WO2022217609A1 - 信号检测网络的确定方法和装置 - Google Patents

信号检测网络的确定方法和装置 Download PDF

Info

Publication number
WO2022217609A1
WO2022217609A1 PCT/CN2021/087901 CN2021087901W WO2022217609A1 WO 2022217609 A1 WO2022217609 A1 WO 2022217609A1 CN 2021087901 W CN2021087901 W CN 2021087901W WO 2022217609 A1 WO2022217609 A1 WO 2022217609A1
Authority
WO
WIPO (PCT)
Prior art keywords
unit
sample
update
input
signal detection
Prior art date
Application number
PCT/CN2021/087901
Other languages
English (en)
French (fr)
Inventor
陈栋
Original Assignee
北京小米移动软件有限公司
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 北京小米移动软件有限公司 filed Critical 北京小米移动软件有限公司
Priority to CN202180000927.3A priority Critical patent/CN115769551A/zh
Priority to PCT/CN2021/087901 priority patent/WO2022217609A1/zh
Priority to US18/287,104 priority patent/US20240137074A1/en
Priority to EP21936484.1A priority patent/EP4325778A4/en
Publication of WO2022217609A1 publication Critical patent/WO2022217609A1/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
    • H04B7/04013—Intelligent reflective surfaces
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/024—Channel estimation channel estimation algorithms
    • H04L25/0254—Channel estimation channel estimation algorithms using neural network algorithms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0613—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
    • H04B7/0615—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
    • H04B7/0665—Feed forward of transmit weights to the receiver
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/06—Dc level restoring means; Bias distortion correction ; Decision circuits providing symbol by symbol detection

