbers respectively
  • M Nt is the total number of constellation points
  • ⁇ 2 is the noise variance
  • n is the number of iterations index.
  • the weight matrix needs to be updated. Therefore, the first iteration process can be calculated According to the calculation Update the weight matrix, namely
  • S4024 Generate an intermediate iteration rate according to the updated weight matrix and the channel characteristic matrix.
  • R n+1 G(H,W n+1 )
  • the intermediate iteration rate can be compared with the initial iteration rate, so as to determine whether this iteration has been completed according to the comparison result.
  • the iterative gradient value can be updated, and the iteration can continue to make the lower bound of mutual information
  • the function gradually approaches the minimum value until the iterative gradient value is less than or equal to the minimum step size factor to extract the corresponding weight matrix to obtain the power allocation weight of each signal in the current downlink channel.
  • step S4025 Under another comparison result in parallel with step S4025, if the intermediate iteration rate is less than the initial iteration rate, it means that the current lower bound function of mutual information G(H, W) has not yet obtained a minimized result. Therefore, it is possible to update the initial iteration rate to the intermediate iteration rate, and jump to step S4023 to continue the iteration until the intermediate iteration rate calculated in the iteration process is greater than or equal to the initial iteration rate calculated in the previous iteration, and the corresponding weight matrix is output.
  • the initial iteration rate may be set to R 2 , and step S4023 may be skipped to continue the iteration.
  • the iteration gradient value can be updated by gradient iteration to further determine whether the minimization result of the mutual information lower bound function G(H, W) is obtained. That is, the MIMA algorithm may include: if the intermediate iteration rate is greater than or equal to the initial iteration rate, updating the iteration gradient value.
  • the iterative gradient value can be updated according to the following formula:
  • ⁇ is the iterative gradient value
  • k is the convergence coefficient
  • the intermediate iteration rate can also be calculated by performing step S4024, and the intermediate iteration rate can be compared with the initial iteration rate by performing step S4025, and repeated iterations can be implemented until the intermediate iteration rate is greater than or It is equal to the initial iteration rate, and the iteration gradient value is less than the set minimum step factor, the iteration is determined to be completed, and the final weight matrix is output.
  • the corresponding element is extracted from the weight matrix as the power allocation weight of each signal in the current downlink channel to implement precoding of the downlink data.
  • the initial weight W 1 and the channel characteristic matrix H in the channel information can be extracted, and the initial weight W 1 and the channel characteristic matrix H can be used as the initialization parameters of the MIMA algorithm.
  • set the initial step size factor ⁇ ⁇ init and set the minimum step size factor ⁇ min .
  • the initial iteration rate is set to the intermediate iteration rate R 2 to continue the iteration, and the weight matrix W 2 is used as the initial weight value, and is compared with the channel characteristic matrix H Together as the initialization parameter of the MIMA algorithm.
  • step factor iteration gradient value ⁇ " is less than or equal to the minimum step factor ⁇ min . If it is determined that the step factor iteration gradient value ⁇ " is less than or equal to the minimum step factor ⁇ min , then the iteration is completed and the weight matrix W n+m at this time is output. Then the channel capacity is calculated according to the output weight matrix, In order to determine the weight and power assigned to each signal.
  • the above implementation method can solve the global optimal solution after a dozen or dozens of iterations. Compared with the thousands of iterations in academic theories, the complexity of the iteration steps can be greatly simplified. This saves solution time, so that the above implementation can be applied in the real-time precoding process of downlink data. That is to say, the current downlink channel can be optimally solved, and the downlink data can be pre-coded according to the solution result, so as to achieve the goal of not reducing the signal transmission efficiency, but also obtaining the optimal throughput, and improving the base station in the MIMO system scenario Downlink data transmission efficiency improves signal transmission quality.
  • Performing precoding can include:
  • S4031 Generate an allocated power value according to the power allocation weight and the channel capacity of the downlink channel.
  • the base station can calculate the allocated power value of the signal in combination with the channel capacity of the current downlink channel.
  • the allocated power value refers to the transmission power of the corresponding digital signal. Different types of signals require different transmission powers, and the corresponding multi-layer constellations have different combinations.
  • S4032 Apply the allocated power value to the downlink data.
  • the optimal solution obtained by calculation can be used to obtain the best transmission mode of the downlink data, so as to make reasonable use of the channel capacity. Therefore, after the allocated power value is generated, the allocated power value can be applied to the downlink data to form a downlink data stream, so that the downlink data is sent to the terminal according to the calculated optimal power and weight, and the application channel capacity is maximized.
  • the downlink precoding method provided in the above embodiments can be applied to a single-user multiple input multiple output system, and it can also be applied to a multi-user multiple input multiple output system. Both can obtain the characteristics of the current downlink channel before transmitting the downlink data. The power distribution weights to obtain the best throughput performance.
  • the system since the system includes multiple user terminals, multiple user terminals can communicate with each other through the base station. Therefore, when one terminal (target terminal) is paired with another terminal (paired terminal), the two terminals The communication data generated between them can also generate downlink data in the base station. For example, after terminal A is paired with terminal B, the data sent by terminal A to terminal B is first sent to the base station; the base station then sends the corresponding data as downlink data to terminal B through the downlink channel.
  • downlink data refers to data sent from a base station to a terminal.
  • the downlink data includes not only the communication data sent by the paired terminal to the target terminal, but also other data sent to the target terminal. , Such as data sent from the Internet to the target terminal.
  • the pairing relationship is not limited to a pairing relationship between two terminals, and may also be a pairing relationship between multiple terminals.
  • the base station may obtain channel characteristics related to the paired terminal as channel information to form a channel characteristic matrix. Therefore, as shown in FIG. 7, in order to adapt to a multi-user multiplexing scenario, the step of acquiring channel information of a downlink channel shown in step S401 may include:
  • the pairing information is used to indicate the corresponding channel characteristic data of the terminal having the pairing relationship, and may include one or more of modulation and coding strategy information, modulation scheme information, and constellation point type information of the target terminal and the paired terminal.
  • the modulation and coding strategy information is usually expressed as the MCS index value, which can be used to implement the rate configuration in LTE.
  • the MCS information the concerned factors affecting the communication rate can be used as the columns of the table, and the MCS index value can be used as the rows to form a rate table.
  • each MCS index actually corresponds to a physical transmission rate under a set of parameters.
  • the modulation scheme information is used to characterize the digital modulation parameters when the current base station transmits digital signals, such as the parameters of the carrier signal and the related parameters such as source coding, encryption, and equalization.
  • digital signals such as the parameters of the carrier signal and the related parameters such as source coding, encryption, and equalization.
  • Different base stations, terminals and even downlink channels can use different forms of modulation schemes.
  • the modulation/demodulation method of downlink channel data can be determined through the modulation scheme information, so that the transmitted signal has stronger anti-interference performance and anti-channel loss performance, and has Better security.
  • the constellation point type information is data used to characterize the data mapping state.
  • digital signals are represented on a complex plane to intuitively represent the signals and the relationship between the signals. Therefore, after the data is channel-coded, it is mapped on the constellation diagram.
  • the different types of constellation points will affect the result of data mapping on the constellation diagram.
  • the channel characteristic data may include modulation and coding strategy information, modulation scheme information, and constellation point type information, etc. Therefore, the pairing information may be extracted from the channel characteristic data corresponding to the target terminal and the paired terminal, respectively.
  • S4012 Generate a channel characteristic matrix.
  • the base station After extracting the pairing information, the base station can use the pairing information to generate a channel characteristic matrix.
  • corresponding data such as one of modulation and coding strategy information, modulation scheme information, and constellation point type information, can be extracted from the pairing information based on the original channel characteristic matrix. Multiple types, and add the extracted data to the channel feature matrix.
  • the above three kinds of information can be combined to form a channel characteristic matrix used to characterize the characteristics of the downlink channel.
  • the channel information is not limited to the above three types, and can also include other types of information, such as the number of antennas, sky size, channel capacity, etc., which can constitute different channel information according to key factors considered in the actual signal transmission process.
  • the channel characteristics related to the paired terminal can be added to the generated channel characteristic matrix.
  • the results of power calculation and weight calculation can have the channel characteristics of the paired terminals, and the two terminals paired with each other can use the corresponding channel characteristics for equalization and detection And decoding, thereby improving the signal transmission efficiency between the two paired terminals.
  • pairing information on the one hand, it can be used as channel information to form a channel characteristic matrix, and on the other hand, it can also be sent to each terminal having a pairing relationship through the base station. If multiple users can learn information related to data transmission such as the corresponding MCS, modulation scheme, constellation point type, etc., it is possible to directly adopt a suitable data processing method after pairing. For example, the same modulation/demodulation method can be used to reduce the problem that the terminal cannot demodulate data or use a blind solution method for demodulation, which can speed up the terminal’s data equalization, detection and decoding process, and facilitate the realization of the maximum channel capacity The purpose of the value.
  • the base station can send signaling to the target terminal and the paired terminal respectively to notify the target terminal and the paired terminal to obtain channel characteristics from each other, that is, as shown in Figure 8, in one implementation,
  • the downlink precoding method also includes:
  • S404 Send downlink signaling.
  • the base station may generate downlink signaling for the target user and the paired user, and may send the downlink signaling to the target terminal and the paired terminal respectively, so as to notify the target terminal and the paired terminal of the pairing information. That is, the target terminal obtains pairing information such as modulation and coding strategy information, modulation scheme information, and constellation point type information corresponding to the paired terminal notified by the base station by receiving downlink signaling, and then uses the received pairing information to perform equalization, detection, and decoding. Similarly, the paired terminal can also obtain the pairing information corresponding to the target terminal notified by the base station by receiving downlink signaling, and use the pairing information to perform equalization, detection, and decoding. Among them, the downlink signaling may include pairing information such as MCS, modulation scheme, and constellation point type.
  • the downlink signaling can be directly parsed by the terminal, so that pairing information such as modulation and coding strategy information, modulation scheme information, and constellation point type information of the paired terminal can be extracted from the downlink signaling.
  • the downlink signaling is carried in radio resource control (radio resource control, RRC) and/or DCI signaling.
  • RRC radio resource management
  • RRM radio resource management
  • RA radio resource assignment
  • Corresponding RRC signaling is the signaling when the base station implements radio resource management, control, and scheduling.
  • the signaling can be directly transmitted between the terminal and the base station without sending it in the form of downlink data.
  • the DCI signaling can be carried by the downlink physical control channel and is dedicated to sending downlink control information to the terminal, including uplink and downlink resource allocation, hybrid automatic repeat request information, power control, etc., and it does not need to be sent in the form of downlink data.
  • the terminal's downlink signaling can be carried in RRC signaling or DCI signaling or received through a combination of RRC signaling and DCI signaling.
  • the terminal can use the corresponding downlink signaling to perform equalization, detection, and decoding processes to obtain optimal power allocation and throughput performance.
  • the downlink precoding method provided in the foregoing embodiment and the downlink precoding method provided in various implementations or the steps included in the method can be combined with each other to obtain more implementations of the precoding method. Repeat it again.
  • a downlink precoding device is also provided.
  • the downlink precoding apparatus may be used to implement the downlink precoding method provided in the foregoing embodiment.
  • the downlink precoding device includes: an acquisition module 1, a weight calculation module 2, and a precoding module 3, which are respectively configured to perform step S1 and step S2 in the foregoing embodiment. And step S3 to perform precoding on the downlink data.
  • the acquisition module 1 is used to acquire channel information of a downlink channel and send the channel information to the weight calculation module 2, where the channel information includes the initial weight of the downlink channel and the channel characteristic matrix of the downlink channel.
  • the weight calculation module 2 is configured to receive the channel information sent by the acquisition module 1, and use the MIMA algorithm to iteratively generate the power allocation weight of the downlink channel.
  • the objective function of the MIMA algorithm is the lower bound function of the mutual information of the downlink channel
  • the initialization parameters of the MIMA algorithm are the initial weight and the channel characteristic matrix.
  • the weight calculation module 2 may also send the power allocation weight to the precoding module 3 after calculating and obtaining the power allocation weight.
  • the precoding module 3 is configured to receive power allocation weights, and perform precoding on downlink data according to the power allocation weights.
  • the aforementioned downlink precoding device can obtain channel information through the acquisition module 1 before sending the downlink data. Then the weight calculation module 2 executes the MIMA algorithm, and iteratively calculates the optimal solution of the power and the weight according to the channel information, and obtains the power allocation weight. Finally, the precoding module 3 performs precoding on the downlink data according to the power allocation weight, and determines the optimal combination mode of the multi-layer constellation, that is, the optimal throughput is obtained.
  • the weight calculation module 2 can also perform the following operations to obtain the power allocation weight: first set the iterative gradient value to the initial step factor and the minimum step factor; then according to the initial weight and channel characteristics Matrix generates the initial iteration rate; if the iteration gradient value is greater than the minimum step size factor, update the weight matrix; then generate the intermediate iteration rate based on the updated weight matrix and channel characteristic matrix; if the intermediate iteration rate is greater than or equal to the initial iteration rate, then Extract the weight matrix to obtain the power distribution weight.
  • the embodiment of the present application provides a downlink precoding method, which can use MIMA algorithm to iteratively generate the power allocation weight of the downlink channel after acquiring the channel information of the downlink channel, and then execute the downlink data according to the power allocation weight Precoding realizes the transmission of downlink data.
  • the downlink precoding method takes the lower bound function of mutual information as the objective function, and completes the calculation of power and weight through iterations to determine the optimal constellation combination, and ensures that the channel capacity allocation is maintained in the discrete modulation constellation under a small number of iterations The global optimal solution at time.
  • each module in the downlink precoding device in the above implementation is only a logical function division, which can be fully or partially integrated into a physical entity in the actual implementation process, or can be physically separated .
  • the acquisition module 1 can be implemented by a signal transceiver
  • the weight calculation module 2 and the precoding module 3 can be implemented by a controller.
  • the information is solved iteratively by MIMA algorithm, and the downlink data is pre-coded, and finally the downlink data is sent to the terminal through the signal transceiver.
  • a base station 100 is further provided, which may include a signal transceiver station 101 and a controller 102, where the signal transceiver station 101 may be configured to acquire channel information, and the controller 102 may be configured to Execute the MIMA algorithm and precoding the downlink data.
  • the signal transceiver station 101 is connected to the controller 102 to send the acquired channel information to the controller 102; the controller 102 is configured to execute operation instructions to implement the downlink precoding method, and control the signal transceiver station to respond to the downlink
  • the data is pre-encoded.
  • the controller 102 may have a built-in processor and a memory, wherein the memory may store a control program corresponding to the above-mentioned precoding method, the processor may call the corresponding control program from the memory, and pre-code the downlink data by executing the control program.
  • the processor may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP.
  • the processor may also further include a hardware chip.
  • the aforementioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof.
  • the above-mentioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a general array logic (generic array logic, GAL), or any combination thereof.
  • CPLD complex programmable logic device
  • FPGA field-programmable gate array
  • GAL general array logic
  • the memory may include volatile memory, such as random-access memory (RAM); the memory may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk Or solid state hard disk; memory can also include a combination of the above types of memory.
  • volatile memory such as random-access memory (RAM)
  • non-volatile memory such as read-only memory (ROM), flash memory, hard disk Or solid state hard disk
  • ROM read-only memory
  • the channel information corresponding to the terminal 200 may be obtained through the signal transceiver 101, and the channel information may be sent to the controller 102.
  • the processor in the controller 102 After the processor in the controller 102 receives the channel information, it can first call the MIMA algorithm-related application in the memory, and extract the initial weight and channel characteristic matrix in the channel information as the input of the MIMA algorithm program, so as to pass The MIMA algorithm is executed to generate power distribution weights.
  • the precoding related application program can be called from the memory, and the power allocation weight is used as the input of the precoding related application program to perform precoding on the downlink data. Finally, the precoding result is sent to the terminal 200 through the signal transceiver 101.
  • a communication device is also provided, and the communication device may be a terminal or a chip or a system on a chip in the terminal.
  • the communication device can implement the functions performed by the terminal in each of the foregoing aspects or in each possible implementation manner, and the functions can be implemented by hardware.
  • the communication device may include a processor and a communication interface, and the processor may be used to support the communication device to implement the aforementioned downlink precoding method.
  • a computer-readable storage medium may be a readable non-volatile storage medium, and the computer-readable storage medium stores instructions when it is stored in When running on a computer, the computer can execute the aforementioned downlink precoding method.
  • a computer program product containing instructions is also provided, which when running on a computer, enables the computer to execute the aforementioned downlink precoding method.
  • a communication device is also provided.
  • the communication device may be a terminal or a chip or a system on a chip in the terminal.
  • the communication device includes one or more processors and one or more memories.
  • the one or more memories are coupled with the one or more processors, and the one or more memories are used to store computer program codes, and the computer program codes include computer instructions.
  • the communication device is caused to execute the aforementioned downlink precoding method.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions may be transmitted from a website, computer, server, or data center. Transmission to another website, computer, server or data center via wired, such as coaxial cable, optical fiber, digital subscriber line, or wireless, such as infrared, wireless, microwave, etc.
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state hard disk).
  • These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device.
  • the device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment.
  • the instructions provide steps for implementing functions specified in a flow or multiple flows in the flowchart and/or a block or multiple blocks in the block diagram.