Definitions

  • the present disclosure relates to the field of communication technologies, and in particular, to a method for determining a signal detection network, a device for determining a signal detection network, a communication device, and a computer-readable storage medium.
  • the receiving end needs to perform signal detection on the transmitted signal of the transmitting end, so as to perform subsequent operations according to the estimated transmitted signal.
  • the method used for signal detection is mainly linear algorithm, such as minimum mean square error estimation MMSE (Minimum Mean Squareerror Estimate).
  • MMSE Minimum Mean Squareerror Estimate
  • the decoding matrix can be determined according to the channel matrix from the transmitter to the receiver, and then the decoding matrix and the receiver can be determined according to the channel matrix.
  • the received signal estimates the transmitted signal at the transmitter.
  • the embodiments of the present disclosure propose a method for determining a signal detection network, a device for determining a signal detection network, a communication device, and a computer-readable storage medium to solve the technical problems in the related art.
  • a method for determining a signal detection network including:
  • the initial neural network is trained to obtain a signal detection network, wherein, the input of the initial neural network is the sample communication parameter, and the output of the initial neural network is the estimated value of the transmission signal of the sample transmitter.
  • the sample communication parameters include at least one of the following:
  • the first channel matrix from the sample transmitter to the sample IRS the phase matrix of the sample IRS
  • a second channel matrix from the sample IRS to the sample receiver a third channel matrix from the sample transmitter to the sample receiver; and a sample receive signal from the sample receiver.
  • the initial neural network includes a plurality of cascaded update units, and the input of the update unit includes a unit common input and a unit related input; the method further includes: determining the unit correlation according to a gradient descent method The relationship between the input and the update value of the relevant input of the unit; the common input of the unit is determined according to the parameter used to characterize the update value in the relationship, and the output of the update unit is determined according to the update value; wherein , the output of the update unit is used as the unit-dependent input of the next update unit of the cascade.
  • the update unit includes two unit common inputs; the first unit common input of the two unit common inputs is based on the first channel matrix, the phase matrix, and the second channel matrix and the third channel matrix; the common input of the second unit in the common input of the two units is based on the first channel matrix, the phase matrix, the second channel matrix, the third channel matrix and the sample received signal is determined.
  • the update unit includes three fully connected layers, and the three fully connected layers include an input layer, a hidden layer, and an output layer.
  • the update unit further includes a short-circuit direct connection structure, wherein the starting point of the short-circuit direct connection structure is the unit-related input, and the end point is the hidden layer.
  • the network parameters of the fully connected layer include unit-related weights, and at least some of the unit-related weights in the update units are related to the ordering of the update units in the plurality of cascaded update units ; wherein, the higher the ranking of the update unit in the plurality of cascaded update units, the greater the unit correlation weight in the update unit.
  • the unit-related weight is a preset value
  • the unit-related weight is less than the preset value, and the higher the ranking of the update unit in the plurality of cascaded update units, the higher the ranking of the update unit in the update unit.
  • a signal detection method which includes: in response to receiving a received signal from an IRS, the received signal is a signal converted by the IRS from a transmit signal sent by a transmitting end, and according to The above-mentioned method for determining the signal detection network determines that the signal detection network determines the transmission signal.
  • an apparatus for determining a signal detection network including:
  • the parameter determination module is configured to determine the sample communication parameters for the sample transmitter and the sample receiver to communicate through the sample intelligent reflective surface IRS;
  • the network training module is configured to determine the initial neural network based on the training sample set formed by the sample communication parameters. Perform training to obtain a signal detection network, wherein the input of the initial neural network is the sample communication parameter, and the output of the initial neural network is an estimated value of the transmitted signal of the sample transmitter.
  • the sample communication parameters include at least one of the following: a first channel matrix from the sample transmitter to the sample IRS; a phase matrix of the sample IRS; the sample IRS to the sample receiver The second channel matrix from the sample transmitter end; the third channel matrix from the sample transmitter end to the sample receiver end; and the sample receive signal from the sample receiver end.
  • the initial neural network includes a plurality of cascaded update units, and the inputs of the update units include unit common inputs and unit related inputs.
  • the apparatus further includes: a relationship determination module configured to determine a relationship between the unit-related input and an updated value of the unit-related input according to a gradient descent method; an input-output determination module configured to determine a relationship according to the relationship
  • the parameter used to characterize the update value determines the common input of the unit, and the output of the update unit is determined according to the update value; wherein, the output of the update unit is used as the unit-related input of the next update unit of the cascade.
  • the update unit includes two unit common inputs; the first unit common input of the two unit common inputs is based on the first channel matrix, the phase matrix, and the second channel matrix and the third channel matrix; the common input of the second unit in the common input of the two units is based on the first channel matrix, the phase matrix, the second channel matrix, the third channel matrix and the sample received signal is determined.
  • the three-layer fully connected layer includes an input layer, a hidden layer and an output layer.
  • the update unit further includes a short-circuit direct connection structure, wherein the starting point of the short-circuit direct connection structure is the unit-related input, and the end point is the hidden layer.
  • the network parameters of the fully connected layer include unit-related weights, and at least some of the unit-related weights in the update units are related to the ordering of the update units in the plurality of cascaded update units ; wherein, the higher the ranking of the update unit in the plurality of cascaded update units, the greater the unit correlation weight in the update unit.
  • the unit-related weight in an update unit whose ranking is higher than a preset order, is a preset value; in an update unit whose ranking is lower than or equal to the preset order, the unit-related weight is less than the preset value, And the higher the ranking of the update unit in the plurality of cascaded update units, the greater the unit correlation weight in the update unit.
  • a signal detection apparatus including: a signal estimation module, configured to respond to receiving a received signal from an IRS, where the received signal is a transmit signal sent by a transmitting end through the For the signal converted by the IRS, the transmitted signal is determined according to the signal detection network determined by the device for determining the signal detection network.
  • a communication apparatus comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the above method for determining a signal detection network.
  • a communication device comprising: a processor; a memory for storing a computer program; wherein, when the computer program is executed by the processor, the above signal detection method is implemented.
  • a computer-readable storage medium for storing a computer program, and when the program is executed by a processor, the steps in the above method for determining a signal detection network are implemented.
  • a computer-readable storage medium for storing a computer program, and when the computer program is executed by a processor, the steps in the above signal detection method are implemented.
  • the initial neural network can be trained based on the training sample set to obtain the signal detection network. Since the neural network is not a linear algorithm, it is not limited to perform operations on linear relationships, even in the case of a large number of nonlinear factors. In complex communication environments, it can also be effectively used for signal detection to obtain high-precision estimation results for the transmitted signal.
  • the training sample set is constructed based on the sample communication parameters that the sample transmitter and the sample receiver communicate through the sample IRS
  • the signal detection network trained based on the training sample is more suitable for the communication between the receiver and the transmitter through the IRS. scene, signal detection can be performed accurately for this scene.
  • FIG. 1 is a schematic flowchart of a method for determining a signal detection network according to an embodiment of the present disclosure.
  • FIG. 2 is a schematic flowchart of applying the signal detection network according to an embodiment of the present disclosure.
  • FIG. 3 is a schematic flowchart of still another method for determining a signal detection network according to an embodiment of the present disclosure.
  • FIG. 4A is a schematic partial structure diagram of an initial neural network according to an embodiment of the present disclosure.
  • FIG. 4B is a schematic partial structure diagram of another initial neural network according to an embodiment of the present disclosure.
  • FIG. 5A is a schematic partial structural diagram of an update unit according to an embodiment of the present disclosure.
  • FIG. 5B is a schematic partial structure diagram of another update unit according to an embodiment of the present disclosure.
  • FIG. 5C is a schematic partial structural diagram of yet another update unit according to an embodiment of the present disclosure.
  • FIG. 6 is a schematic block diagram of an apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • FIG. 7 is a schematic block diagram of another apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • Fig. 8 is a schematic block diagram of an apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • FIG. 9 is a schematic block diagram of another apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • FIG. 1 is a schematic flowchart of a method for determining a signal detection network according to an embodiment of the present disclosure.
  • the determination method shown in this embodiment can be used to determine a signal detection network, and the signal detection network can be used by the receiver to perform signal detection. For example, after the transmitter sends a transmit signal, the receiver can receive a corresponding received signal, and the receiver can The signal detection network estimates the transmitted signal.
  • the transmitting end and the receiving end may be communication devices such as terminals, base stations, satellites, unmanned aerial vehicles, and core networks.
  • the following embodiments mainly illustrate the present disclosure in the case that the transmitting end is a base station and the receiving end is a terminal.
  • the terminal includes, but is not limited to, a communication device such as a mobile phone, a tablet computer, a wearable device, a sensor, and an Internet of Things device.
  • the terminal may communicate with a base station as a user equipment, and the base station includes but is not limited to a 4G base station, a 5G base station, and a 6G base station.
  • the method for determining the signal detection network may include the following steps:
  • step S101 determine the sample communication parameters that the sample transmitting end and the sample receiving end communicate through the sample intelligent reflecting surface IRS (Intelligent Reflecting Surface);
  • step S102 an initial neural network is trained based on the training sample set composed of the sample communication parameters to obtain a signal detection network, wherein the input of the initial neural network is the sample communication parameters, and the input of the initial neural network is the sample communication parameter.
  • the output is an estimate of the transmitted signal at the transmitter of the sample.
  • the transmitting end and the receiving end may communicate directly or communicate through the intelligent reflective surface IRS.
  • the method for determining the signal detection network described in the embodiments of the present disclosure may be applied in a scenario where the transmitting end and the receiving end communicate through the IRS.
  • the smart reflective surface is a planar array composed of a large number of reconfigurable passive components.
  • each passive component can independently generate a certain phase shift on the incident signal. , so that the reflected signal reflected by the smart reflective surface changes the propagation characteristics relative to the incident signal, such as changing the phase and direction.
  • an initial neural network and a training sample set may be constructed, and the initial neural network may be trained based on the training sample set to obtain a signal detection network.
  • the deep learning algorithm used in the training process can be selected as needed, for example, the Adam optimization algorithm and the end-to-end learning method can be selected.
  • sample transmitters may be predetermined, and then the sample transmitters may be used to communicate with the sample receivers through the sample IRSs, that is, the transmit signals of the sample transmitters may be transmitted to the samples via the sample IRSs.
  • the sample communication parameters through which the sample transmitter and the sample receiver communicate through the sample IRS can be determined, and a training sample set formed based on these sample communication parameters can be determined.
  • the sample communication parameters include at least one of the following:
  • the samples at the sample receiving end receive the signal y.
  • sample communication parameters are all known quantities for the sample receiving end.
  • the sample receiving signal y is the signal received by the sample receiving end from the sample transmitting end.
  • the first channel matrix H 1 and the second channel matrix H 2 can be based on the sample
  • the transmitter uses the sample IRS to determine the channel state information (Channel State Information, CSI) of the channel used by the sample receiver to transmit the signal, and the third channel matrix H3 can be directly transmitted to the sample receiver according to the sample receiver.
  • CSI Channel State Information
  • the CSI of the channel used by the signal It is determined that the phase matrix ⁇ of the sample IRS may be determined before or during the communication between the transmitter and the receiver, and may be specifically determined according to the state of the passive elements in the sample IRS.
  • the initial neural network can be trained based on the training sample set to obtain the signal detection network. Since the neural network is not a linear algorithm, it is not limited to perform operations on linear relationships, even in the case of a large number of nonlinear factors. In complex communication environments, it can also be effectively used for signal detection to obtain high-precision estimation results for the transmitted signal.
  • the training sample set is constructed based on the sample communication parameters that the sample transmitter and the sample receiver communicate through the sample IRS
  • the signal detection network trained based on the training sample is more suitable for the communication between the receiver and the transmitter through the IRS. scene, signal detection can be performed accurately for this scene.
  • the data in the training sample set can also be divided into three sets, the first set is used as a sample set for training, and the second set is used as a test set for performing the training results.
  • the third set is used as a validation set to verify the training results.
  • the ratio of samples in the first set, the second set, and the third set can be 96:2:2.
  • FIG. 2 is a schematic flowchart of applying the signal detection network according to an embodiment of the present disclosure.
  • the receiving signal in response to the receiving end receiving a receiving signal from the IRS, the receiving signal is a signal converted by the IRS from the transmitting signal sent by the transmitting end, and is determined according to the signal detection network. the transmitted signal.
  • signal detection may be performed through the signal detection network to determine the transmit signal sent by the transmitter.
  • the IRS can inform the receiving end of the overall phase matrix of the IRS, so that the receiving end can determine the phase matrix of the IRS.
  • the receiving end can also determine the channel matrix from the transmitting end to the IRS and the channel matrix from the IRS to the receiving end according to the channel state information of the channel used by the transmitting end to transmit signals to the receiving end through the IRS, and directly send the information to the receiving end according to the transmitting end.
  • the channel state information of the channel used to transmit the signal determines the channel matrix from the sample transmitter to the sample receiver.
  • the receiving end can determine the input quantity based on the received signal, the above-mentioned phase matrix and the three channel matrices to input into the signal detection network determined in the above-mentioned embodiment, so as to obtain the estimated value of the transmitted signal of the transmitting end.
  • FIG. 3 is a schematic flowchart of still another method for determining a signal detection network according to an embodiment of the present disclosure.
  • the initial neural network includes a plurality of cascaded update units, and the input of the update unit includes unit common input and unit related input;
  • the method also includes:
  • step S301 the relationship between the unit-related input and the updated value of the unit-related input is determined according to the gradient descent method
  • step S302 the common input of the unit is determined according to the parameter used to characterize the update value in the relationship, and the output of the update unit is determined according to the update value;
  • the output of the update unit is used as the unit-related input of the next update unit of the cascade.
  • a plurality of cascaded update units may be set to form an initial neural network, and the inputs of the update units include unit common inputs and unit related inputs. Among them, the unit common input as the input of each update unit remains unchanged; the unit related input varies based on the update unit.
  • the output of the update unit may be a gradient descent-based update value of the unit-related input of the update unit, for example, for the ith update unit in the n+1 update units, the unit-related input is x i , the output is x i+1 , n ⁇ 0, 0 ⁇ i ⁇ n.
  • the relationship between the received signal x received by the receiver and the transmitted signal y sent by the transmitter is:
  • n Gaussian white noise, which is a random variable
  • H the channel matrix
  • the channel matrix H in the above formula can be based on the first channel matrix from the sample transmitter to the sample IRS.
  • H H 2 ⁇ H 1 +H 3 ;
  • the embodiments of the present disclosure are mainly for obtaining make x and The gap between them is as small as possible.
  • the update in this embodiment, the input is updated by the update unit Update, for example, for the i-th update unit, you can update to get relative to It is closer to the transmitted signal x actually sent by the transmitting end.
  • ⁇ i is the ith update unit pair the step size for updating, is a preset value as required.
  • FIG. 4A is a schematic partial structure diagram of an initial neural network according to an embodiment of the present disclosure.
  • FIG. 4B is a schematic partial structure diagram of another initial neural network according to an embodiment of the present disclosure.
  • the input is The output is can be directly As the input of the next update unit, that is, the i+1th update unit U i+1 And so on until you get
  • the input of the i-th update unit U i can also be and output Weighted summation (the weights used by each update unit for weighting can be preset as needed), and the weighted summation result is used as the input of the i+1th update unit U i+1
  • Weighted summation the weights used by each update unit for weighting can be preset as needed
  • the weighted summation result is used as the input of the i+1th update unit U i+1
  • the structure shown in FIG. 4B is more complicated than the structure shown in FIG. 4A , more factors of the previous unit can be introduced into the input of the next update unit, which is beneficial to avoid the excessively large update amplitude each time, which seriously deviates from the expected value .
  • the update unit includes two unit common inputs;
  • the common input of the first unit in the common input of the two units is determined based on the first channel matrix, the phase matrix, the second channel matrix and the third channel matrix, for example, (H in the above formula 2) 2 ⁇ H 1 +H 3 ) T (H 2 ⁇ H 1 +H 3 );
  • a second unit common input of the two unit common inputs is determined based on the first channel matrix, the phase matrix, the second channel matrix, the third channel matrix and the sample received signal, eg (H 2 ⁇ H 1 +H 3 ) Ty in the above formula 2.
  • each update unit in the initial neural network only needs 3 inputs, and the number of inputs is small, so that the number of connections in the network is small, which is beneficial to reduce the complexity of the network and make the network suitable for complex channel environments , the results can also be output faster.
  • FIG. 5A is a schematic partial structural diagram of an update unit according to an embodiment of the present disclosure.
  • the update unit includes three fully connected layers including an input layer, a hidden layer and an output layer.
  • the update unit may be designed to include three fully connected layers, and the three fully connected layers may specifically include an input layer, a hidden layer, and an output layer.
  • the input layer and hidden layer in the i-th update unit U i are mainly shown, wherein the corresponding weight of the input layer is w i1 , the corresponding bias is b i1 , and the corresponding activation function is ⁇ ;
  • the corresponding weight of the hidden layer is w i2 , the corresponding bias is b i2 , and the corresponding activation function is ⁇ .
  • Concatenated Concat refers to the concatenation method, which can concatenate strings, arrays, vectors, etc. in the input method.
  • the result of multiplication and (H 2 ⁇ H 1 +H 3 ) Ty are input to concat in series, then the series result of the three can be obtained.
  • the number of antennas at the transmitting end is M (known at the receiving end), and the transmitted signal can be a vector of M*1.
  • M is an M*1 vector, (H 2 ⁇ H 1 +H 3 ) T (H 2 ⁇ H 1 +H 3 ) and
  • the result of multiplication is also a vector of M*1, (H 2 ⁇ H 1 +H 3 ) T y is also a vector of M*1, then the result of concatenating the three is a vector of 3M*1, so you can set the input
  • the input of the layer is 3M*1.
  • the input is multiplied by w i1 in the input layer and added to bi1 , and then passed through the activation function ⁇ , and then output to the hidden layer, where the activation function ⁇ can be a sigmod function.
  • the number of inputs to the hidden layer it can be set as required, and is generally greater than or equal to the number of inputs to the input layer, for example, it can be 4M*1.
  • the input is multiplied by w i2 in the hidden layer and added to b i2 , and then passed through the activation function ⁇ , and then output to the output, where the activation function ⁇ can be a tanh function.
  • the output number of the output layer is M*1, which can form an M*1 vector input to the next update unit.
  • the activation function is not limited to the examples exemplified in the above-mentioned embodiments, and can be specifically selected as required, for example, the relu function can also be selected as the activation function.
  • FIG. 5B is a schematic partial structure diagram of another update unit according to an embodiment of the present disclosure.
  • the update unit further includes a short-circuit direct connection structure, wherein the starting point of the short-circuit direct connection structure is the cell-related input, and the end point is the hidden layer.
  • the full A short-cut (also can be translated as direct, shortcut) structure is added to the connection layer.
  • the starting point of the short-cut structure is the relevant input of the unit, and the end point is the hidden layer. It is added to the output of the input layer (either directly or by weighted summation), and the result of the addition is used as the input of the hidden layer.
  • a signal detection network can be obtained by training an initial neural network formed by cascading update units in the above embodiment based on a training sample set.
  • the network parameters of the fully connected layer include cell-related weights, and the cell-related weights in at least some of the update units are related to the ordering of the update units in the plurality of cascaded update units ; Wherein, the higher the order of the update unit in the multiple cascaded update units, the greater the unit correlation weight in the update unit, the update unit in the multiple cascaded update units The later the order in the update unit is, the smaller the unit-related weight in the update unit is.
  • unit-dependent weights ie weights associated with update units, may be introduced in the input layer of the update unit.
  • FIG. 5C is a schematic partial structural diagram of yet another update unit according to an embodiment of the present disclosure.
  • the unit-related weight of the i-th update unit U i is ⁇ i .
  • U i is ranked higher in the n cascaded update units, that is, the smaller i is. , the larger ⁇ i is.
  • the position of the unit-related weight ⁇ i in the fully connected layer can be set as required, and generally can be set in the input layer and/or hidden layer, for example, in the input layer, which can be set as shown in Figure 5C
  • the input is further multiplied by the result of multiplying w i1 in the input layer, and if it is set in the hidden layer, it can be set after w i2 , and the input is multiplied by w i2 in the hidden layer.
  • the multiplied result is further multiplied.
  • the earlier update has a greater impact on the overall update process than the later update, so when multiple update units are used to update the
  • a relatively large unit-related weight can be set for the update unit ranked earlier, and a relatively small unit-related weight can be set for the update unit ranked later, which is beneficial to reduce the complexity of the training process.
  • the content related to the unit-related weights is not limited to that shown in FIG. 5C , but is applied on the basis of the embodiment shown in FIG. 5B , and can also be directly applied on the basis of the embodiment shown in FIG. 5A .
  • the unit-related weight is a preset value.
  • the unit-related weight is less than the preset value, and the higher the ranking of the update unit in the plurality of cascaded update units, the higher the ranking of the update unit in the update unit. The larger the unit-related weight is, the later the update unit is ranked in the plurality of cascaded update units, and the smaller the unit-related weight in the update unit is.
  • the unit correlation it is possible to set the unit correlation to remain unchanged for some update units, and only for another part of the update units to change with the order.
  • the unit-related weight can be set to decrease as the ranking is later.
  • the change of the relative weights of some units can be reduced, which is beneficial to reduce the complexity of the training process.
  • the unit correlation weight ⁇ i can be a semi-exponential function of updating the unit order i, and its form can be as follows:
  • the unit-related weights in the first half of the update units can remain unchanged, while the unit-related weights in the second half of the update units can decrease as the update units are ranked later.
  • the embodiment of the present disclosure also proposes a signal detection method, and the signal detection method can be implemented by the receiving end in the communication process between the transmitting end and the receiving end.
  • the transmitting end and the receiving end may be communication devices such as terminals, base stations, satellites, unmanned aerial vehicles, and core networks.
  • the receiving end can be either a base station or a terminal.
  • the transmitting end is a base station
  • the receiving end is a terminal
  • the transmitting end is a terminal
  • the receiving end is a base station.
  • the terminal includes, but is not limited to, a communication device such as a mobile phone, a tablet computer, a wearable device, a sensor, and an Internet of Things device.
  • the terminal may communicate with a base station as a user equipment, and the base station includes but is not limited to a 4G base station, a 5G base station, and a 6G base station.
  • the signal detection method may include the following steps:
  • the received signal is a signal converted by the IRS from the transmit signal sent by the transmitter, and the transmit signal is determined by the signal detection network determined by the method in any of the foregoing embodiments.
  • the signal detection network may be used for signal detection to determine the signal sent by the transmitter. transmit a signal.
  • the IRS can inform the receiving end of the overall phase matrix of the IRS, so that the receiving end can determine the phase matrix of the IRS.
  • the receiving end can also determine the channel matrix from the transmitting end to the IRS and the channel matrix from the IRS to the receiving end according to the channel state information of the channel used by the transmitting end to transmit signals to the receiving end through the IRS, and directly send the information to the receiving end according to the transmitting end.