Abstract

本申请实施例提供一种下行预编码方法、装置及基站。所述下行预编码方法在获取下行信道的信道信息后，使用MIMA算法迭代生成下行信道的功率分配权值，其中，MIMA算法的目标函数为下行信道的互信息下界函数，MIMA算法的初始化参数为信道信息中的初始权值和信道特征矩阵；再按照功率分配权值对下行数据执行预编码实现对下行数据的传递。所述下行预编码方法以互信息下界函数为目标函数，通过迭代完成对功率和权值的计算，以便确定最优星座组合，在较小的迭代次数下，保证信道容量分配维持在离散调制星座时的全局最优解。

Description

一种下行预编码方法、装置及基站
本申请要求于2020年05月27日提交到国家知识产权局、申请号为202010460800.7、发明名称为“一种下行预编码方法、装置及基站”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及数字通信技术领域，尤其涉及一种下行预编码方法、装置及基站。
背景技术
预编码一个数字通信的信号处理环节，是在已知信道状态信息的情况下，在发送端对发送的信号做一个预先的处理。预编码环节可用于通过多个发射天线并行传输多个数据流，以提高峰值传输速率。在多输入多输出(multiple input multiple output，MIMO)场景中，如果天线数目和天线尺寸不变，会约束场景中信号传输的性能，使得单用户(single user，SU)和多用户(multiple user，MU)的容量受限。
为了提高信号的传输性能，可以通过调整预编码环节中下行权值，分配各信道的传输功率。典型预编码方法中，可以采用基于香农公式最大化的方式实现单用户的多输入多输出(single user-multiple input multiple output，SU-MIMO)和多用户的多输入多输出(multiple user-multiple input multiple output，MU-MIMO)的下行权值计算。即信号容量C＝Wlog[1+Ps/(N 0W)]；其中，W为下行权值，Ps为输入信号功率，N 0为噪声的双边功率谱密度。需保证1+Ps/(N 0W)最大化，以获得最大的容量。通过多次迭代后，求解并输出权值和功率分配结果，以达到最优频谱效率。但是，上述预编码方法需要进行多次迭代，导致求解全局最优解的复杂度很高，需要消耗大量的运算资源和时间。
发明内容
本申请提供了一种下行预编码方法、装置及基站，以解决传统预编码方法需要进行多次迭代的问题。
第一方面，本申请实施例提供一种下行预编码方法，所述方法包括：
获取下行信道的信道信息，所述信道信息包括下行信道的初始权值和下行信道的信道特征矩阵；
使用互信息最大化(mutual information maximization，MIMA)算法迭代生成下行信道的功率分配权值，所述MIMA算法的目标函数为下行信道的互信息下界函数，所述MIMA算法的初始化参数为所述初始权值和信道特征矩阵；
按照所述功率分配权值对下行数据执行预编码，所述下行数据为所述下行信道中传递的数据。
基站在向终端传递下行数据时，可以先获取终端对应下行信道的信道信息，并从下行信道信息中提取初始权值和信道特征矩阵，以便输入互信息最大化算法迭代生成下行信道的功率分配权值。在MIMA算法中可以设置目标函数为下行信道的互信息下界函数，初始化参数为初始权值和信道特征矩阵。通过基于初始权值和信道特征矩阵的MIMA算法迭代，可以最 终获得功率分配权值，从而按照功率分配权值对下行数据执行预编码，以将待传递的下行数据发送给终端。
采用上述预编码方法，可以在传递数据前根据下行信道的信道信息实时确定功率权值，实现对信道功率进行分配，能够减少功率浪费、缓解吞吐量损失。以及，通过将下行信道的互信息下界函数作为MIMA算法的目标函数进行优化求解，并结合梯度迭代简化迭代步骤，有效地保证最优的信道容量分配。
上述预编码方法中，使用的信道特征矩阵中包含能够表征当前下行信道特征的信息，在一种实现方式中，可以针对多用户复用场景在信道信息中对配对信息进行提取，以生成信道特征矩阵。即获取下行信道的信道信息的步骤可以包括：
获取配对信息，所述配对信息包括目标终端和配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种；
生成信道特征矩阵，所述信道特征矩阵中的元素为在所述配对信息中提取的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种。
通过信道特征矩阵，可以在MIMA算法中对迭代过程形成约束，以获得更符合当前下行信道特点的迭代结果。
结合上述下行预编码方法，在第一方面的一种可能的实现方式中，使用MIMA算法迭代生成下行信道的功率分配权值，包括：
先设置迭代梯度值为初始步长因子以及最小步长因子。再根据初始权值和信道特征矩阵生成初始迭代速率。
如果迭代梯度值大于最小步长因子，则更新权值矩阵。并根据更新后的权值矩阵和信道特征矩阵生成中间迭代速率。直到中间迭代速率大于或等于初始迭代速率时，更新所述迭代梯度值；如果迭代梯度值大于最小步长因子，则继续更新权值矩阵，并按照更新的权值矩阵计算中间迭代速率，重复迭代；如果所述迭代梯度值小于或等于所述最小步长因子，则提取权值矩阵，获得功率分配权值。
该实现方式提供一种基于MIMA算法的迭代方式，通过该迭代方式，能够以初始权值和信道特征矩阵为输入，通过多次迭代，计算迭代速率以将目标函数推演至互信息下界函数的最小值，从而获得迭代后的最优功率分配权值，进行相关预编码操作。
在迭代过程中，可以分别求得每次的中间迭代速率，通过将求得的中间迭代速率与初始迭代速率进行比较，以根据两者的大小关系确定是否完成迭代。其中，如果中间迭代速率小于初始迭代速率，则更新初始迭代速率为中间迭代速率，即可以根据更新后的初始迭代速率进行下一次迭代过程，直至中间迭代速率大于或等于初始迭代速率，更新迭代梯度值，以便在迭代梯度值小于或等于最小步长因子时结束迭代提取权值矩阵，获得功率分配权值。
所述初始迭代速率是每次迭代过程中计算的目标判断值，初始迭代速率可以按照下式计算生成：
R n＝G(H,W n)
其中，R n为初始迭代速率；
G(H，W)为目标函数，即互信息下界函数：
Figure PCTCN2021092744-appb-000001
H为信道特征矩阵；W n为初始权值矩阵；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
根据上述计算公式，可以分别求出每次迭代过程中的迭代速率，由于其中目标函数为互信息下界函数，并包含当前信道特征矩阵和功率分配权值，因此通过迭代计算使目标函数逐渐趋于下界值，并随着迭代过程逐渐更新权值矩阵。
所述权值矩阵是用于指导下行信道中信号功率分配的矩阵，矩阵中每个元素可以对应于每个信号传输过程的功率分配具体情况，因此，在一种实现方式中，可以采用下式更新权值矩阵：
Figure PCTCN2021092744-appb-000002
其中，α为预设影响系数；H为信道特征矩阵；W n为权值矩阵；
Figure PCTCN2021092744-appb-000003
为互信息下界函数的一阶导数；
Figure PCTCN2021092744-appb-000004
e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引；μ为迭代梯度值。
在迭代过程中可以根据上述公式对权值进行更新，为了获取最终采用的功率分配权值，可以在中间迭代速率大于或等于初始迭代速率时，更新所述迭代梯度值，重复迭代，直至迭代梯度值小于或等于最小步长因子时提取当前的权值矩阵，以获得所述功率分配权值。并按照输出的权值矩阵计算信道容量，以根据信道容量设置信号传输方式。可见，上述迭代过程可以通过类梯度迭代的方式简化迭代步骤和迭代过程的复杂度。
在一种实现方式中，可以按照下式更新所述迭代梯度值：
Figure PCTCN2021092744-appb-000005
其中，μ为迭代梯度值；k为收敛系数。通过对迭代梯度的更新，可以重复对迭代梯度值与最小步长因子进行判断的目的，并继续进行后续迭代过程，计算中间迭代速率以及对更新初始迭代速率，直到迭代梯度值小于或等于最小步长因子时，输出权值矩阵。
在上述MIMA算法中，需要将从信道信息中获取的初始权值输入MIMA算法，以开始首次迭代。在一种实现方式中，所述初始权值为所述下行信道中单用户权值；或者，所述下行信道的估计权值。单用户权值和估计权值可以通过历史信息或在下行信道的信道特征矩阵中获取，通过单用户权值或者估计权值，可以作为首次迭代的参考值，以便最终获得MIMA算法的输出。
在获得MIMA算法的输出后，可以将MIMA算法的输出作为预编码的输入，对下行数据执行预编码。因此，在一种实现方式中，按照所述功率分配权值对下行数据执行预编码包括：先根据所述功率分配权值和所述下行信道的信道容量生成分配功率值；再应用所述分配功率值至所述下行数据，以按照计算获得的功率和权值发送下行数据。
对于多用户复用场景，基站还可以生成包含配对信息的下行信令；并将下行信令发送给对应的终端，实现将配对信息分别发送给目标终端和配对终端，以便于终端可以根据配对信息对传输的数据进行均衡、检测和译码。其中，所述下行信令承载在无线资源控制信令和/或下行控制信息信令中。
第二方面，本申请实施例中还提供一种下行预编码装置，可以包括用于执行第一方面所提供下行预编码方法的获取模块、权值计算模块以及预编码模块。也可以包括用于 执行第一方面各实现方式中方法步骤的模块。
具体的，获取模块用户获取下行信道的信道信息，所述信道信息包括下行信道的初始权值和下行信道的信道特征矩阵；
权值计算模块用于使用MIMA算法迭代生成下行信道的功率分配权值，所述MIMA算法的目标函数为下行信道的互信息下界函数，所述MIMA算法的初始化参数为初始权值和信道特征矩阵；
预编码模块用于按照功率分配权值对下行数据执行预编码，所述下行数据为所述下行信道中传递的数据。
所述下行预编码装置可以通过获取模块获取信道信息，并输入给权值计算模块，使权值计算模块使用MIMA算法进行迭代求解最优解。在求解最优解以后，可以通过预编码模块对下行数据执行预编码。由于权值计算模块中所运行的MIMA算法的目标函数为下行信道的互信息下界函数，初始化参数为信道信息中的初始权值和信道特征矩阵，因此所述下行预编码装置能够根据信道特征确定获得最优功率分配方案，并简化迭代过程，快速获得最优的功率分配权值。
在一种实现方式中，所述权值计算模块所获取的信道信息中，所述初始权值为下行信道中单用户权值；或者，下行信道的估计权值。通过初始权值，可以作为MIMA算法首次迭代的参考，以便进行后续迭代。
在一种实现方式中，所述权值计算模块，具体用于设置迭代梯度值为初始步长因子以及最小步长因子；并根据初始权值和信道特征矩阵生成初始迭代速率；如果迭代梯度值大于最小步长因子，更新权值矩阵；根据更新后的权值矩阵和信道特征矩阵生成中间迭代速率；如果中间迭代速率大于或等于初始迭代速率，则更新所述迭代梯度值，以及在所述迭代梯度值小于或等于所述最小步长因子时提取权值矩阵，获得功率分配权值。
可见，上述实现方式中权值计算模块可以在MIMA算法中运用梯度迭代算法，简化迭代次数，缩短MIMA算法的迭代时间。
在一种实现方式中，所述权值计算模块具体用于如果中间迭代速率小于初始迭代速率，则更新初始迭代速率为中间迭代速率。即权值计算模块可以在中间迭代速率小于初始迭代速率时通过更新初始迭代速率继续进行下一次迭代，直至目标函数达到互信息下界。
在一种实现方式中，权值计算模块使用MIMA算法进行迭代求解最优解时，所述初始迭代速率可以按照下式计算生成：
R n＝G(H,W n)
其中，R n为初始迭代速率；G(H，W)为互信息下界函数；
Figure PCTCN2021092744-appb-000006
H为信道特征矩阵；W为初始权值矩阵；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
在一种实现方式中，权值计算模块使用MIMA算法进行迭代求解最优解时，可以按照下式更新权值矩阵：
Figure PCTCN2021092744-appb-000007
其中，α为预设影响系数；H为信道特征矩阵；W n为权值矩阵；
Figure PCTCN2021092744-appb-000008
为互信息 下界函数的一阶导数；μ为迭代梯度值；