  • the channel state information of the channel used to transmit the signal determines the channel matrix from the sample transmitter to the sample receiver. Then the receiving end can determine the input quantity (for example, the input quantity corresponding to the three input quantities of the embodiment shown in FIG. 4A and FIG.
  • the estimated value of the signal transmitted by the transmitter is obtained.
  • the operation of determining the signal detection network can be performed by the receiving end or by other devices. This is not limited by the embodiments of the present disclosure. For example, the operation of determining the signal detection network can be performed by other devices. Send the signal detection network to the receiver.
  • the present disclosure also provides an embodiment of an apparatus for determining a signal detection network.
  • FIG. 6 is a schematic block diagram of an apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • the determining device shown in this embodiment can be used to determine a signal detection network, and the signal detection network can be used by the receiving end to perform signal detection.
  • the signal detection network estimates the transmitted signal.
  • the transmitting end and the receiving end may be communication devices such as terminals, base stations, satellites, unmanned aerial vehicles, and core networks.
  • the following embodiments mainly illustrate the present disclosure in the case that the transmitting end is a base station and the receiving end is a terminal.
  • the terminal includes, but is not limited to, a communication device such as a mobile phone, a tablet computer, a wearable device, a sensor, and an Internet of Things device.
  • the terminal may communicate with a base station as a user equipment, and the base station includes but is not limited to a 4G base station, a 5G base station, and a 6G base station.
  • the device for determining the signal detection network may include: a parameter determination module 601 configured to determine sample communication parameters for communication between the sample transmitter and the sample receiver through the sample intelligent reflective surface IRS; the network training module 602 , is configured to train an initial neural network based on the training sample set composed of the sample communication parameters to obtain a signal detection network, wherein the input of the initial neural network is the sample communication parameter, and the output of the initial neural network is the estimated value of the transmitted signal at the sample transmitter.
  • the sample communication parameters include at least one of the following: a first channel matrix from the sample transmitter to the sample IRS; a phase matrix of the sample IRS; the sample IRS to the sample receiver The second channel matrix from the sample transmitter end; the third channel matrix from the sample transmitter end to the sample receiver end; and the sample receive signal from the sample receiver end.
  • FIG. 7 is a schematic block diagram of yet another apparatus for determining a signal detection network according to an embodiment of the present disclosure.
  • the initial neural network includes a plurality of cascaded update units, and the inputs of the update units include unit common inputs and unit related inputs.
  • the apparatus further includes: a relationship determination module 701, configured to determine the relationship between the unit-related input and the updated value of the unit-related input according to a gradient descent method; an input-output determination module 702, configured to The parameter used to characterize the update value in the relationship determines the common input of the unit, and the output of the update unit is determined according to the update value; wherein, the output of the update unit is related to the unit of the next update unit in the cascade enter.
  • a relationship determination module 701 configured to determine the relationship between the unit-related input and the updated value of the unit-related input according to a gradient descent method
  • an input-output determination module 702 configured to The parameter used to characterize the update value in the relationship determines the common input of the unit, and the output of the update unit is determined according to the update value; wherein, the output of the update unit is related to the unit of the next update unit in the cascade enter.
  • the update unit includes two unit common inputs; a first unit common input of the two unit common inputs is based on the first channel matrix, the phase matrix, the second channel matrix and the third channel matrix; the common input of the second unit in the common input of the two units is based on the first channel matrix, the phase matrix, the second channel matrix, the third channel matrix and the sample received signal is determined.
  • the update unit includes three fully connected layers including an input layer, a hidden layer and an output layer.
  • the update unit further includes a short-circuit direct connection structure, wherein the start point of the short-circuit direct connection structure is the cell-related input, and the end point is the hidden layer.
  • the network parameters of the fully connected layer include cell-related weights, and the cell-related weights in at least some of the update units are related to the ordering of the update units in the plurality of cascaded update units ;
  • the higher the ranking of the update unit in the multiple cascaded update units the greater the unit correlation weight in the update unit, the update unit in the multiple cascaded update units
  • the later the order of the smaller the unit-related weight in the update unit.
  • the unit-related weight is a preset value
  • the unit-related weight is less than the preset value, and the higher the ranking of the update unit in the plurality of cascaded update units, the higher the ranking of the update unit in the update unit.
  • Embodiments of the present disclosure further provide a signal detection device, which can be applied to a receiving end in a communication process between a transmitting end and a receiving end.
  • the transmitting end and the receiving end may be communication devices such as terminals, base stations, satellites, unmanned aerial vehicles, and core networks.
  • the receiving end can be either a base station or a terminal.
  • the transmitting end is a base station
  • the receiving end is a terminal
  • the transmitting end is a terminal
  • the receiving end is a base station.
  • the terminal includes, but is not limited to, a communication device such as a mobile phone, a tablet computer, a wearable device, a sensor, and an Internet of Things device.
  • the terminal may communicate with a base station as a user equipment, and the base station includes but is not limited to a 4G base station, a 5G base station, and a 6G base station.
  • the signal detection device may include:
  • a signal estimation module configured to respond to receiving a received signal from the IRS, the received signal is a signal converted by the IRS from the transmit signal sent by the transmitting end, and detect the signal determined by the device according to any one of the above embodiments
  • the network determines the transmitted signal.
  • the apparatus embodiments since they basically correspond to the method embodiments, reference may be made to the partial descriptions of the method embodiments for related parts.
  • the device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in One place, or it can be distributed over multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment. Those of ordinary skill in the art can understand and implement it without creative effort.
  • An embodiment of the present disclosure also provides a communication device, comprising: a processor; a memory for storing a computer program; wherein, when the computer program is executed by the processor, the signal detection network described in any of the foregoing embodiments is implemented method of determination.
  • Embodiments of the present disclosure also provide a communication device, including: a processor; a memory for storing a computer program; wherein, when the computer program is executed by the processor, the signal detection method described in the above embodiments is implemented.
  • Embodiments of the present disclosure further provide a computer-readable storage medium for storing a computer program, when the computer program is executed by a processor, the method for determining a signal detection network according to any of the foregoing embodiments is implemented. step.
  • Embodiments of the present disclosure also provide a computer-readable storage medium for storing a computer program, when the computer program is executed by a processor, the steps in the signal detection method described in the foregoing embodiments are implemented.
  • FIG. 8 is a schematic block diagram of another apparatus 800 for determining a signal detection network according to an embodiment of the present disclosure.
  • the apparatus 800 may be provided as a base station.
  • apparatus 800 includes a processing component 822, a wireless transmit/receive component 824, an antenna component 826, and a signal processing portion specific to a wireless interface, and the processing component 822 may further include one or more processors.
  • One of the processors in the processing component 822 may be configured to implement the method for determining a signal detection network described in any of the foregoing embodiments.
  • one of the processors in the processing component 822 may be configured to implement the steps of performing signal detection based on the signal detection network described in the foregoing embodiments.
  • FIG. 9 is a schematic block diagram of an apparatus 900 for determining a signal detection network according to an embodiment of the present disclosure.
  • apparatus 900 may be a mobile phone, computer, digital broadcast terminal, messaging device, game console, tablet device, medical device, fitness device, personal digital assistant, and the like.
  • the apparatus 900 may include one or more of the following components: a processing component 902, a memory 904, a power supply component 906, a multimedia component 908, an audio component 910, an input/output (I/O) interface 912, a sensor component 914, And the communication component 916 .
  • the processing component 902 generally controls the overall operation of the apparatus 900, such as operations associated with display, phone calls, data communications, camera operations, and recording operations.
  • the processing component 902 may include one or more processors 920 to execute instructions to complete all or part of the steps of the above-described method for determining a signal detection network.
  • the processor 920 may further execute an instruction to complete the step of performing signal detection based on the signal detection network described in the embodiment
  • processing component 902 may include one or more modules to facilitate interaction between processing component 902 and other components.
  • processing component 902 may include a multimedia module to facilitate interaction between multimedia component 908 and processing component 902.
  • Memory 904 is configured to store various types of data to support operations at device 900 . Examples of such data include instructions for any application or method operating on device 900, contact data, phonebook data, messages, pictures, videos, and the like. Memory 904 may be implemented by any type of volatile or non-volatile storage device or combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read only memory
  • EPROM erasable Programmable Read Only Memory
  • PROM Programmable Read Only Memory
  • ROM Read Only Memory
  • Magnetic Memory Flash Memory
  • Magnetic Disk Magnetic Disk or Optical Disk.
  • Power supply assembly 906 provides power to various components of device 900 .
  • Power supply components 906 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to device 900 .
  • Multimedia component 908 includes a screen that provides an output interface between the device 900 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user.
  • the touch panel includes one or more touch sensors to sense touch, swipe, and gestures on the touch panel. The touch sensor may not only sense the boundaries of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action.
  • the multimedia component 908 includes a front-facing camera and/or a rear-facing camera. When the apparatus 900 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each of the front and rear cameras can be a fixed optical lens system or have focal length and optical zoom capability.
  • Audio component 910 is configured to output and/or input audio signals.
  • audio component 910 includes a microphone (MIC) that is configured to receive external audio signals when device 900 is in operating modes, such as call mode, recording mode, and voice recognition mode. The received audio signal may be further stored in memory 904 or transmitted via communication component 916 .
  • audio component 910 also includes a speaker for outputting audio signals.
  • the I/O interface 912 provides an interface between the processing component 902 and a peripheral interface module, which may be a keyboard, a click wheel, a button, or the like. These buttons may include, but are not limited to: home button, volume buttons, start button, and lock button.
  • Sensor assembly 914 includes one or more sensors for providing status assessment of various aspects of device 900.
  • the sensor assembly 914 can detect the open/closed state of the device 900, the relative positioning of components, such as the display and keypad of the device 900, and the sensor assembly 914 can also detect a change in the position of the device 900 or a component of the device 900 , the presence or absence of user contact with the device 900 , the orientation or acceleration/deceleration of the device 900 and the temperature change of the device 900 .
  • Sensor assembly 914 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact.
  • Sensor assembly 914 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor assembly 914 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • Communication component 916 is configured to facilitate wired or wireless communication between apparatus 900 and other devices.
  • Device 900 may access wireless networks based on communication standards, such as WiFi, 2G or 3G, 4G LTE, 5G NR, or a combination thereof.
  • the communication component 916 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication component 916 also includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module may be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • apparatus 900 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable A gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components are implemented for implementing the above-mentioned determination method of the signal detection network.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA field programmable A gate array
  • controller a controller
  • microcontroller a microcontroller
  • microprocessor or other electronic components are implemented for implementing the above-mentioned determination method of the signal detection network.
  • a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 904 including instructions, and the instructions can be executed by the processor 920 of the apparatus 900 to complete the above method for determining a signal detection network .
  • the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Health & Medical Sciences (AREA)
  • Signal Processing (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Power Engineering (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Image Analysis (AREA)
  • Filters That Use Time-Delay Elements (AREA)

Abstract

本公开涉及信号检测网络的确定方法和装置，方法包括：确定样本发射端和样本接收端通过样本智能反射面IRS进行通信的样本通信参数（S101）；基于样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，初始神经网络的输入为样本通信参数，初始神经网络的输出为对样本发射端的发射信号的估计值（S102）。根据本公开，可以基于训练样本集对初始神经网络进行训练以得到信号检测网络，由于神经网络并非线性算法，并不局限于对线性关系进行运算，即使在非线性因素较多的复杂通信环境中，也能有效地用于信号检测，对发射信号得到高精度的估计结果。

Description

信号检测网络的确定方法和装置 技术领域
本公开涉及通信技术领域，具体而言，涉及信号检测网络的确定方法、信号检测网络的确定装置、通信装置和计算机可读存储介质。
背景技术
在基站与终端通信时，接收端需要对发射端的发射信号进行信号检测，以便根据估计确定的发射信号进行后续操作。
目前用于信号检测的方式主要为线性算法，例如最小均方误差估计MMSE(Minimum Mean Squareerror Estimate)，基于这种方式可以根据发射端到接收端的信道矩阵确定解码矩阵，然后根据解码矩阵以及接收端的接收信号估计出发射端的发射信号。
但是，随着通信技术的发展，在通信过程中，信道环境日益复杂化，信道中非线性因素也日益增多，通过目前的线性算法进行信号检测的精度会受到很大影响，严重影响信号检测的准确度。
发明内容
有鉴于此，本公开的实施例提出了信号检测网络的确定方法、信号检测网络的确定装置、通信装置和计算机可读存储介质，以解决相关技术中的技术问题。
根据本公开实施例的第一方面，提出一种信号检测网络的确定方法，包括：
确定样本发射端和样本接收端通过样本智能反射面(Intelligent Reflecting Surface，IRS)进行通信的样本通信参数；基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
在一个实施例中，所述样本通信参数包括以下至少之一：
所述样本发射端到所述样本IRS的第一信道矩阵；所述样本IRS的相位矩阵；
所述样本IRS到所述样本接收端的第二信道矩阵；所述样本发射端到所述样本 接收端的第三信道矩阵；以及所述样本接收端的样本接收信号。
在一个实施例中，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入；所述方法还包括：根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
在一个实施例中，所述更新单元包括两个单元公共输入；所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定；所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定。
在一个实施例中，所述更新单元包括三层全连接层，所述三层全连接层包括输入层、隐藏层和输出层。
在一个实施例中，所述更新单元还包括短路直连结构，其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
在一个实施例中，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单元中的单元相关权重越大。
在一个实施例中，在排序高于预设次序的更新单元中，单元相关权重为预设值；
在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大。
根据本公开实施例的第二方面，提出一种信号检测方法，包括：响应于接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据上述信号检测网络的确定方法确定信号检测网络确定所述发射信号。
根据本公开实施例的第三方面，提出一种信号检测网络的确定装置，包括：
参数确定模块，被配置为确定样本发射端和样本接收端通过样本智能反射面 IRS进行通信的样本通信参数；网络训练模块，被配置为基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
在一个实施例中，所述样本通信参数包括以下至少之一：所述样本发射端到所述样本IRS的第一信道矩阵；所述样本IRS的相位矩阵；所述样本IRS到所述样本接收端的第二信道矩阵；所述样本发射端到所述样本接收端的第三信道矩阵；以及所述样本接收端的样本接收信号。
在一个实施例中，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入。所述装置还包括：关系确定模块，被配置为根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；输入输出确定模块，被配置为根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
在一个实施例中，所述更新单元包括两个单元公共输入；所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定；所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定。
在一个实施例中，所述三层全连接层包括输入层、隐藏层和输出层。
在一个实施例中，所述更新单元还包括短路直连结构，其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
在一个实施例中，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单元中的单元相关权重越大。
在一个实施例中，在排序高于预设次序的更新单元中，单元相关权重为预设值；在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大。
根据本公开实施例的第四方面，提出一种信号检测装置，包括：信号估计模块，被配置为响应于接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据上述信号检测网络的确定装置确定的信号检测网络确定所述发射信号。
根据本公开实施例的第五方面，提出一种通信装置，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为执行上述信号检测网络的确定方法。
根据本公开实施例的第六方面，提出一种通信装置，包括：处理器；用于存储计算机程序的存储器；其中，当所述计算机程序被处理器执行时，实现上述信号检测方法。
根据本公开实施例的第七方面，提出一种计算机可读存储介质，用于存储计算机程序，所述程序被处理器执行时实现上述信号检测网络的确定方法中的步骤。
根据本公开实施例的第八方面，提出一种计算机可读存储介质，用于存储计算机程序，当所述计算机程序被处理器执行时，实现上述信号检测方法中的步骤。
根据本公开的实施例，可以基于训练样本集对初始神经网络进行训练以得到信号检测网络，由于神经网络并非线性算法，所以并不局限于对线性关系进行运算，即使在非线性因素较多的复杂通信环境中，也能有效地用于信号检测，对发射信号得到高精度的估计结果。
另外，由于训练样本集是基于样本发射端和样本接收端通过样本IRS进行通信的样本通信参数构建的，那么基于训练样本训练得到的信号检测网络，更契合于接收端和发射端通过IRS通信的场景，可以针对该场景准确地进行信号检测。
附图说明
为了更清楚地说明本公开实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是根据本公开的实施例示出的一种信号检测网络的确定方法的示意流程图。
图2是根据本公开的实施例示出的一种应用所述信号检测网络的示意流程图。
图3是根据本公开的实施例示出的又一种信号检测网络的确定方法的示意流程图。
图4A是根据本公开的实施例示出的一种初始神经网络的局部结构示意图。
图4B是根据本公开的实施例示出的另一种初始神经网络的局部结构示意图。
图5A是根据本公开的实施例示出的一种更新单元的局部结构示意图。
图5B是根据本公开的实施例示出的另一种更新单元的局部结构示意图。
图5C是根据本公开的实施例示出的又一种更新单元的局部结构示意图。
图6是根据本公开的实施例示出的一种信号检测网络的确定装置的示意框图。
图7是根据本公开的实施例示出的另一种信号检测网络的确定装置的示意框图。
图8是根据本公开的实施例示出的一种用于确定信号检测网络的装置的示意框图。
图9是根据本公开的实施例示出的另一种用于确定信号检测网络的装置的示意框图。
具体实施方式
下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本公开保护的范围。
图1是根据本公开的实施例示出的一种信号检测网络的确定方法的示意流程图。本实施例所示的确定方法可以用于确定信号检测网络，信号检测网络可以供接收端进行信号检测，例如在发射端发出发射信号后，接收端可以接收到响应的接收信号，接收端可以基于信号检测网络估计发射信号。
在一个实施例中，所述发射端和接收端可以是终端、基站、卫星、无人飞行器、核心网等通信设备。以下实施例主要在发射端为基站，接收端为终端的情况下对本公开进行示例性说明。
在一个实施例中，所述终端包括但不限于手机、平板电脑、可穿戴设备、传感器、物联网设备等通信装置。所述终端可以作为用户设备与基站通信，所述基站包括但不限于4G基站、5G基站、6G基站。
如图1所示，所述信号检测网络的确定方法可以包括以下步骤：
在步骤S101中，确定样本发射端和样本接收端通过样本智能反射面IRS(Intelligent Reflecting Surface)进行通信的样本通信参数；
在步骤S102中，基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
在一个实施例中，发射端和接收端可以直接通信，也可以通过智能反射面IRS通信。本公开的实施例所述的信号检测网络的确定方法可以应用在发射端和接收端通过IRS通信的场景中。
其中，智能反射面是由大量可重构的无源元件组成的平面阵列，针对入射到智能反射面上的入射信号，每个无源元件都能够独立地在入射信号上产生一定大小的相移，从而使得智能反射面反射出的反射信号，相对于入射信号改变传播特性，例如改变相位、方向等。
在一个实施例中，为了训练得到信号检测网络，可以构建初始神经网络和训练样本集，并基于训练样本集对初始神经网络进行训练以得到信号检测网络。
其中，训练过程中所使用到的深度学习算法可以根据需要选择，例如可以选择Adam优化算法和端到端学习方式。
在一个实施例中，可以预先确定大量样本发射端、样本接收端、以及样本IRS，然后使用样本发射端通过样本IRS与样本接收端通信，也即样本发射端的发射信号可以经由样本IRS传输至样本接收端。在这种情况下，可以确定样本发射端和样本接收端通过样本IRS进行通信的样本通信参数，并基于这些样本通信参数构成的训练样本集。
在一个实施例中，所述样本通信参数包括以下至少之一：
所述样本发射端到所述样本IRS的第一信道矩阵H 1；
所述样本IRS的相位矩阵Φ；
所述样本IRS到所述样本接收端的第二信道矩阵H 2；
所述样本发射端到所述样本接收端的第三信道矩阵H 3；
所述样本接收端的样本接收信号y。
这些样本通信参数对于样本接收端而言都是已知量，例如样本接收信号y就是样本接收端接收到的来自样本发射端的信号，第一信道矩阵H 1和第二信道矩阵H 2可以根据样本发射端通过样本IRS向样本接收端发射信号所使用信道的信道状态信息(Channel State Information，CSI)确定，第三信道矩阵H 3可以根据样本发射端直接向样本接收端发射信号所使用信道的CSI确定，样本IRS的相位矩阵Φ可以是发射端和接收端通信之前或通信过程中确定的，具体可以根据样本IRS中的无源元件的状态确定。
根据本公开的实施例，可以基于训练样本集对初始神经网络进行训练以得到信号检测网络，由于神经网络并非线性算法，所以并不局限于对线性关系进行运算，即使在非线性因素较多的复杂通信环境中，也能有效地用于信号检测，对发射信号得到高精度的估计结果。
另外，由于训练样本集是基于样本发射端和样本接收端通过样本IRS进行通信的样本通信参数构建的，那么基于训练样本训练得到的信号检测网络，更契合于接收端和发射端通过IRS通信的场景，可以针对该场景准确地进行信号检测。
在一个实施例中，在训练之前，也可以先将训练样本集中的数据拆分为三个集合，第一个集合作为样本集用于训练，第二个集合作为测试集用于对训练结果进行测试，第三个集合作为验证集用于对训练结果进行验证，例如第一个集合、第二个集合、第三个集合中样本的比例可以为96:2:2。
图2是根据本公开的实施例示出的一种应用所述信号检测网络的示意流程图。如图2所示，在步骤S201中，响应于接收端接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据所述信号检测网络确定所述发射信号。
在一个实施例中，在确定了信号检测网络后，可以在后续实际使用接收端和发射端通过IRS进行通信的场景中，通过信号检测网络进行信号检测，以确定发射端发出的发射信号。
例如IRS每次调整其中无源元件后，都可以将IRS整体的相位矩阵告知接收端， 从而接收端可以确定IRS的相位矩阵。接收端还可以根据发射端通过IRS向接收端发射信号所使用的信道的信道状态信息，确定所述发射端到IRS的信道矩阵，以及IRS到接收端的信道矩阵，并根据发射端直接向接收端发射信号所使用的信道的信道状态信息确定样本发射端到样本接收端的信道矩阵。然后接收端可以基于接收到的接收信号、上述相位矩阵和三个信道矩阵确定输入量输入到上述实施例确定的信号检测网络中，以得到对发射端发射信号的估计值。