Figure PCTCN2021092744-appb-000009
e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
在一种实现方式中，所述权值计算模块具体用于如果中间迭代速率大于或等于初始迭代速率，更新迭代梯度值，以通过调整迭代梯度值进行下一次迭代。如果迭代梯度值小于或等于最小步长因子，则输出权值矩阵；并按照输出的权值矩阵计算信道容量。即权值计算模块在确定迭代梯度值小于或等于最小步长因子时，完成迭代，并按照输出的权值矩阵计算信道容量，以便预编码模块根据输出结果执行预编码。
其中，权值计算模块可以按照下式更新所述迭代梯度值：
Figure PCTCN2021092744-appb-000010
其中，μ为迭代梯度值；k为收敛系数。
对于多用户复用的场景，在一种实现方式中，所述获取模块还用于获取配对信息并根据配对信息生成信道特征矩阵。配对信息包括目标终端和配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种；信道特征矩阵中的元素为在配对信息中提取的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种。获取模块通过在信道特征矩阵增加配对信息，可以互通目标终端与配对终端的信息发送方式，便于在目标终端与配对终端之间进行数据传输。
在一种实现方式中，所述预编码模块，具体用于根据功率分配权值和下行信道的信道容量生成分配功率值；再应用分配功率值至下行数据以按照计算获得的功率和权值发送下行数据。
对于多用户复用场景，所述预编码装置还包括配对模块，所述配对模块用于生成包含配对信息的下行信令，并发送下行信令给目标终端和配对终端。其中，所述下行信令可以承载在无线资源控制信令和/或下行控制信息信令中，以便于实现对下行数据的传输。
第三方面，本申请实施例中还提供一种基站，可以包括信号收发台和控制器，其中，所述信号收发台连接所述控制器；所述控制器被配置为执行操作指令，以实现第一方面所提供下行预编码方法，用于控制所述信号收发台对下行数据执行预编码。
第四方面，本申请实施例中还提供一种通信装置，该通信装置可以为终端或者终端中的芯片或者片上系统。该通信装置可以实现上述各方面或者各可能的实现方式中终端所执行的功能，所述功能可以通过硬件实现。该通信装置可以包括：处理器和通信接口，处理器可以用于支持通信装置实现上述第一方面或者第一方面的任一种可能的实现方式中所涉及的下行预编码方法。
第五方面，本申请实施例中还提供一种计算机可读存储介质，该计算机可读存储介质可以为可读的非易失性存储介质，该计算机可读存储介质中存储有指令，当其在计算机上运行时，使得计算机可以执行上述第一方面或者上述方面的任一种可能的实现方式中所述的下行预编码方法。
第六方面，本申请实施例中还提供一种包含指令的计算机程序产品，当其在计算机上运行时，使得计算机可以执行上述第一方面或者上述方面的任一种可能的实现方式中所述的下 行预编码方法。
第七方面，本申请实施例中还提供一种通信装置，该通信装置可以为终端或者终端中的芯片或者片上系统，该通信装置包括一个或者多个处理器以及和一个或多个存储器。所述一个或多个存储器与所述一个或多个处理器耦合，所述一个或多个存储器用于存储计算机程序代码，所述计算机程序代码包括计算机指令，当所述一个或多个处理器执行所述计算机指令时，使所述通信装置执行如上述第一方面或者第一方面的任一可能的实现方式中所述的下行预编码方法。
为了实现对下行数据的预编码，本申请实施例中，在获取下行信道的信道信息后，使用MIMA算法迭代生成下行信道的功率分配权值，其中，MIMA算法的目标函数为下行信道的互信息下界函数，MIMA算法的初始化参数为信道信息中的初始权值和信道特征矩阵；再按照功率分配权值对下行数据执行预编码实现对下行数据的传递。所述下行预编码方法以互信息下界函数为目标函数，通过迭代完成对功率和权值的计算，以便确定最优星座组合，在较小的迭代次数下，保证信道容量分配维持在离散调制星座时的全局最优解。
附图说明
图1为本申请实施例中MIMO系统场景示意图；
图2为本申请实施例中一种下行预编码方法的流程示意图；
图3为本申请实施例中功率和权值调整后星座组合效果示意图；
图4为本申请实施例中功率分配对比效果示意图；
图5为本申请实施例中一种MIMA算法的流程示意图；
图6为本申请实施例中多用户复用场景下获取信道信息的流程示意图；
图7为本申请实施例中多用户复用场景下执行预编码的流程示意图；
图8为本申请实施例中发送下行信令的流程示意图；
图9为本申请实施例中一种下行预编码装置的结构示意图。
具体实施方式
为便于本申请的技术方案进行说明，以下首先对本申请所用到的一些概念进行简要说明。
在本申请各个实施例中，下行信道是基站向终端发送数据时的信道，包括物理下行控制信道(physical downlink control channel，PDCCH)，和物理下行数据信道(physical downlink shared channel，PDSCH)(或者称为下行数据信道)。
物理下行控制信道，主要用于承载下行控制信息(downlink control information，DCI)，通常表现为DCI信令。DCI可以包括公共控制信息(如：系统信息等)和用户专属信息(如：下行资源分配指示，上行调度，随机接入响应，上行功率控制参数等)等。PDCCH可以通过其承载的DCI调度数据信道，如：DCI可以用于指示数据信道的传输参数(如：数据信道的时域资源位置等)，在传输数据信道之前，基站或其他网络设备可以向终端发送PDCCH，终端接收到PDCCH后，可以先解调PDCCH中的DCI，然后在DCI所指示的时域资源位置上接收或发送数据信道。而物理下行数据信道则用于承载从基站向终端发送的数据(或者称为下行数据)。
PDCCH也可以通过其承载的DCI指示信道状态信息参考信号(channel state information reference signal，CSI-RS)的时域资源位置，以触发非周期(nonperiodic)CSI-RS的发送。其中，CSI-RS用于使终端测量终端与基站间的信道状态，CSI-RS可以包括一个或者多个信道 状态测量资源。例如，基站可以向终端发送用于指示CSI-RS的时域资源位置的DCI以及CSI-RS，终端在DCI所指示的时域资源位置上接收CSI-RS，并对CSI-RS包括的信道状态测量资源进行测量，根据测量结果向基站上报信道状态信息(channel state information，CSI)。
基站也可以根据终端上报的CSI，对下行信道的数据发送方式进行设置，如设置最优星座组合方式。例如，对于多输入多输出系统(以下简称MIMO系统)，基站和终端都包括多个天线，可以在发送下行数据时形成多个数据流。而通过设置下行信道中各数据流的发送功率权值等，可以使基站能够按照设置的星座组合方式向终端发送下行数据，充分利用信道容量。
其中，星座即星座映射，是一种数字调制技术。星座映射的过程，就是将携带数字信息的有限域“比特”序列映射成适于传输的“符号”序列。每个符号的取值空间可以是一维实数空间，二维实数空间(即复数空间)。星座映射包含两个要素，即星座图(constellation)和星座点映射方法(labeling)。星座图代表星座映射输出符号的所有取值组成的集合，其中，星座图的每一个点对应输出符号的一种取值。星座点映射方式代表输入比特(序列/组)到星座点的特定映射关系，或者是星座点到比特(序列/组)的特定映射关系。通常，星座图主要有一维实数空间的脉冲幅度调制(pulse amplitude modulation，PAM)，二维实数空间的正交幅度调制(quadrature amplitude modulation，QAM)，相移键控(phase shift keying，PSK)调制等。
基站按照设置星座组合方式将下行数据进行编码、转化实现星座映射的过程称为预编码。而为了获得最优的预编码方式，基站可以对预编码中所依据的数据，例如功率分配矩阵，进行优化求解，从而获得对应的数据发送方式。
下面对基站设置下行信道的数据发送方式的流程进行说明。
本申请实施例中基站设置下行信道的数据发送方式可应用于MIMO模式的通信系统，如：第四代(4th generation，4G)系统、长期演进技术(long term evolution，LTE)系统、第五代(5th generation，5G)系统、新空口(new radio，NR)系统、NR-车与任何事物通信(vehicle-to-everything，V2X)系统中的任一系统，也可以适用于其他下一代通信系统等。
如图1所示，MIMO系统中包括基站100和多个终端200。在MIMO系统中，多个终端200可以与基站100之间建立通信连接。
在本申请各个实施例中，基站100主要用于实现终端200的资源调度、无线资源管理、无线接入控制等功能。所述基站100也可以是其他网络设备，例如：小型基站、无线接入点、收发点(transmission receive point，TRP)传输点(transmission point，TP)以及某种其它接入节点中的任一节点等。基站100中内置有天线，使基站100可以在运行中接收和发射无线电信号，并形成一个无线电覆盖区域。覆盖区域中的终端200可以通过与基站100之间的无线电信号收发关系连接互联网或者连接其他终端，实现移动通信业务。本申请实施例中，用于实现基站的功能的装置可以是基站，也可以是能够支持基站实现该功能的装置或者功能模块，例如芯片系统。
终端200可以为终端设备(terminal equipment)或者用户设备(user equipment，UE)或者移动台(mobile station，MS)或者移动终端(mobile terminal，MT)等。如：终端可以是手机(mobile phone)、平板电脑或带无线收发功能的电脑，也可以是虚拟现实(virtual reality，VR)终端、增强现实(augmented reality，AR)终端、工业控制中的无线终端、无人驾驶中的无线终端、远程医疗中的无线终端、智能电网中的无线终端、智慧城市(smart city)中的无线终端、智能家居、车载终端等。
在终端200接入基站100后，基站100会作为发送端向终端200发送数据，所发送的数 据称为下行数据，相应发送数据的通道称为下行信道。本申请实施例中，用于实现终端的功能的装置可以是终端，也可以是能够支持终端实现该功能的装置，例如芯片系统。
在MIMO系统中，基站100和终端200中都可以使用多个天线，由于使用多个天线，因此下行数据可以同时通过多个天线分别进行发送，形成多个数据流，从而提高下行数据的发送效率。同时，基站100也可以与多个终端200之间建立连接关系，构成多用户复用模式，以实现多个终端200之间的通信。
根据不同通信标准，基站中传输的信号形式也存在差异。例如，现行移动通信技术在基站100与终端200之间多采用数字信号进行传递，如LTE技术和NR技术。LTE技术和NR技术定义的信号为离散均匀分布信号，包括QPSK、16QAM、64QAM、256QAM等形式。通常，在多输入多输出的数字信号传递过程中，可采用单码字传输方式。例如，LTE技术在2流以上输出时采用单码字到多层映射的传输方式。而NR技术对于不大于4流的传输，也都采用单码子传输方式。
这种单码字传输方式在天线数目不变和天面尺寸约束的情况下，多输入多输出场景的单用户或多用户容量受限，已经接近理论香农公式的极限值，如果需要进一步提升性能，需要增加天线数目或者天面尺寸。
然而在部分使用场景下，即使增加了天线数目或者天面尺寸，由于针对下行数据进行预编码时所依据的权值矩阵(或预编码矩阵)依旧为根据香农公式确定的通用权值，导致功率分配与下行信道中各信号的传输特点不相符，因此，为每个下行数据流配置的星座组合并不是最优的组合方式。即表现为部分传输层的解调性能受限于星座间的最小码距，出现传输错误；而部分传输层被分配较高的功率而无法获得较高的吞吐量。
例如，通常终端的天线数目大于2，平均传输层数也大于2，并支持组播侦听发现协议(multicast listener discover，MLD)接收机或者Turbo接收机配合低密度奇偶校验码(low-density parity-check，LDPC)译码规则。因此根据信息论原理，最优权值设计与输入信号特征有关，则根据香农公式，在输入信号服从高斯分布时，等效信道容量最大化。