图3是根据本公开的实施例示出的又一种信号检测网络的确定方法的示意流程图。如图3所示，在一些实施例中，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入；
所述方法还包括：
在步骤S301中，根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；
在步骤S302中，根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；
其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
在一个实施例中，可以设置多个级联的更新单元来构成初始神经网络，更新单元的输入包括单元公共输入和单元相关输入。其中，单元公共输入作为每个更新单元的输入，是保持不变的；单元相关输入基于更新单元不同而有所不同。
在一个实施例中，更新单元的输出可以是更新单元的单元相关输入的基于梯度下降法的更新值，例如针对n+1个更新单元中的第i个更新单元而言，单元相关输入为x i，输出为x i+1，n≥0，0≤i≤n。
接收端接收到的接收信号x与发射端发出的发射信号y之间的关系为：
y＝Hx+n；
这里的n是高斯白噪声，属于随机变量，H是信道矩阵。
而在考虑发射端与接收端通过IRS通信的情况下，例如样本发射端通过样本IRS与样本接收端通信，那么上式中的信道矩阵H，可以根据样本发射端到样本IRS的第一信道矩阵H 1、样本IRS的相位矩阵Φ、样本IRS到样本接收端的第二信道矩阵H 2、样本发射端到样本接收端的第三信道矩阵H 3来确定：
H＝H 2ΦH 1+H 3；
那么y＝(H 2ΦH 1+H 3)x+n；
由于在实际应用中接收端并不能直接确定发射信号x，而是需要通过信号检测来估计出
Figure PCTCN2021087901-appb-000001
本公开的实施例主要是为了求得
Figure PCTCN2021087901-appb-000002
使得x与
Figure PCTCN2021087901-appb-000003
之间的差距尽可能的小。
而为了求得更为接近x的
Figure PCTCN2021087901-appb-000004
可以根据梯度下降法对
Figure PCTCN2021087901-appb-000005
进行更新，在本实施例中通过更新单元对输入
Figure PCTCN2021087901-appb-000006
进行更新，例如对于第i个更新单元而言，可以对
Figure PCTCN2021087901-appb-000007
进行更新得到
Figure PCTCN2021087901-appb-000008
相对于
Figure PCTCN2021087901-appb-000009
更接近发射端实际发出的发射信号x。
基于梯度下降法，
Figure PCTCN2021087901-appb-000010
与
Figure PCTCN2021087901-appb-000011
之间的关系为：
Figure PCTCN2021087901-appb-000012
其中，
Figure PCTCN2021087901-appb-000013
η i为第i个更新单元对
Figure PCTCN2021087901-appb-000014
进行更新的步长，
Figure PCTCN2021087901-appb-000015
为根据需要预先设定的值。
展开公式1得到
Figure PCTCN2021087901-appb-000016
与
Figure PCTCN2021087901-appb-000017
之间的关系为：
Figure PCTCN2021087901-appb-000018
基于公式2可见，第i个更新单元对
Figure PCTCN2021087901-appb-000019
进行更新得到
Figure PCTCN2021087901-appb-000020
需要基于
Figure PCTCN2021087901-appb-000021
(H 2ΦΗ 1+H 3) Ty和(H 2ΦH 1+H 3) T(H 2ΦH 1+H 3)这三个量进行计算，因此可以将这3个量作为第i个更新单元的输入量，其中(H 2ΦH 1+H 3) Ty和(H 2ΦH 1+H 3) T(H 2ΦH 1+H 3)并不随着i改变，对于每个更新单元都是相同的，因此可以作为单元公共输入，而
Figure PCTCN2021087901-appb-000022
则会随着i改变，针对每个更新单元都有所不同，因此可以作为单元相关输入。
图4A是根据本公开的实施例示出的一种初始神经网络的局部结构示意图。图4B是根据本公开的实施例示出的另一种初始神经网络的局部结构示意图。
如图4A所示，针对第i个更新单元U i，输入为
Figure PCTCN2021087901-appb-000023
输出为
Figure PCTCN2021087901-appb-000024
可以直接将
Figure PCTCN2021087901-appb-000025
作为下一个更新单元，也即第i+1个更新单元U i+1的输入
Figure PCTCN2021087901-appb-000026
以此类推，直至得到
Figure PCTCN2021087901-appb-000027
如图4B所示，还可以在图4A所示结构的基础上，将第i个更新单元U i的输入
Figure PCTCN2021087901-appb-000028
与输出
Figure PCTCN2021087901-appb-000029
加权求和(每个更新单元在加权所用权值可以根据需要预先设定)，将加权求和结果作为第i+1个更新单元U i+1的输入
Figure PCTCN2021087901-appb-000030
虽然图4B所示结构相对于图4A所示结构较为复杂，但是可以在下一个更新单元的输入中，引入更多上一个单元的因素，有利于避免每次更新幅度过大，而严重偏离预期值。
在一个实施例中，所述更新单元包括两个单元公共输入；
所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定，例如上述公式2中的(H 2ΦH 1+H 3) T(H 2ΦH 1+H 3)；
所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定，例如上述公式2中的(H 2ΦH 1+H 3) Ty。
根据本公开的实施例，初始神经网络中的每个更新单元只需要3个输入，输入数量较少，使得网络中的连接数较少，有利于降低网络复杂度，使得网络适用于复杂信道环境中，也能够较快地输出结果。
图5A是根据本公开的实施例示出的一种更新单元的局部结构示意图。在一些实施例中，所述更新单元包括三层全连接层，所述三层全连接层包括输入层、隐藏层和输出层。
在一个实施例中，可以设计更新单元包括三层全连接层，三层全连接层可以具体包括输入层、隐藏层和输出层。
如图5A所示，主要示出了第i个更新单元U i中的输入层和隐藏层，其中，输入层对应的权重为w i1，对应的偏置为b i1，对应的激活函数为ρ；隐藏层对应的权重为w i2，对应的偏置为b i2，对应的激活函数为ψ。
输入层的输入主要根据发射信号确定，串联Concat是指串联方法，可以将输入该方法中的字符串、数组、向量等进行串联。可以将
Figure PCTCN2021087901-appb-000031
(H 2ΦH 1+H 3) T(H 2ΦH 1+H 3)与
Figure PCTCN2021087901-appb-000032
相乘的结果，以及(H 2ΦH 1+H 3) Ty三者输入串联Concat，则可以得到三者的串联结果。
例如发射端的天线数量为M(接收端已知)，发射信号可以是一个M*1的向量，相应地，
Figure PCTCN2021087901-appb-000033
就是一个M*1的向量，(H 2ΦH 1+H 3) T(H 2ΦH 1+H 3)与
Figure PCTCN2021087901-appb-000034
相乘的结果也是一个M*1的向量，(H 2ΦH 1+H 3) Ty也是一个M*1的向量，那么将三者串联的结果就是一个3M*1的向量，因此可以设置输入层的输入3M*1个。
输入在输入层中经过与w i1相乘后与b i1相加，再经过激活函数ρ后，输出到隐藏层，其中，激活函数ρ可以为sigmod函数。
而关于隐藏层的输入的数量，则可以根据需要设置，一般大于或等于输入层的输入的数量，例如可以为4M*1个。输入在隐藏层中经过与w i2相乘后与b i2相加，再经过激活函数ψ后，输出到输出，其中，激活函数ψ可以为tanh函数。
输出层的输出数量M*1，可以构成M*1的向量输入到下一个更新单元。
需要说明的是，激活函数并不限于上述实施例所举例的，具体可以根据需要进行选择，例如还可以选择relu函数作为激活函数。
图5B是根据本公开的实施例示出的另一种更新单元的局部结构示意图。如图5B所示，在一些实施例中，所述更新单元还包括短路直连结构，其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
在一个实施例中，仅仅依靠全连接网络，在训练过程中，难以充分挖掘复杂信道环境中的非线性关系，因此可以如图5B所示，在图5A所示更新单元的基础上，在全连接层中加入短路直连(short cut，也可以译作直连、捷径)结构，短路直连结构的起点为所述单元相关输入，终点为所述隐藏层，也即可以将
Figure PCTCN2021087901-appb-000035
与输入层的输出相加(可以是直接相加，也可以是加权求和)，并将相加的结果作为隐藏层的输入。
据此，便于在训练过程中充分挖掘复杂信道环境中的非线性关系，并且可以缓解多个更新单元级联而带来的梯度发散效应，有利于保证训练结果的合理性。
另外，基于上述实施例可知，在
Figure PCTCN2021087901-appb-000036
与
Figure PCTCN2021087901-appb-000037
的关系中，由三项构成，其中两项包含
Figure PCTCN2021087901-appb-000038
因此相对于其他输入，
Figure PCTCN2021087901-appb-000039
对
Figure PCTCN2021087901-appb-000040
的影响更大一些，而通过在全连接层中添加短路直连，使得
Figure PCTCN2021087901-appb-000041
(也即
Figure PCTCN2021087901-appb-000042
)不仅作为全连接层的初始输入，还可以作为全连接层中隐藏层的输入，确保了
Figure PCTCN2021087901-appb-000043
更新单元中可以具有更大的影响力，与
Figure PCTCN2021087901-appb-000044
和
Figure PCTCN2021087901-appb-000045
的关系相契合，也有利于保证训练结果的合理性。
在一个实施例中，基于训练样本集对上述实施例中更新单元级联构成的初始神经网络进行训练，可以得到信号检测网络。
其中，训练过程中所使用到的深度学习算法可以根据需要选择，例如可以选择Adam优化算法和端到端学习方式；训练过程中所使用的损失函数也可以根据需要设置，例如可以设置为
Figure PCTCN2021087901-appb-000046
也即从i＝1到i＝n，对
Figure PCTCN2021087901-appb-000047
进行加权求和，权重值为lgi。
在一些实施例中，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单元中的单元相关权重越大，所述更新单元在所述多个级联的更新单元中的排序越靠后，所述更新单元中的单元相关权重越小。
在一个实施例中，可以在更新单元的输入层中引入单元相关权重，也即与更新单元相关的权重。
图5C是根据本公开的实施例示出的又一种更新单元的局部结构示意图。如图5C所示，第i个更新单元U i的单元相关权重为β i，对于至少部分更新单元而言，U i在n个级联的更新单元中排序越靠前，也即i越小，β i就越大。
需要说明的是，单元相关权重为β i在全连接层中的位置可以根据需要设置，一般可以设置在输入层和/或隐藏层中，例如设置在输入层中，可以如图5C所示设置在w i1之后，与输入在输入层中经过与w i1相乘后的结果进一步相乘，而若设置在隐藏层中，则可以设置在w i2之后，与输入在隐藏层中经过与w i2相乘后的结果进一步相乘。
由于在基于梯度下降算法更新
Figure PCTCN2021087901-appb-000048
时，靠前的更新相对后靠后的更新，对于整体更新过程的影响更大，所以在通过多个更新单元对
Figure PCTCN2021087901-appb-000049
进行更新时，可以针对排序靠前的更新单元设置相对较大的单元相关权重，而对于排序靠后的更新单元则可以设置相对较小的单元相关权重，有利于降低训练过程的复杂度。
需要说明的是，有关单元相关权重的内容，并不限于如图5C所示，应用在图5B所示实施例的基础上，也可以直接应用在图5A所示实施例的基础上。
在一个实施例中，在排序高于预设次序的更新单元中，单元相关权重为预设值。 在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大，该更新单元在所述多个级联的更新单元中的排序越靠后，该更新单元中的单元相关权重越小。
在一个实施例中，可以设置单元相关中针对部分更新单元保持不变，仅针对另一部分更新单元随排序改变，例如针对排序高于预设次序的更新单元，单元相关权重保持为预设值不变，而针对排序低于或等于预设次序的更新单元，则可以设置单元相关权重随着排序靠后而降低。
据此，在确保排序靠前的更新单元的重要性的基础上，可以减少对一部分单元相关权重的变化，有利于降低训练过程的复杂度。
例如单元相关权重β i可以是更新单元排序i的半指数函数，其形式例如可以如下所示：
Figure PCTCN2021087901-appb-000050
也即在n个更新单元中，前一半更新单元中的单元相关权重可以保持不变，而后一半更新单元中的单元相关权重可以随着更新单元的排序i靠后而降低。
本公开的实施例还提出了一种信号检测方法，所述信号检测方法可以由发射端与接收端通信过程中的接收端实现。
在一个实施例中，所述发射端和接收端可以是终端、基站、卫星、无人飞行器、核心网等通信设备。例如在基站与终端的通信过程中，接收端，既可以是基站，也可以是终端，例如发射端为基站时，那么接收端为终端，发射端为终端时，接收端为基站。
在一个实施例中，所述终端包括但不限于手机、平板电脑、可穿戴设备、传感器、物联网设备等通信装置。所述终端可以作为用户设备与基站通信，所述基站包括但不限于4G基站、5G基站、6G基站。
所述信号检测方法可以包括以下步骤：
响应于接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经 过所述IRS转换后的信号，根据上述任一实施例所述方法确定的信号检测网络确定所述发射信号。
在一个实施例中，在基于前述实施例确定了信号检测网络后，可以在后续实际使用接收端和发射端通过IRS进行通信的场景中，通过信号检测网络进行信号检测，以确定发射端发出的发射信号。
例如IRS每次调整其中无源元件后，都可以将IRS整体的相位矩阵告知接收端，从而接收端可以确定IRS的相位矩阵。接收端还可以根据发射端通过IRS向接收端发射信号所使用的信道的信道状态信息，确定所述发射端到IRS的信道矩阵，以及IRS到接收端的信道矩阵，并根据发射端直接向接收端发射信号所使用的信道的信道状态信息确定样本发射端到样本接收端的信道矩阵。然后接收端可以基于接收到的接收信号、上述相位矩阵和三个信道矩阵确定输入量输(例如与图4A图4B所示实施例的3个输入量相对应的输入量)入到上述实施例确定的信号检测网络中，以得到对发射端发射信号的估计值。
需要说明的是，确定信号检测网络的操作可以由接收端执行，也可由其他设备执行，对此，本公开的实施例并不限制，例如由其他设备执行，那么可以在得到信号检测网络后，将信号检测网络发送给接收端。
与前述的信号检测网络的确定方法的实施例相对应，本公开还提供了信号检测网络的确定装置的实施例。
图6是根据本公开的实施例示出的一种信号检测网络的确定装置的示意框图。本实施例所示的确定装置可以用于确定信号检测网络，信号检测网络可以供接收端进行信号检测，例如在发射端发出发射信号后，接收端可以接收到响应的接收信号，接收端可以基于信号检测网络估计发射信号。
在一个实施例中，所述发射端和接收端可以是终端、基站、卫星、无人飞行器、核心网等通信设备。以下实施例主要在发射端为基站，接收端为终端的情况下对本公开进行示例性说明。
在一个实施例中，所述终端包括但不限于手机、平板电脑、可穿戴设备、传感器、物联网设备等通信装置。所述终端可以作为用户设备与基站通信，所述基站包括但不限于4G基站、5G基站、6G基站。
如图6所示，所述信号检测网络的确定装置可以包括：参数确定模块601，被 配置为确定样本发射端和样本接收端通过样本智能反射面IRS进行通信的样本通信参数；网络训练模块602，被配置为基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
在一些实施例中，所述样本通信参数包括以下至少之一：所述样本发射端到所述样本IRS的第一信道矩阵；所述样本IRS的相位矩阵；所述样本IRS到所述样本接收端的第二信道矩阵；所述样本发射端到所述样本接收端的第三信道矩阵；以及所述样本接收端的样本接收信号。
图7是根据本公开的实施例示出的又一种信号检测网络的确定装置的示意框图。如图7所示，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入。
所述装置还包括：关系确定模块701，被配置为根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；输入输出确定模块702，被配置为根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
在一些实施例中，所述更新单元包括两个单元公共输入；所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定；所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定。
在一些实施例中，所述更新单元包括三层全连接层，所述三层全连接层包括输入层、隐藏层和输出层。
在一些实施例中，所述更新单元还包括短路直连结构，其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
在一些实施例中，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；
其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单 元中的单元相关权重越大，所述更新单元在所述多个级联的更新单元中的排序越靠后，所述更新单元中的单元相关权重越小。
在一些实施例中，在排序高于预设次序的更新单元中，单元相关权重为预设值；
在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大，该更新单元在所述多个级联的更新单元中的排序越靠后，该更新单元中的单元相关权重越小。
本公开的实施例还提出一种信号检测装置，所述信号检测装置可以应用于发射端与接收端通信过程中的接收端。
在一个实施例中，所述发射端和接收端可以是终端、基站、卫星、无人飞行器、核心网等通信设备。例如在基站与终端的通信过程中，接收端，既可以是基站，也可以是终端，例如发射端为基站时，那么接收端为终端，发射端为终端时，接收端为基站。
在一个实施例中，所述终端包括但不限于手机、平板电脑、可穿戴设备、传感器、物联网设备等通信装置。所述终端可以作为用户设备与基站通信，所述基站包括但不限于4G基站、5G基站、6G基站。
所述信号检测装置可以包括：
信号估计模块，被配置为响应于接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据上述任一实施例所述装置确定的信号检测网络确定所述发射信号。
关于上述实施例中的装置，其中各个模块执行操作的具体方式已经在相关方法的实施例中进行了详细描述，此处将不做详细阐述说明。
对于装置实施例而言，由于其基本对应于方法实施例，所以相关之处参见方法实施例的部分说明即可。以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的模块可以是或者也可以不是物理上分开的，作为模块显示的部件可以是或者也可以不是物理模块，即可以位于一个地方，或者也可以分布到多个网络模块上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性劳动的情况下，即可以理解并实施。
本公开的实施例还提出一种通信装置，包括：处理器；用于存储计算机程序的存储器；其中，当所述计算机程序被处理器执行时，实现上述任一实施例所述的信号检测网络的确定方法。
本公开的实施例还提出一种通信装置，包括：处理器；用于存储计算机程序的存储器；其中，当所述计算机程序被处理器执行时，实现上述实施例所述的信号检测方法。
本公开的实施例还提出一种计算机可读存储介质，用于存储计算机程序，当所述计算机程序被处理器执行时，实现如上述任一实施例所述的信号检测网络的确定方法中的步骤。
本公开的实施例还提出一种计算机可读存储介质，用于存储计算机程序，当所述计算机程序被处理器执行时，实现如上述实施例所述的信号检测方法中的步骤。
如图8所示，图8是根据本公开的实施例示出的另一种用于确定信号检测网络的装置800的示意框图。装置800可以被提供为一基站。参照图8，装置800包括处理组件822、无线发射/接收组件824、天线组件826、以及无线接口特有的信号处理部分，处理组件822可进一步包括一个或多个处理器。处理组件822中的其中一个处理器可以被配置为实现上述任一实施例所述的信号检测网络的确定方法。
在所述装置800作为接收端时，所述处理组件822中的其中一个处理器可以可以被配置为实现上述实施例所述的基于所述信号检测网络进行信号检测的步骤。
图9是根据本公开的实施例示出的一种用于确定信号检测网络的装置900的示意框图。例如，装置900可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等。
参照图9，装置900可以包括以下一个或多个组件：处理组件902，存储器904，电源组件906，多媒体组件908，音频组件910，输入/输出(I/O)的接口912，传感器组件914，以及通信组件916。
处理组件902通常控制装置900的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件902可以包括一个或多个处理器920来执行指令，以完成上述信号检测网络的确定方法的全部或部分步骤。在所述装置900作为接收端是，处理器920还可以执行指令来完成实施例所述的基于所述信号检测网络进行信号检测的步骤
此外，处理组件902可以包括一个或多个模块，便于处理组件902和其他组件之间的交互。例如，处理组件902可以包括多媒体模块，以方便多媒体组件908和处理组件902之间的交互。
存储器904被配置为存储各种类型的数据以支持在装置900的操作。这些数据的示例包括用于在装置900上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器904可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件906为装置900的各种组件提供电力。电源组件906可以包括电源管理***，一个或多个电源，及其他与为装置900生成、管理和分配电力相关联的组件。
多媒体组件908包括在所述装置900和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件908包括一个前置摄像头和/或后置摄像头。当装置900处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜***或具有焦距和光学变焦能力。
音频组件910被配置为输出和/或输入音频信号。例如，音频组件910包括一个麦克风(MIC)，当装置900处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器904或经由通信组件916发送。在一些实施例中，音频组件910还包括一个扬声器，用于输出音频信号。
I/O接口912为处理组件902和***接口模块之间提供接口，上述***接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件914包括一个或多个传感器，用于为装置900提供各个方面的状态 评估。例如，传感器组件914可以检测到装置900的打开/关闭状态，组件的相对定位，例如所述组件为装置900的显示器和小键盘，传感器组件914还可以检测装置900或装置900一个组件的位置改变，用户与装置900接触的存在或不存在，装置900方位或加速/减速和装置900的温度变化。传感器组件914可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件914还可以包括光传感器，如CMOS或CCD图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件914还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件916被配置为便于装置900和其他设备之间有线或无线方式的通信。装置900可以接入基于通信标准的无线网络，如WiFi，2G或3G，4G LTE、5G NR或它们的组合。在一个示例性实施例中，通信组件916经由广播信道接收来自外部广播管理***的广播信号或广播相关信息。在一个示例性实施例中，所述通信组件916还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
在示例性实施例中，装置900可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述信号检测网络的确定方法。
在示例性实施例中，还提供了一种包括指令的非临时性计算机可读存储介质，例如包括指令的存储器904，上述指令可由装置900的处理器920执行以完成上述信号检测网络的确定方法。例如，所述非临时性计算机可读存储介质可以是ROM、随机存取存储器(RAM)、CD-ROM、磁带、软盘和光数据存储设备等。
本领域技术人员在考虑说明书及实践这里公开的公开后，将容易想到本公开的其它实施方案。本公开旨在涵盖本公开的任何变型、用途或者适应性变化，这些变型、用途或者适应性变化遵循本公开的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的，本公开的真正范围和精神由下面的权利要求指出。
应当理解的是，本公开并不局限于上面已经描述并在附图中示出的精确结构，并且可以在不脱离其范围进行各种修改和改变。本公开的范围仅由所附的权利要求来 限制。
需要说明的是，在本文中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
以上对本公开实施例所提供的方法和装置进行了详细介绍，本文中应用了具体个例对本公开的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本公开的方法及其核心思想；同时，对于本领域的一般技术人员，依据本公开的思想，在具体实施方式及应用范围上均会有改变之处，综上所述，本说明书内容不应理解为对本公开的限制。