基于此，为了达到信道容量最大化效果，本申请求解最优权值方法的一个示意性实施方式如下：
S301：设置目标函数。
为了获得信道容量最大化的效果，可以将信道容量对应的函数设置为目标函数，并通过迭代求解，计算获得在信道容量最大化时，对应的功率权值。
例如，取互信息最大值：
Figure PCTCN2021092744-appb-000011
由于，最大信道容量为：
Figure PCTCN2021092744-appb-000012
因此，只要保证香农公式最大化即可保证信道容量最大，即：
Figure PCTCN2021092744-appb-000013
Figure PCTCN2021092744-appb-000014
其中，h(Y)为输出信号Y的熵信息，h(Y|X)为输入信号为X条件下，输出信号为Y的条件熵信息，P Y为输出信号Y的方差，P X为输入信号X的方差，P S为输入信号功率，W为系统带宽，N 0为白噪声的双边功率谱密度，N为带宽W下的计算互信息的样点数目。
利用下式：
Y＝X+n
则得到输入信号为X条件下，输出信号为Y的条件熵信息：
h(Y|X)＝h(n)
由于熵信息满足下式：
Figure PCTCN2021092744-appb-000015
如果假设高斯白噪声
Figure PCTCN2021092744-appb-000016
则期望和方差分别满足：
Figure PCTCN2021092744-appb-000017
Figure PCTCN2021092744-appb-000018
即在输出信号Y服从高斯分布时，满足下式：
Figure PCTCN2021092744-appb-000019
且：P Y＝P X+σ 2；
因此，可以采用基于香农公式最大化方式来实现MIMO系统的功率分配权值计算。对于单用户采用对信道自相关矩阵进行奇异值分解(singular value decomposition，SVD)，并利用SVD分解后的特征矢量来进行预编码，即：
[D,V]＝SVD(R)
结合香农公式：
Figure PCTCN2021092744-appb-000020
可知，保证信号干扰噪声比(signal to interference plus noise ratio，SINR)最大化，即可获得最优的容量。
为了能使SINR最大化，在白噪声场景下，只需要保证信号功率最大化。由此，可以获得单用户权值的最优值为取特征矢量对应的最大层数个特征值对应的特征矢量。
同理，对于多用户场景下的权值，为了保证SINR最大化，可以采用迫零的方式来减少干扰项的功率；或者，采用SINR和信漏噪比(signal to loss plus noise ratio，SLNR)最大化的方式获得信号功率和干扰功率的最优组合。
例如：基于SINR最大化的迫零权值：
W k＝[h 0,h 1,…,h K-1] + (:,k)
其中，h k(k＝0,1,…,K-1)为多用户中用户k的信道信息，[] +为求伪逆操作，(:,k)代表取矩阵的第k列，W k为计算出的第k个用户的权值。
基于SLNR最大化的权值计算：
Figure PCTCN2021092744-appb-000021
其中，R k为第k个用户信道自相关矩阵，R i为第i个用户的自相关矩阵，W k为计算出的第k个用户的权值。
可见，保证输出信号Y服从高斯分布，则要求输入信号的方差服从高斯分布。但在现行的LTE或NR通信系统内，发送的信号X并不服从高斯分布，而是服从离散均匀分布。因此，用香农公式求得的最大信道容量仅仅为次优解，而不是最优解。
S302：设置初始化迭代参数。
在设置目标函数后，可以根据下行信道中的信道信息，设置初始化迭代参数。例如，为了获得最优解可以采用功率控制和空间旋转的算法求解最优值，即：
设置矩阵元素p j属于集合
Figure PCTCN2021092744-appb-000022
其中，p j共K+1个可选值，(K+1) Rank种组合。
通过计算目标终端的信道信息的奇异值分解结果：
h＝*d*v′
结合计算调制分集矩阵：
Figure PCTCN2021092744-appb-000023
其中，
Figure PCTCN2021092744-appb-000024
在集合
Figure PCTCN2021092744-appb-000025
中选取方差σ j，j＝0,…,Rank-1，进行功率归一化操作，得到：
Figure PCTCN2021092744-appb-000026
假设Σ P的各个对角线元素p j属于集合
Figure PCTCN2021092744-appb-000027
其中K为量化粒度参数，对其(K+1) Nt种组合进行归一化，得到各个对角元素值：
Figure PCTCN2021092744-appb-000028
再计算各
Figure PCTCN2021092744-appb-000029
与Σ P之间的欧式距离，将
Figure PCTCN2021092744-appb-000030
最近的一组功率分配矩阵作为备选集。
通过选择适当的q MOD值，计算
Figure PCTCN2021092744-appb-000031
设置W＝V HΣ PV MOD。其中，p j为第j层调整后的功率值，Rank为层数，h为目标终端的信道信息。
S303：生成迭代结果。
经过迭代后，可以生成迭代结果，该迭代结果可以作为下行数据的功率分配权值。即，输出在目标函数即信道容量最大化时，对应的功率分配权值，并按照功率分配权值执行预编码，从而可以达到最优频谱效率。
通过上述采用迭代方法生成功率分配权值的过程可以看出，如果采用全局最优求解，通过迭代多次局部最优求解，计算获得全局最优的计算复杂度高，需要较长的迭代时间才能获得权值的最优解，不利于工程实现。也可以采用简单最优局部最优解求解，即采用功率控制 和空间旋转的思路进行求解，以降低工程实现过程中的复杂度，但是这种求解方式对信号传输的性能增益较小，并且在低信噪比条件下容易出现较大的负增益。
综上所述，如果采用香农公式求得的最大信道容量仅仅为次优解，而不是最优解；而通过功率控制和空间旋转算法进行全局最优求解的迭代复杂度较高，迭代次数也较多。因此，本申请实施例提供一种下行预编码方法，用于在对下行数据进行预编码的过程中，能够通过较少的迭代次数，获得全局最优解，以便根据功率分配权值对下行数据执行预编码操作。
下面结合附图，对本申请下行预编码方法进行说明。
如图2所示，本申请一个示意性实施例中提供的下行预编码方法，包括以下步骤：
S401：获取下行信道的信道信息。
在与用户终端建立连接时，基站可以获取下行信道的信道信息。例如，通过PDCCH向终端发送用于获取信道信息的DCI，以使终端对信道状态测量资源进行测量，并将测量结果反馈给基站。信道信息包括信道特征矩阵和初始权值。对于MIMO系统，由于包括多个终端，且每个终端设有多个天线，因此可以通过反馈结果组成一个信道特征矩阵。
信道特征矩阵中的每个元素用于反映对应终端与基站之间的信号传输特点，例如，信道特征矩阵中可以包括当前下行信道的调制与编码策略(modulation and coding scheme，MCS)信息、调制方案信息以及星座点类型信息等。信道特征数据可以利用特定量化方式，转化为具体的值，以体现为上述信道特征矩阵中的每个元素。
初始权值可以是单用户权值或者估计权值。其中，单用户权值是指在非MIMO场景下，通过下行信道与终端进行信号传递时的权值。估计权值是指根据场景中终端的接入特点，如终端数量、天线数量以及天面尺寸，结合信道容量进行预估所获得的权值。
在步骤S401中获取的信道信息，可以作为MIMA算法的初始化参数，以使用MIMA算法迭代求解功率分配权值的最优解。
S402：以信道信息作为初始化参数，使用MIMA算法迭代生成下行信道的功率分配权值。
在获取信道信息后，基站可以通过调用MIMA算法进行功率分配权值的求解计算。为了使用MIMA算法迭代求解功率分配权值，调用的MIMA算法的目标函数为下行信道的互信息下界函数，同时将信道信息中的初始权值和信道特征矩阵输入至MIMA算法中，以执行MIMA算法进行迭代优化求解。
其中，互信息下界函数具体表示如下：
根据互信息公式：
Figure PCTCN2021092744-appb-000032
其中，
Figure PCTCN2021092744-appb-000033
其中，码距e ij＝x i-x j。利用杰森(Jensen)不等式，求解互信息下界，即：
Figure PCTCN2021092744-appb-000034
由于：
Figure PCTCN2021092744-appb-000035
因此：
Figure PCTCN2021092744-appb-000036
将f(H,W)代入I(y；x)，可得互信息下界：
Figure PCTCN2021092744-appb-000037
其中：
Figure PCTCN2021092744-appb-000038
其中，G(H，W)为互信息下界函数；H为信道特征矩阵；W n为初始权值矩阵；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
因此，互信息最大化问题，可以转换成G(H,W)最小化的问题，即：
Figure PCTCN2021092744-appb-000039
相应的，关于互信息最大化问题的数学模型可等价为互信息下界函数G(H,W)最小化的问题，即设置MIMA算法的目标函数如下：
Figure PCTCN2021092744-appb-000040
可见，在上述的一种可行的实施方式中，可将互信息最大化问题转化为互信息下界问题，以进行功率分配和权值计算。由于互信息下界函数G(H,W)的迭代计算复杂程度相对于互信息最大化函数的迭代计算复杂程度更低，因此可以简化迭代过程中的计算复杂程度，即降低工程实现中的复杂度，有利于快速确定信道容量在离散调制星座时的全局最优解。
在上述互信息下界函数G(H,W)中，由于H为信道特征矩阵，即在步骤S401中获取的信道特征矩阵，W为权值矩阵，会随着后续迭代过程不断更新。为了完成首次迭代，可以将W设置为初始权值，以作为前几次迭代计算的判断约束值。可见，通过将信道特征矩阵和初始权值作为MIMA算法的输入，可以获得符合当前信道特点的迭代输出结果，即最优解状态下的功率分配权值矩阵。
在确定上述互信息下界函数G(H,W)后，可以针对互信息下界函数G(H,W)进行迭代 优化求解。在一种实现方式中，为了简化后续迭代过程，也可以利用梯度迭代法更新权值矩阵W n，即：
Figure PCTCN2021092744-appb-000041
其中，α为预设影响系数；H为信道特征矩阵；W n为权值矩阵；μ为迭代梯度，
Figure PCTCN2021092744-appb-000042
为互信息下界函数G(H,W)关于权值W的一阶导数，因此：
Figure PCTCN2021092744-appb-000043
即：
Figure PCTCN2021092744-appb-000044
其中，e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
基于上述公式，可以互信息下界函数G(H,W)为目标函数，以信道特征矩阵H和初始权值W 1作为迭代初始化参数，进行梯度迭代，以获得目标函数G(H,W)最小化的情况下对应的权值矩阵W。
S403：按照所述功率分配权值对下行数据执行预编码。
获得MIMA算法的迭代计算输出的权值矩阵W后，可以将步骤S402的输出结果作为预编码的输入，对下行数据执行预编码。由于基站与终端之间可以利用多层星座组合的方式进行数据发送，因此可以利用迭代求解获得的功率分配权值与数据的组合，实现最优多层星座的组合，并且保证组合后星座可以获得最优吞吐量。
例如，可利用多层星座的组合进行数据发送，如图3所示，通过上述实施例中提供的MIMA算法进行求解计算，可以保证组合后的星座方式能够获得最优吞吐量。如在图3中，两个QPSK星座通过星座汇聚组合成离散高斯化星座，更符合高斯分布。
又例如，如图4所示，对于一种情况，在两层传输时，如果基站利用传统方法给层0(Layer0)分配功率P0，给层1(Layer1)分配功率P1，由于传输的数据为离散数据，解调性能受限于星座间的最小码距，层1分配的功率P1较小，因此容易导致层1传输错误。而利用上述实施例中提供的MIMA算法后，由于假设数据为离散数据，可以给层0分配功率P0+P1，而给层1分配0功率，从而保证最优的吞吐量性能。