Claims (22)

  1. 一种信号检测网络的确定方法，其特征在于，包括：
    确定样本发射端和样本接收端通过样本智能反射面IRS进行通信的样本通信参数；
    基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
  2. 根据权利要求1所述的方法，其特征在于，所述样本通信参数包括以下至少之一：
    所述样本发射端到所述样本IRS的第一信道矩阵；
    所述样本IRS的相位矩阵；
    所述样本IRS到所述样本接收端的第二信道矩阵；
    所述样本发射端到所述样本接收端的第三信道矩阵；以及
    所述样本接收端的样本接收信号。
  3. 根据权利要求2所述的方法，其特征在于，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入；
    所述方法还包括：
    根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；
    根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；
    其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
  4. 根据权利要求3所述的方法，其特征在于，所述更新单元包括两个单元公共输入；
    所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定；
    所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定。
  5. 根据权利要求3所述的方法，其特征在于，所述更新单元包括三层全连接层，所述三层全连接层包括输入层、隐藏层和输出层。
  6. 根据权利要求5所述的方法，其特征在于，所述更新单元还包括短路直连结构， 其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
  7. 根据权利要求5所述的方法，其特征在于，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；
    其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单元中的单元相关权重越大。
  8. 根据权利要求7所述的方法，其特征在于，在排序高于预设次序的更新单元中，单元相关权重为预设值；
    在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大。
  9. 一种信号检测方法，其特征在于，包括：
    接收来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据权利要求1至8中任一项所述方法确定信号检测网络确定所述发射信号。
  10. 一种信号检测网络的确定装置，其特征在于，包括：
    参数确定模块，被配置为确定样本发射端和样本接收端通过样本智能反射面IRS进行通信的样本通信参数；
    网络训练模块，被配置为基于所述样本通信参数构成的训练样本集对初始神经网络进行训练以得到信号检测网络，其中，所述初始神经网络的输入为所述样本通信参数，所述初始神经网络的输出为对样本发射端的发射信号的估计值。
  11. 根据权利要求10所述的装置，其特征在于，所述样本通信参数包括以下至少之一：
    所述样本发射端到所述样本IRS的第一信道矩阵；
    所述样本IRS的相位矩阵；
    所述样本IRS到所述样本接收端的第二信道矩阵；
    所述样本发射端到所述样本接收端的第三信道矩阵；以及
    所述样本接收端的样本接收信号。
  12. 根据权利要求11所述的装置，其特征在于，所述初始神经网络包括多个级联的更新单元，所述更新单元的输入包括单元公共输入和单元相关输入；
    所述装置还包括：
    关系确定模块，被配置为根据梯度下降法确定所述单元相关输入与所述单元相关输入的更新值之间的关系；
    输入输出确定模块，被配置为根据所述关系中用于表征所述更新值的参数确定所述单元公共输入，根据所述更新值确定所述更新单元的输出；
    其中，所述更新单元的输出作为级联的下一个更新单元的单元相关输入。
  13. 根据权利要求12所述的装置，其特征在于，所述更新单元包括两个单元公共输入；
    所述两个单元公共输入中的第一单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵和所述第三信道矩阵确定；
    所述两个单元公共输入中的第二个单元公共输入基于所述第一信道矩阵、所述相位矩阵、所述第二信道矩阵、所述第三信道矩阵和所述样本接收信号确定。
  14. 根据权利要求12所述的装置，其特征在于，所述三层全连接层包括输入层、隐藏层和输出层。
  15. 根据权利要求14所述的装置，其特征在于，所述更新单元还包括短路直连结构，其中，所述短路直连结构的起点为所述单元相关输入，终点为所述隐藏层。
  16. 根据权利要求14所述的装置，其特征在于，所述全连接层的网络参数包括单元相关权重，且至少部分所述更新单元中的单元相关权重与所述更新单元在所述多个级联的更新单元中的排序相关；
    其中，所述更新单元在所述多个级联的更新单元中的排序越靠前，所述更新单元中的单元相关权重越大。
  17. 根据权利要求16所述的装置，其特征在于，在排序高于预设次序的更新单元中，单元相关权重为预设值；
    在排序低于或等于预设次序的更新单元中，单元相关权重小于所述预设值，且该更新单元在所述多个级联的更新单元中的排序越靠前，该更新单元中的单元相关权重越大。
  18. 一种信号检测装置，其特征在于，包括：
    信号估计模块，被配置为响应于接收到来自IRS的接收信号，所述接收信号为发射端发出的发射信号经过所述IRS转换后的信号，根据权利要求10至17中任一项所述装置确定的信号检测网络确定所述发射信号。
  19. 一种通信装置，其特征在于，包括：
    处理器；
    用于存储计算机程序的存储器；
    其中，当所述计算机程序被处理器执行时，实现如权利要求1至8中任一项所述的信号检测网络的确定方法。
  20. 一种通信装置，其特征在于，包括：
    处理器；
    用于存储计算机程序的存储器；
    其中，当所述计算机程序被处理器执行时，实现如权利要求9所述的信号检测方法。
  21. 一种计算机可读存储介质，用于存储计算机程序，其特征在于，当所述计算机程序被处理器执行时，实现如权利要求1至8中任一项所述的信号检测网络的确定方法中的步骤。
  22. 一种计算机可读存储介质，用于存储计算机程序，其特征在于，当所述计算机程序被处理器执行时，实现如权利要求9所述的信号检测方法中的步骤。
PCT/CN2021/087901 2021-04-16 2021-04-16 信号检测网络的确定方法和装置 WO2022217609A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
CN202180000927.3A CN115769551A (zh) 2021-04-16 2021-04-16 信号检测网络的确定方法和装置
PCT/CN2021/087901 WO2022217609A1 (zh) 2021-04-16 2021-04-16 信号检测网络的确定方法和装置
US18/287,104 US20240137074A1 (en) 2021-04-16 2021-04-16 Method and apparatus for determining signal detection network
EP21936484.1A EP4325778A4 (en) 2021-04-16 2021-04-16 METHOD AND APPARATUS FOR DETERMINING A SIGNAL DETECTION NETWORK

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/087901 WO2022217609A1 (zh) 2021-04-16 2021-04-16 信号检测网络的确定方法和装置

Publications (1)

Publication Number Publication Date
WO2022217609A1 true WO2022217609A1 (zh) 2022-10-20

Family

ID=83639445

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/087901 WO2022217609A1 (zh) 2021-04-16 2021-04-16 信号检测网络的确定方法和装置

Country Status (4)

Country Link
US (1) US20240137074A1 (zh)
EP (1) EP4325778A4 (zh)
CN (1) CN115769551A (zh)
WO (1) WO2022217609A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111817768A (zh) * 2020-06-03 2020-10-23 北京交通大学 一种用于智能反射表面无线通信的信道估计方法
KR102192234B1 (ko) * 2019-10-28 2020-12-17 성균관대학교 산학협력단 지능형 반사 평면을 포함하는 무선 통신 시스템의 통신 방법 및 이를 위한 장치
CN112134816A (zh) * 2020-09-27 2020-12-25 杭州电子科技大学 一种基于智能反射表面的elm-ls联合信道估计方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10460747B2 (en) * 2016-05-10 2019-10-29 Google Llc Frequency based audio analysis using neural networks

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102192234B1 (ko) * 2019-10-28 2020-12-17 성균관대학교 산학협력단 지능형 반사 평면을 포함하는 무선 통신 시스템의 통신 방법 및 이를 위한 장치
CN111817768A (zh) * 2020-06-03 2020-10-23 北京交通大学 一种用于智能反射表面无线通信的信道估计方法
CN112134816A (zh) * 2020-09-27 2020-12-25 杭州电子科技大学 一种基于智能反射表面的elm-ls联合信道估计方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
ELBIR AHMET M.; PAPAZAFEIROPOULOS ANASTASIOS; KOURTESSIS PANDELIS; CHATZINOTAS SYMEON: "Deep Channel Learning for Large Intelligent Surfaces Aided mm-Wave Massive MIMO Systems", IEEE WIRELESS COMMUNICATIONS LETTERS, IEEE, PISCATAWAY, NJ, USA, vol. 9, no. 9, 8 May 2020 (2020-05-08), Piscataway, NJ, USA , pages 1447 - 1451, XP011807800, ISSN: 2162-2337, DOI: 10.1109/LWC.2020.2993699 *
LIU SHIYU, LEI MING, ZHAO MIN-JIAN: "Deep Learning Based Channel Estimation for Intelligent Reflecting Surface Aided MISO-OFDM Systems", 2020 IEEE 92ND VEHICULAR TECHNOLOGY CONFERENCE (VTC2020-FALL), IEEE, 1 November 2020 (2020-11-01) - 16 December 2020 (2020-12-16), pages 1 - 5, XP055977125, ISBN: 978-1-7281-9484-4, DOI: 10.1109/VTC2020-Fall49728.2020.9348697 *
See also references of EP4325778A4 *

Also Published As

Publication number Publication date
CN115769551A (zh) 2023-03-07
EP4325778A4 (en) 2024-06-12
US20240137074A1 (en) 2024-04-25
EP4325778A1 (en) 2024-02-21

Similar Documents

Publication Publication Date Title
US10930304B2 (en) Processing voice
KR102365890B1 (ko) 신경망의 훈련 방법, 장치 및 저장 매체
US20200387795A1 (en) Super network training method and device
US10885905B2 (en) Predicting user actions on ubiquitous devices
TWI769635B (zh) 網路訓練、行人重識別方法、電子設備及電腦可讀存儲介質
WO2021196401A1 (zh) 图像重建方法及装置、电子设备和存储介质
TWI772668B (zh) 一種目標對象處理方法、裝置、電子設備及儲存介質
CN110970046B (zh) 一种音频数据处理的方法及装置、电子设备、存储介质
US11337173B2 (en) Method and device for selecting from a plurality of beams
CN111160448B (zh) 一种图像分类模型的训练方法及装置
US10027785B2 (en) Method for switching screen state of terminal, terminal thereof, and computer-readable medium thereof
US20230217212A1 (en) Ranging capacity request method, ranging capacity sending method, and ranging capacity receiving method
CN109698794A (zh) 一种拥塞控制方法、装置、电子设备及存储介质
CN106250430A (zh) 智能设备列表的排序方法及装置
WO2020192113A1 (zh) 图像处理方法及装置、电子设备和存储介质
CN114095437B (zh) 发送数据包的方法、装置、电子设备和存储介质
RU2721181C1 (ru) Способ построения суперсети, способ использования, устройство и носитель информации
WO2018201439A1 (zh) 随机接入方法及装置、用户设备和计算机可读存储介质
WO2022217609A1 (zh) 信号检测网络的确定方法和装置
CN109214175A (zh) 基于样本特征训练分类器的方法、装置及存储介质
US20210065004A1 (en) Method and device for subnetwork sampling, and method and device for building a hypernetwork topology
WO2022261842A1 (zh) 预编码矩阵确定方法、装置、用户设备、基站及存储介质
WO2020224448A1 (zh) 交互方法及装置、音箱、电子设备和存储介质
CN107317599B (zh) 一种自干扰信号的消除方法及装置
WO2023056646A1 (zh) 一种信道状态信息上报方法、装置及存储介质

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: 21936484

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18287104

Country of ref document: US

WWE Wipo information: entry into national phase

Ref document number: 2021936484

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2021936484

Country of ref document: EP

Effective date: 20231116