对于另一种情况，在两层传输时，利用传统方法给层0分配功率P0+P1，给层1分配P2，由于数据离散特性，给层0分配更高的功率无法获得更高的吞吐量，从而导致功率浪费和吞吐量损失。而利用上述实施例中提供的MIMA算法后，由于假设数据为离散数据，可以给层0分配功率P0，L1分配功率P0+P1，从而保证最优的吞吐量性能。
下面结合附图对前述实施例中所示的MIMA算法方案进行进一步说明。
在一种实现方式中，如图5所示，步骤2所示的使用MIMA算法迭代生成下行信道的功率分配权值的步骤，具体可以包括如下步骤：
S4021：设置迭代梯度值为初始步长因子以及设置最小步长因子。
在开始迭代前，需要对迭代参数进行设置，初始步长因子和最小步长因子可以根据实际信号传输需要进行设置，即设置初始化权值矩阵W 1为信道特征矩阵，同时，设置初 始步长因子μ＝μ init，设置最小步长因子μ min。通过设置迭代梯度值，可以确定后续迭代过程的初始值。在后续迭代过程中，可以按照初始步长因子开始迭代，并通过与最小步长因子进行比较，确定完成迭代的时机。
S4022：根据所述初始权值和信道特征矩阵生成初始迭代速率。
在设置迭代梯度值以后，可以开始首次迭代，即根据初始权值和信道特征矩阵生成初始迭代速率R n＝G(H,W n)，其中，所述初始迭代速率按照下式计算生成：
R n＝G(H,W n)
式中，R n为初始迭代速率；G(H，W)为互信息下界函数，即：
Figure PCTCN2021092744-appb-000045
其中，H为信道特征矩阵；W n为初始权值；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
例如，在首次迭代过程中，可以直接在互信息下界函数G(H,W)输入当前下行信道的初始权值W 1和信道特征矩阵H，并计算初始迭代速率R 1＝G(H,W 1)。计算初始迭代速率后，则进入迭代步骤中的判断环节，在上述示意性实施例中，需要分别关于迭代梯度值和迭代速率进行两次判断。
S4023：如果所述迭代梯度值大于所述最小步长因子，更新权值矩阵。
在生成初始迭代速率后，可以对迭代梯度值与最小步长因子进行比较，从而确定是否已完成迭代。如果迭代梯度值大于最小步长因子，则未完成本次迭代过程，因此可以更新权值矩阵W n并继续进行下一轮迭代。如果迭代梯度值小于或等于最小步长因子，则确定已完成本次迭代，因此可以输出此时的权值矩阵，以便按照输出的权值矩阵，计算信道容量。
对于首次迭代过程，由于设置初始步长因子μ＝μ init，所设置的初始步长因子必然大于设置的最小步长因子μ min，因此在首次迭代过程中，迭代梯度值大于最小步长因子，即进行更新权值矩阵。
在一种实现方式中，可以按照下式更新权值矩阵：
Figure PCTCN2021092744-appb-000046
其中，μ为迭代梯度值；α为预设的影响系数；
Figure PCTCN2021092744-appb-000047
为互信息下界函数G(H,W)关于权值W的一阶导数，即：
Figure PCTCN2021092744-appb-000048
式中，H为信道特征矩阵；W为初始权值；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
例如，对于首次迭代过程，由于判断迭代梯度值大于最小步长因子，需要更新权值矩阵。因此可以先计算首次迭代过程中的
Figure PCTCN2021092744-appb-000049
再根据计算的
Figure PCTCN2021092744-appb-000050
更新权值矩阵，即
Figure PCTCN2021092744-appb-000051
S4024：根据更新后的权值矩阵和所述信道特征矩阵生成中间迭代速率。
在对权值矩阵进行更新后，以更新后的权值矩阵W n+1和当前下行信道的信道特征矩阵H为输入，再次根据迭代速率的计算公式计算中间迭代速率，即：
R n+1＝G(H,W n+1)
例如，对于首次迭代过程中获得的更新后的权值矩阵W 2可以将更新后的权值矩阵W 2和信道特征矩阵H输入R n＝G(H,W n)计算中间迭代速率R 2，以便进行后续对迭代速率的判断，从而实现迭代过程。
在计算出中间迭代速率后，可以对中间迭代速率与初始迭代速率进行比较，从而根据比较结果确定是否已完成本次迭代。
S4025：如果所述中间迭代速率大于或等于所述初始迭代速率，则更新所述迭代梯度值，以及在所述迭代梯度值小于或等于所述最小步长因子时提取权值矩阵，获得所述功率分配权值。
通过中间迭代速率与初始迭代速率的比较，如果中间迭代速率大于或等于初始迭代速率，则说明当前互信息下界函数已经接近最小化结果，因此可以更新迭代梯度值，继续进行迭代，使互信息下界函数逐渐逼近最小值，直至迭代梯度值小于或等于最小步长因子时提取对应的权值矩阵，以获得当前下行信道中各信号的功率分配权值。
在与步骤S4025并列的另一种比较结果下，如果中间迭代速率小于初始迭代速率，则说明当前互信息下界函数G(H,W)还未获得最小化结果。因此可以更新初始迭代速率为中间迭代速率，并跳转至步骤S4023继续进行迭代，直到迭代过程中计算的中间迭代速率大于或等于上一次迭代计算的初始迭代速率，输出对应的权值矩阵。
例如，对于首次迭代过程对应计算的中间迭代速率R 2，如果通过对比确定中间迭代速率R 2＜R 1，则确定当前互信息下界函数R 1＝G(H,W 1)未获得最小化结果，因此，可以将初始迭代速率设置为R 2，并跳转执行步骤S4023继续进行迭代。
对于步骤S4025，如果中间迭代速率大于或等于所述初始迭代速率，可以通过梯度迭代的方式，更新迭代梯度值，以进一步判断是否获得互信息下界函数G(H,W)的最小化结果。即所述MIMA算法可以包括：如果中间迭代速率大于或等于初始迭代速率，更新迭代梯度值。
可以按照下式更新所述迭代梯度值：
Figure PCTCN2021092744-appb-000052
其中，μ为迭代梯度值；k为收敛系数。
在迭代过程中，可以通过设置的收敛系数k对迭代梯度值进行更新，通常可以设置收敛系数k＝3。在更新迭代梯度值以后，可以跳转至步骤S4023，继续对迭代梯度值进行判断，以根据判断结果再进行迭代，以及对权值矩阵进行更新。
例如，在对于首次迭代过程对应计算的中间迭代速率R 2，如果通过对比确定首次迭代的中间迭代速率R 2≥R 1，则可以对迭代梯度值进行更新，即μ＝1/3μ init，相应的在对更新后的迭代梯度值与设置的最小步长因子继续进行比较。如果更新后的迭代梯度值μ仍然大于最小步长因子μ min，则进行继续权值矩阵进行更新。
同理，在对权值矩阵进行更新后，也可以通过执行步骤S4024，计算中间迭代速率，以及通过执行步骤S4025对中间迭代速率与初始迭代速率进行比较，实施反复迭代，直到中间迭代速率大于或等于初始迭代速率，并且迭代梯度值小于设置的最小步长因子，确定完成迭代，输出最终的权值矩阵。最后从权值矩阵中提取对应的元素，作为当前下行信道中每个信号的功率分配权值，实现对下行数据执行预编码。
可见，在上述实现方式中，通过将MIMA算法中的目标函数设置为采用互信息下界 函数，进行功率和权值的优化求解，可以简化每一次迭代中的计算复杂度。并相应利用了杰森不等式、类梯度迭代、类牛顿迭代相结合的迭代方式，减少迭代次数，简化迭代步骤复杂度，使得上述MIMA算法能够在工程中进行应用。
下面结合一个具体的实施示例，对上述MIMA算法的迭代过程进行描述，例如：
获取到信道信息后，可以提取信道信息中的初始权值W 1和信道特征矩阵H，并将初始权值初始权值W 1和信道特征矩阵H作为MIMA算法的初始化参数。同时设置初始步长因子μ＝μ init，设置最小步长因子μ min。
根据初始权值W 1和信道特征矩阵H计算初始迭代速率R 1＝G(H,W 1)，通过对比，确定初始步长因子μ init大于最小步长因子μ min，则更新权值矩阵
Figure PCTCN2021092744-appb-000053
再根据更新后的权值矩阵W 2和所述信道特征矩阵H生成中间迭代速率R 2＝G(H,W 2)。
再通过对比，确定中间迭代速率R 2小于初始迭代速率R 1，则将初始迭代速率设置为中间迭代速率R 2继续进行迭代，将权值矩阵W 2作为初始权值，并与信道特征矩阵H一同作为MIMA算法的初始化参数。
此时，由于初始步长因子μ init大于最小步长因子μ min，因此需要更新权值矩阵
Figure PCTCN2021092744-appb-000054
并计算中间迭代速率R 3＝G(H,W 3)。再通过对比，确定中间迭代速率R 3小于初始迭代速率R 2，则将初始迭代速率设置为中间迭代速率R 3继续进行迭代，将权值矩阵W 3作为初始权值，并与信道特征矩阵H作为MIMA算法的初始化参数。
同理，依次通过迭代计算获得权值矩阵W 4、W 5、W 6……，以及中间迭代速率R 4、R 5、R 6……，并进行对比，直到中间迭代速率R n大于或等于初始迭代速率R n-1时，更新迭代梯度值μ’＝1/3μ。
再通过对比，确定步长因子迭代梯度值μ’大于最小步长因子μ min，再通过更新权值矩阵
Figure PCTCN2021092744-appb-000055
以及计算中间迭代速率R n+1＝G(H,W n+1)，并继续进行对比以及迭代计算，获得权值矩阵W n+2、W n+3、W n+4……，以及中间迭代速率R n+2、R n+3、R n+4……，至中间迭代速率R n+m大于或等于初始迭代速率R n+m-1时，更新迭代梯度值μ”＝1/3μ’。
通过对比，如果确定步长因子迭代梯度值μ”小于或等于最小步长因子μ min，则完成迭代，输出此时的权值矩阵W n+m。再根据输出的权值矩阵计算信道容量，从而确定每个信号对应分配的权值以及功率。
通过实际检测分析可知，上述实现方式可以在经过十几次或几十次迭代即可求解出全局最优解，相对于学术理论中的成千上万的迭代次数可以大大简化迭代步骤复杂度，从而节省求解时间，使得上述实现方式可以应用在下行数据的实时预编码过程中。即可以有针对性的对当前下行信道进行最优求解，并按照求解结果对下行数据执行预编码，达到既不降低信号传输效率，又能获得最优吞吐量的目的，提高基站在MIMO系统场景下的下行数据发送效率，改善信号传输质量。
为了将下行数据发送给终端，需要按照计算获得的功率分配权值对下行数据执行预编码，如图6所示，在一种实现方式中，步骤S403示出的按照功率分配权值对下行数据执行预编码可以包括：
S4031：根据所述功率分配权值和所述下行信道的信道容量生成分配功率值。
通过迭代获得功率分配权值后，基站可以结合当前下行信道的信道容量，计算信号的分配功率值。其中，分配功率值是指对于对应数字信号的发送功率，不同类型的信号所需要的发送功率不同，相应多层星座的组合方式也不同。
S4032：应用所述分配功率值至所述下行数据。
根据功率分配权值确定的分配功率值，可以利用计算获得的最优解获得下行数据的最佳发送方式，从而合理利用信道容量。因此，在生成分配功率值后，可以将分配功率值应用至下行数据，以形成下行数据流，实现按照计算获得的最优功率和权值将下行数据发送至终端，最大化应用信道容量。
上述实施例提供的下行预编码方法，可以应用于单用户的多输入多输出系统中，也可以应用在多用户的多输入多输出系统中，均能够在传递下行数据前获得符合当前下行信道特点的功率分配权值，以获得最优的吞吐量性能。
对于多用户的MIMO系统，由于系统中包括多个用户终端，多个用户终端之间可以通过基站进行相互通信，因此在一个终端(目标终端)配对另一个终端(配对终端)时，两个终端之间产生的通信数据也可以在基站中产生下行数据。例如，终端A在与终端B配对后，终端A向终端B发送的数据，会先发送至基站；基站再将对应数据作为下行数据通过下行信道发送给终端B。
需要说明的是，下行数据是指从基站发送给终端的数据，在多用户的MIMO场景中，所述下行数据不仅包括配对终端向目标终端发送的通信数据，还包括发送给目标终端的其他数据，例如互联网发送给目标终端的数据等。并且，所述配对关系不局限于两个终端之间的配对关系，还可以是多个终端之间的配对关系。
在传输具有配对关系的终端对应的数据时，基站可以获取与配对终端有关的信道特征作为信道信息，以形成信道特征矩阵。因此，如图7所示，为了适应多用户复用场景，步骤S401示出的获取下行信道的信道信息的步骤可以包括：
S4011：获取配对信息。
所述配对信息用于表示具有配对关系的终端对应信道特征数据，可以包括目标终端和配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种。
其中，调制与编码策略信息通常表现为MCS索引值，可用来实现LTE中速率的配置。MCS信息中可以将所关注的影响通讯速率的因素作为表的列，将MCS索引值作为行，形成一张速率表。其中，每一个MCS索引其实对应一组参数下的物理传输速率。
调制方案信息用于表征当前基站传输数字信号时的数字调制参数，例如载波信号的参量以及信源编码、加密和均衡等相关的参数等。不同基站、终端甚至下行信道可采用不同形式的调制方案，通过调制方案信息可以确定下行信道数据的调制/解调方式，以便传输的信号拥有更强的抗干扰性能和抗信道损耗性能，以及拥有更好的安全性。
星座点类型信息是用于表征数据映射状态的数据，由于数字通信领域中，会将数字信号在复平面上表示，以直观的表示信号以及信号之间的关系。因此数据经过信道编码后，被映射在星座图上，星座点的类型不同，将影响数据在星座图上的映射结果。
对于每个终端，其在连接至基站后，根据不同的硬件特点，分别对应有不同的信道特征数据。而在信道特征数据中，可以包含有调制与编码策略信息、调制方案信息以及星座点类型信息等，因此配对信息可以分别在目标终端和配对终端对应的信道特征数据中提取获得。
S4012：生成信道特征矩阵。
在提取配对信息后，基站可以利用配对信息生成信道特征矩阵。相对于上述实施例，本实施例中可以在原本信道特征矩阵的基础上，在所述配对信息中提取相应的数据，如 调制与编码策略信息、调制方案信息以及星座点类型信息的一种或多种，并将提取的数据添加至信道特征矩阵。
上述三种信息可以通过组合形成一个用于表征下行信道特点的信道特征矩阵。显然，信道信息并不局限于上述三种，也可以包括其他类型的信息，例如天线数量、天面尺寸、信道容量等，可以根据实际信号传输过程所考虑的重点因素，构成不同的信道信息。
因此，通过上述信道特征矩阵的生成方法，可以在生成的信道特征矩阵中增加配对终端相关的信道特征。在两个相互配对的终端之间发生通信时，可以使得功率计算和权值计算的结果中具有与之配对的终端信道特征，并且相互配对的两个终端可以利用相应的信道特征进行均衡、检测和译码，从而提高两个配对终端之间的信号传输效率。
对于上述配对信息，一方面可以作为信道信息形成信道特征矩阵，另一方面还可以通过基站发送至具有配对关系的各终端中。如果多用户之间能够相互获知对应的MCS、调制方案、星座点类型等与数据传输相关的信息，则可以实现在配对后直接采用相适应的数据处理方式。例如可以采用相同的调制/解调方式等，从而减少终端无法解调数据或者采用盲解方法进行解调的问题，能够加快终端对数据的均衡、检测和译码过程，便于实现达到信道容量最大值的目的。
因此，在多用户的MIMO系统中，基站可以向目标终端和配对终端分别发送信令，以通知目标终端和配对终端相互获得信道特征，即如图8所示，在一种实现方式中，所述下行预编码方法还包括：
S404：发送下行信令。
对于多用户复用场景，基站可以针对目标用户和配对用户生成下行信令，并且可以向目标终端和配对终端分别发送下行信令，以便将配对信息通知到目标终端和配对终端。即目标终端通过接收下行信令获得基站通知的配对终端对应的调制与编码策略信息、调制方案信息以及星座点类型信息等配对信息，然后利用接收到的配对信息进行均衡、检测和译码。同理，配对终端也可以通过接收下行信令获得基站通知的目标终端对应的配对信息，并利用配对信息进行均衡、检测和译码。其中，下行信令可以包括诸如MCS、调制方案、星座点类型等配对信息。
下行信令可以直接被终端进行解析，从而可以在下行信令中提取配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息等配对信息。在一种实现方式中，所述下行信令承载在无线资源控制信令(radio resource control，RRC)，和/或DCI信令中。
其中，RRC又称为无线资源管理(radio resource management，RRM)或者无线资源分配(radio resource assignment RRA)，是指通过一定的策略和手段进行无线资源管理、控制和调度，在满足服务质量的要求下，充分利用有限的无线网络资源，确保到达规划的覆盖区域，提高业务容量和资源利用率。
相应的RRC信令即在基站实施无线资源管理、控制和调度时的信令，该信令可以直接在终端与基站之间进行传输，无需通过下行数据的方式进行发送。同样，DCI信令可以由下行物理控制信道承载，专用于向终端发送下行控制信息，包括上下行资源分配、混合自动重传请求信息、功率控制等，也无需通过下行数据的方式发送。
终端对下行信令可以承载在RRC信令中或者DCI信令或者通过RRC信令和DCI信令的组合进行接收。终端可以利用对应下行信令，进行均衡、检测和译码过程，从而获得最优的功率分配和吞吐量性能。
上述实施例中所提供的下行预编码方法以及各种实现方式中提供的下行预编码方法或方法所包含的步骤之间，可以相互结合以获得更多的预编码方法的实现方式，此处不再赘述。
基于上述实施例提供的下行预编码方法，在本申请的一种示意性实施方式中，还提供一种下行预编码装置。所述下行预编码装置可以用于实施上述实施例中提供的下行预编码方法。如图9所示，在一种实现方式中，所述下行预编码装置包括：获取模块1、权值计算模块2以及预编码模块3，分别用于执行上述实施例中的步骤S1、步骤S2以及步骤S3，以便对下行数据执行预编码。
例如，获取模块1用于获取下行信道的信道信息，并将信道信息发送给权值计算模块2，其中，所述信道信息包括下行信道的初始权值和下行信道的信道特征矩阵。
权值计算模块2用于接收获取模块1发送的信道信息，并使用MIMA算法迭代生成下行信道的功率分配权值。其中，所述MIMA算法的目标函数为下行信道的互信息下界函数，所述MIMA算法的初始化参数为所述初始权值和信道特征矩阵。权值计算模块2在计算获得功率分配权值后，也可以将功率分配权值发送给预编码模块3。
预编码模块3用于接收功率分配权值，并按照所述功率分配权值对下行数据执行预编码。
可见，上述下行预编码装置，可以在下行数据发送前，先通过获取模块1获得信道信息。再通过权值计算模块2执行MIMA算法，并根据信道信息迭代计算求解功率和权值的最优解，获得功率分配权值。最后通过预编码模块3对下行数据按照功率分配权值执行预编码，确定多层星座的最优组合方式，即获得最优吞吐量。
其中权值计算模块2为了执行MIMA算法，也可以通过具体执行以下操作，以获得功率分配权值：先设置迭代梯度值为初始步长因子以及最小步长因子；再根据初始权值和信道特征矩阵生成初始迭代速率；如果迭代梯度值大于最小步长因子，更新权值矩阵；然后根据更新后的权值矩阵和信道特征矩阵生成中间迭代速率；如果中间迭代速率大于或等于初始迭代速率，则提取权值矩阵，获得功率分配权值。
综上所述，本申请实施例提供一种下行预编码方法，可在获取下行信道的信道信息后，使用MIMA算法迭代生成下行信道的功率分配权值，再按照功率分配权值对下行数据执行预编码实现对下行数据的传递。所述下行预编码方法以互信息下界函数为目标函数，通过迭代完成对功率和权值的计算，以便确定最优星座组合，在较小的迭代次数下，保证信道容量分配维持在离散调制星座时的全局最优解。
对于上述实施例中的其他实现方式，可以在上述下行预编码装置的基础上，通过配置不同的功能单元实施相对应的实现方式，此处不再赘述。
需要说明的是，上述实现方式中的下行预编码装置中各模块的划分仅仅是一种逻辑功能的划分，在实际实现过程中可以全部或部分集成到一个物理实体上，也可以在物理上分开。例如，所述获取模块1可以由信号收发器实现，而权值计算模块2以及预编码模块3可以由控制器实现，即信号收发器获取信道信息并发送给控制器，使得控制器再根据信道信息执行MIMA算法进行迭代求解，并对下行数据执行预编码，最后再通过信号收发器将下行数据发送给终端。
为此，在一种示意性实施方式中，还提供一种基站100，可以包括信号收发台101和控制器102，其中，信号收发台101可以被配置为获取信道信息，控制器102被配置为执行MIMA 算法以及对下行数据的预编码。信号收发台101连接控制器102，以向控制器102发送所获取的信道信息；所述控制器102被配置为执行操作指令，以实现所述下行预编码方法，控制所述信号收发台对下行数据执行预编码。
控制器102可以内置处理器和存储器，其中存储器中可以储存有上述预编码方法对应的控制程序，处理器可以从存储器中调用对应的控制程序，并通过执行该控制程序对下行数据进行预编码。其中，处理器可以是中央处理器(central processing unit，CPU)，网络处理器(network processor，NP)或者CPU和NP的组合。处理器也可以进一步包括硬件芯片。上述硬件芯片可以是专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。
上述PLD可以是复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)或其任意组合。
存储器可以包括易失性存储器，例如随机存取存储器(random-access memory，RAM)；存储器也可以包括非易失性存储器，例如只读存储器(read-only memory，ROM)，快闪存储器，硬盘或固态硬盘；存储器也可以包括上述种类的存储器的组合。
例如，当基站100中有终端200接入时，可以通过信号收发台101获取终端200对应的信道信息，并将信道信息发送给控制器102。控制器102中的处理器在接收到信道信息后，可以在存储器中先调用MIMA算法相关的应用程序，并提取信道信息中的初始权值和信道特征矩阵作为该MIMA算法程序的输入，从而通过执行该MIMA算法，生成功率分配权值。
在生成功率分配权值后，可以再从存储器中调用预编码相关应用程序，并以功率分配权值作为预编码相关应用程序的输入，对下行数据执行预编码。最后将预编码结果再通过信号收发台101发送给终端200。
对于上述实施例中的其他实现方式，只需要在基站100的存储器中相应存储特定的应用程序。并在达到相应条件时，由处理器直接调用执行该应用程序即可实施其他实现方式，此处不再赘述。
在一个示意性实施例中，还提供一种通信装置，该通信装置可以为终端或者终端中的芯片或者片上系统。该通信装置可以实现上述各方面或者各可能的实现方式中终端所执行的功能，所述功能可以通过硬件实现。该通信装置可以包括：处理器和通信接口，处理器可以用于支持通信装置实现上述下行预编码方法。
在一个示意性实施例中，还提供一种计算机可读存储介质，该计算机可读存储介质可以为可读的非易失性存储介质，该计算机可读存储介质中存储有指令，当其在计算机上运行时，使得计算机可以执行上述下行预编码方法。
在一个示意性实施例中，还提供一种包含指令的计算机程序产品，当其在计算机上运行时，使得计算机可以执行上述下行预编码方法。
在一个示意性实施例中，还提供一种通信装置，该通信装置可以为终端或者终端中的芯片或者片上系统，该通信装置包括一个或者多个处理器以及和一个或多个存储器。所述一个或多个存储器与所述一个或多个处理器耦合，所述一个或多个存储器用于存储计算机程序代码，所述计算机程序代码包括计算机指令，当所述一个或多个处理器执行所述计算机指令时，使所述通信装置执行如上述下行预编码方法。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当 使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本发明实施例所述的流程或功能。
所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线，例如同轴电缆、光纤、数字用户线，或无线，例如红外、无线、微波等，方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘)等。
本发明实施例是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然，本领域的技术人员可以对本发明实施例进行各种改动和变型而不脱离本申请的精神和范围。这样，倘若本发明实施例的这些修改和变型属于本申请权利要求及其等同技术的范围之内，则本申请也意图包含这些改动和变型在内。

Claims (27)

  1. 一种下行预编码方法，其特征在于，包括：
    获取下行信道的信道信息，所述信道信息包括下行信道的初始权值和下行信道的信道特征矩阵；
    使用互信息最大化算法迭代生成下行信道的功率分配权值，所述互信息最大化算法的目标函数为下行信道的互信息下界函数，所述互信息最大化算法的初始化参数为所述初始权值和信道特征矩阵；
    按照所述功率分配权值对下行数据执行预编码，所述下行数据为所述下行信道中传递的数据。
  2. 根据权利要求1所述的下行预编码方法，其特征在于，使用互信息最大化算法迭代生成下行信道的功率分配权值，包括：
    设置迭代梯度值为初始步长因子以及最小步长因子；
    根据所述初始权值和信道特征矩阵生成初始迭代速率；
    如果所述迭代梯度值大于所述最小步长因子，更新权值矩阵；
    根据更新后的权值矩阵和所述信道特征矩阵生成中间迭代速率；
    如果所述中间迭代速率大于或等于所述初始迭代速率，则更新所述迭代梯度值，以及在所述迭代梯度值小于或等于所述最小步长因子时提取权值矩阵，获得所述功率分配权值。
  3. 根据权利要求2所述的下行预编码方法，其特征在于，使用互信息最大化算法迭代生成下行信道的功率分配权值，包括：
    如果所述中间迭代速率小于所述初始迭代速率，则更新所述初始迭代速率为所述中间迭代速率。
  4. 根据权利要求2所述的下行预编码方法，其特征在于，所述初始迭代速率按照下式计算生成：
    R n＝G(H,W n)；
    其中，R n为初始迭代速率；G(H，W)为互信息下界函数，
    Figure PCTCN2021092744-appb-100001
    Figure PCTCN2021092744-appb-100002
    H为信道特征矩阵；W n为初始权值矩阵；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
  5. 根据权利要求4所述的下行预编码方法，其特征在于，按照下式更新权值矩阵：
    Figure PCTCN2021092744-appb-100003
    其中，α为预设影响系数；H为信道特征矩阵；W n为权值矩阵；
    Figure PCTCN2021092744-appb-100004
    为互信息下界函数的一阶导数；
    Figure PCTCN2021092744-appb-100005
    e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引；μ为迭代梯度值。
  6. 根据权利要求2所述的下行预编码方法，其特征在于，根据所述初始权值和信道特征矩阵生成初始迭代速率后，包括：
    如果所述迭代梯度值小于或等于所述最小步长因子，则输出权值矩阵；
    按照输出的权值矩阵计算信道容量。
  7. 根据权利要求2所述的下行预编码方法，其特征在于，按照下式更新所述迭代梯度值：
    Figure PCTCN2021092744-appb-100006
    其中，μ为迭代梯度值；k为收敛系数。
  8. 根据权利要求1所述的下行预编码方法，其特征在于，所述初始权值为所述下行信道中单用户权值；或者，所述下行信道的估计权值。
  9. 根据权利要求1所述的下行预编码方法，其特征在于，获取下行信道的信道信息包括：
    获取配对信息，所述配对信息包括目标终端和配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种；
    生成信道特征矩阵，所述信道特征矩阵中的元素为在所述配对信息中提取的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种。
  10. 根据权利要求1所述的下行预编码方法，其特征在于，按照所述功率分配权值对下行数据执行预编码包括：
    根据所述功率分配权值和所述下行信道的信道容量生成分配功率值；
    应用所述分配功率值至所述下行数据。
  11. 根据权利要求9或10所述的下行预编码方法，其特征在于，所述方法还包括：
    发送下行信令，所述下行信令包括配对信息。
  12. 根据权利要求11所述的下行预编码方法，其特征在于，所述下行信令承载在无线资源控制信令和/或下行控制信息信令中。
  13. 一种下行预编码装置，其特征在于，包括：
    获取模块，用户获取下行信道的信道信息，所述信道信息包括下行信道的初始权值和下行信道的信道特征矩阵；
    权值计算模块，用于使用互信息最大化算法迭代生成下行信道的功率分配权值，所述互信息最大化算法的目标函数为下行信道的互信息下界函数，所述互信息最大化算法的初始化参数为所述初始权值和信道特征矩阵；
    预编码模块，用于按照所述功率分配权值对下行数据执行预编码，所述下行数据为所述下行信道中传递的数据。
  14. 根据权利要求13所述的下行预编码装置，其特征在于，
    所述权值计算模块，具体用于设置迭代梯度值为初始步长因子以及最小步长因子；
    根据所述初始权值和信道特征矩阵生成初始迭代速率；
    如果所述迭代梯度值大于所述最小步长因子，更新权值矩阵；
    根据更新后的权值矩阵和所述信道特征矩阵生成中间迭代速率；
    如果所述中间迭代速率大于或等于所述初始迭代速率，则更新所述迭代梯度值，以及在所述迭代梯度值小于或等于所述最小步长因子时提取权值矩阵，获得所述功率分配权值。
  15. 根据权利要求14所述的下行预编码装置，其特征在于，
    所述权值计算模块，具体用于如果所述中间迭代速率小于所述初始迭代速率，则更新所述初始迭代速率为所述中间迭代速率。
  16. 根据权利要求14所述的下行预编码装置，其特征在于，所述初始迭代速率按照 下式计算生成：
    R n＝G(H,W n)；
    其中，R n为初始迭代速率；G(H，W)为互信息下界函数，
    Figure PCTCN2021092744-appb-100007
    Figure PCTCN2021092744-appb-100008
    H为信道特征矩阵；W n为初始权值矩阵；e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引。
  17. 根据权利要求16所述的下行预编码装置，其特征在于，按照下式更新权值矩阵：
    Figure PCTCN2021092744-appb-100009
    其中，α为预设影响系数；H为信道特征矩阵；W n为权值矩阵；
    Figure PCTCN2021092744-appb-100010
    为互信息下界函数的一阶导数；
    Figure PCTCN2021092744-appb-100011
    e ij为星座点码距；i，j分别为星座点编号；M Nt为星座点总数量；σ 2为噪声方差；n为迭代次数索引；μ为迭代梯度值。
  18. 根据权利要求14所述的下行预编码装置，其特征在于，
    所述权值计算模块，具体用于如果所述迭代梯度值小于或等于所述最小步长因子，则输出权值矩阵；
    按照输出的权值矩阵计算信道容量。
  19. 根据权利要求16所述的下行预编码装置，其特征在于，按照下式更新所述迭代梯度值：
    Figure PCTCN2021092744-appb-100012
    其中，μ为迭代梯度值；k为收敛系数。
  20. 根据权利要求13所述的下行预编码装置，其特征在于，所述初始权值为所述下行信道中单用户权值；或者，所述下行信道的估计权值。
  21. 根据权利要求13所述的下行预编码装置，其特征在于，
    所述获取模块，还用于获取配对信息，所述配对信息包括目标终端和配对终端的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种；
    生成信道特征矩阵，所述信道特征矩阵中的元素为在所述配对信息中提取的调制与编码策略信息、调制方案信息以及星座点类型信息中的一种或多种。
  22. 根据权利要求13所述的下行预编码装置，其特征在于，
    所述预编码模块，具体用于根据所述功率分配权值和所述下行信道的信道容量生成分配功率值；
    应用所述分配功率值至所述下行数据。
  23. 根据权利要求21或22所述的下行预编码装置，其特征在于，还包括配对模块；
    所述配对模块，用于发送下行信令，所述下行信令包括配对信息。
  24. 根据权利要求23所述的下行预编码装置，其特征在于，所述下行信令承载在无线资源控制信令和/或下行控制信息信令中。
  25. 一种基站，其特征在于，包括信号收发台和控制器，所述信号收发台连接所述控制器；所述控制器被配置为执行操作指令，以实现权利要求1-13任一项所述的方法，用于控制所述信号收发台对下行数据执行预编码。
  26. 一种计算机存储介质，其特征在于，包括计算机指令，当所述计算机指令在网络设备上运行时，使得所述网络设备执行如权利要求1-13中任一项所述的方法。
  27. 一种计算机程序产品，其特征在于，当所述计算机程序产品在计算机上运行时，使得所述计算机执行如权利要求1-13中任一项所述的方法。
PCT/CN2021/092744 2020-05-27 2021-05-10 一种下行预编码方法、装置及基站 WO2021238634A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010460800.7A CN113746512B (zh) 2020-05-27 2020-05-27 一种下行预编码方法、装置及基站
CN202010460800.7 2020-05-27

Publications (1)

Publication Number Publication Date
WO2021238634A1 true WO2021238634A1 (zh) 2021-12-02

Family

ID=78722970

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/092744 WO2021238634A1 (zh) 2020-05-27 2021-05-10 一种下行预编码方法、装置及基站

Country Status (2)

Country Link
CN (1) CN113746512B (zh)
WO (1) WO2021238634A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114629556A (zh) * 2022-03-29 2022-06-14 西北工业大学 一种低复杂度的最优功率分配方法
CN114665928A (zh) * 2022-03-14 2022-06-24 北京邮电大学 一种基于mimo-vlc系统中的电功率分配与qr-osic预编码方法
CN116418397A (zh) * 2023-06-12 2023-07-11 南昌大学 一种用户公平性的速率分集辅助的可见光通信方法及系统
WO2023185434A1 (zh) * 2022-03-31 2023-10-05 华为技术有限公司 一种数据传输方法以及装置
US11963111B2 (en) 2022-08-24 2024-04-16 Nokia Solutions And Networks Oy Low complexity precoder

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103931110A (zh) * 2011-08-31 2014-07-16 三星电子株式会社 具有每个天线的功率约束的多天线发送
CN109922487A (zh) * 2019-03-28 2019-06-21 南京邮电大学 一种下行mimo-noma网络下的资源分配方法
CN110190881A (zh) * 2019-05-27 2019-08-30 南京邮电大学 权重速率最优的下行mimo-noma功率分配方法

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10187133B2 (en) * 2004-04-02 2019-01-22 Rearden, Llc System and method for power control and antenna grouping in a distributed-input-distributed-output (DIDO) network
CN102546125B (zh) * 2011-12-14 2015-03-04 清华大学 低复杂度的预编码调制矩阵生成方法及其预编码调制方法
CN103036651B (zh) * 2012-12-17 2016-02-03 广东省电信规划设计院有限公司 自适应mimo预编码传输的方法及系统
CN103763072B (zh) * 2014-01-23 2017-02-22 东南大学 一种mimo链路自适应传输方法
CN105162504B (zh) * 2015-09-21 2019-01-29 华南理工大学 一种快速mimo系统发射端预编码方法
CN108471325B (zh) * 2018-03-23 2020-09-15 北京理工大学 一种稀疏射频/基带混合预编码方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103931110A (zh) * 2011-08-31 2014-07-16 三星电子株式会社 具有每个天线的功率约束的多天线发送
CN109922487A (zh) * 2019-03-28 2019-06-21 南京邮电大学 一种下行mimo-noma网络下的资源分配方法
CN110190881A (zh) * 2019-05-27 2019-08-30 南京邮电大学 权重速率最优的下行mimo-noma功率分配方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
IOANNA IOANNOU; CHARALAMBOS D CHARALAMBOUS; STOJAN DENIC: "Mutual Information Expansion for MIMO Systems and Capacity Formulae at Low SNR", IEEE COMMUNICATIONS LETTERS, vol. 15, no. 3, 31 January 2011 (2011-01-31), pages 329 - 331, XP011350728, ISSN: 1089-7798, DOI: 10.1109/LCOMM.2011.012511.102277 *
SUN YANJING , LIU WEN , CAO QI , WANG XIAOLIN , ZHOU SHU , LI SONG: "Minimum Rate Optimization with Joint Power Allocation for Full-Duplex MIMO Relay", JOURNAL OF XIDIAN UNIVERSITY (NATURAL SCIENCE EDITION), vol. 45, no. 2, 30 April 2018 (2018-04-30), CN, pages 103 - 109, XP055872312, ISSN: 1001-2400, DOI: 10.3969/j.issn.1001-2400.2018.02.018 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114665928A (zh) * 2022-03-14 2022-06-24 北京邮电大学 一种基于mimo-vlc系统中的电功率分配与qr-osic预编码方法
CN114629556A (zh) * 2022-03-29 2022-06-14 西北工业大学 一种低复杂度的最优功率分配方法
CN114629556B (zh) * 2022-03-29 2024-01-30 西北工业大学 一种低复杂度的最优功率分配方法
WO2023185434A1 (zh) * 2022-03-31 2023-10-05 华为技术有限公司 一种数据传输方法以及装置
US11963111B2 (en) 2022-08-24 2024-04-16 Nokia Solutions And Networks Oy Low complexity precoder
CN116418397A (zh) * 2023-06-12 2023-07-11 南昌大学 一种用户公平性的速率分集辅助的可见光通信方法及系统
CN116418397B (zh) * 2023-06-12 2023-09-05 南昌大学 一种用户公平性的速率分集辅助的可见光通信方法及系统

Also Published As

Publication number Publication date
CN113746512A (zh) 2021-12-03
CN113746512B (zh) 2023-04-07

Similar Documents

Publication Publication Date Title
WO2021238634A1 (zh) 一种下行预编码方法、装置及基站
Liu et al. Optimized uplink transmission in multi-antenna C-RAN with spatial compression and forward
Liu et al. Multi-antenna wireless powered communication with energy beamforming
Liu et al. Joint power and antenna selection optimization in large cloud radio access networks
US20170180097A1 (en) System and Method for Using Semi-Orthogonal Multiple Access in Wireless Local Area Networks
WO2019196801A1 (zh) 数据传输的方法、通信装置及系统
US11968005B2 (en) Provision of precoder selection policy for a multi-antenna transmitter
WO2012084201A1 (en) Precoding matrix index selection process for a mimo receiver based on a near-ml detection, and apparatus for doing the same
US9712310B2 (en) System and method for channel quality indicator and acknowledgement joint detection
CN109068382B (zh) 一种基于时延QoS的NOMA跨层功率分配方法
US10141989B2 (en) System and method for quantization of angles for beamforming feedback
CN104092519A (zh) 一种基于加权和速率最大化的多用户mimo协作传输方法
Sangdeh et al. LB-SciFi: Online learning-based channel feedback for MU-MIMO in wireless LANs
Sindhwani et al. FFOAS: antenna selection for MIMO wireless communication system using firefly optimisation algorithm and scheduling
US10993232B2 (en) Apparatus and method for channel feedback in wireless communication system
WO2017005086A1 (zh) 一种预编码方法及装置
WO2012145923A1 (zh) 多用户多输入多输出系统中选择多用户配对的方法和装置
WO2020135534A1 (zh) 预编码方法和装置及信息传输方法和装置
WO2020210845A2 (en) Methods and apparatus for power allocation
CN108282201B (zh) 一种用户终端调度方法及装置、通信系统
US20230361820A1 (en) Downlink symbol-level precoding as part of multiuser miso communication
Lasaulce et al. Power allocation game for fading MIMO multiple access channels with antenna correlation
CN113258985A (zh) 单站多星mimo上注系统能效优化方法
Askri et al. Distributed Learning Assisted Fronthaul Compression for Multi-Antenna C-RAN
WO2022111119A1 (zh) 一种数据检测方法及通信装置

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: 21811757

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: 21811757

Country of ref document: EP

Kind code of ref document: A1