Joint channel estimation and detection method and device for unlicensed access of multi-beam satellites
Through the target neural network model, the pilot matrix and pilot sequence are processed, and the advantages of model-driven and data-driven are complementary, the problem of low accuracy in the joint channel estimation and detection of multi-beam satellite authorization-free access is solved, and the detection performance is improved.
Patent Information
- Application Number
- CN202411196282.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-29
AI Technical Summary
The existing joint channel estimation and detection methods for authorization-free access of multi-beam satellites have the problem of low accuracy of detection results, especially when the channel prior distribution is unknown or time-varying, the signal estimation time is long and the time complexity is high.
The target neural network model is used to process the predefined pilot matrix and pilot sequence. The model-driven network module and the data-driven network module in the multi-layer neural network unit complement each other, reducing the dependence on prior information and improving detection performance.
The accuracy of joint channel estimation and detection of multi-beam satellites without authorization is improved, the dependence on prior information is reduced, the detection performance is improved, and the problem of low accuracy of detection results is alleviated.
Smart Images

Figure CN119094281A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of communications, and in particular to a joint channel estimation and detection method and device for multi-beam satellite unlicensed access. Background Art
[0002] Massive Machine Type Communication (mMTC) has become an integral part of the fifth generation (5G) communications and will continue to play a key role in the sixth generation (6G) communication systems to enable seamless connectivity for Internet of Things (IoT) applications. However, existing cellular networks are mainly deployed in densely populated areas. Therefore, supporting access to mMTC devices in remote areas such as oceans and mountains remains a major challenge for current terrestrial networks due to limited coverage. To this end, satellite communication technology offers a promising solution, especially in the context of low earth orbit (LEO) satellites, to support the massively exploding connectivity demand and provide global wide-area coverage and high-rate communications. However, satellite-aided mMTC technology has encountered several obstacles. For example, scarce spectrum resources and interference caused by the increasing number of ground devices have severely affected the efficiency of satellite connectivity. Considering the sporadic activities of IoT applications, unlicensed random access technology can be applied to satellite IoT systems. Specifically, compared with the classic license-based random access, the preamble overhead and access delay of the unlicensed scheme can be greatly reduced.
[0003] A key task of uplink unlicensed random access is to perform joint activity detection and channel estimation (JADCE) based on the pilot sequence received at the LEO satellite side. By leveraging the sporadic activity characteristics in mMTC scenarios, JADCE can be formulated as a problem of solving a single measurement vector or multiple measurement vectors (MMV), depending on the number of receiving antennas or beams. Compressed sensing (CS) techniques have been developed to address these problems and can be divided into three categories: convex optimization-based methods, greedy algorithms, and Bayesian methods. However, most of these works adopt an iterative structure and depend on the prior distribution of the channel, which may be unknown or time-varying in practice. On the other hand, the signal estimation time is limited by the number of iterations and has a high time complexity. In other words, the existing joint channel estimation and detection methods for multi-beam satellite unlicensed access have the technical problem of low accuracy of detection results. Summary of the invention
[0004] The object of the present invention is to provide a joint channel estimation and detection method and device for unlicensed access of multi-beam satellites, so as to alleviate the technical problem of low accuracy of detection results in the existing joint channel estimation and detection method for unlicensed access of multi-beam satellites.
[0005] In a first aspect, the present invention provides a joint channel estimation and detection method for unlicensed access of multi-beam satellites, comprising: obtaining a predefined pilot matrix, a pilot sequence received by the multi-beam satellite, and the number of beams generated by the multi-beam satellite; wherein the predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite; using a target neural network model to process the predefined pilot matrix and the pilot sequence to obtain an equivalent channel matrix for all ground users; wherein the target neural network model comprises multiple layers of neural network units, each layer of the neural network unit comprises: a model-driven network module and a data-driven network module, the input of the model-driven network module comprises: the input of the target neural network model and the output of the data-driven network module in the neural network unit of the previous layer, and the output of the model-driven network module is the input of the data-driven network module of this layer; based on the equivalent channel matrix, the number of beams, and a preset active decision threshold, determining the beam space channel matrix and the active matrix of all ground users, and using the beam space channel matrix and the active matrix as the joint channel estimation and detection results for unlicensed access of multi-beam satellites.
[0006] In an optional embodiment, it also includes: obtaining multiple groups of training samples; wherein each group of training samples includes: a predefined pilot matrix, pilot sequence samples received by a multi-beam satellite, beam space channel matrix samples and active matrix samples of all ground users; the beam space channel matrix samples and active matrix samples of all ground users are sample labels; inputting the multiple groups of training samples into the initial neural network model to perform iterative training on the initial neural network model module by module and layer by layer to obtain a target neural network model; wherein, in the process of iterative training module by module and layer by layer, for the nth layer neural network unit in the initial neural network model, firstly update the network parameters of the model-driven network module based on the output result of its model-driven network module, then update the network parameters of the nth layer neural network unit based on the output result of its data-driven network module, and finally update the network parameters of the 1st to nth layer neural network units based on the output result of the nth layer neural network unit after the network parameters are updated.
[0007] In an optional embodiment, the network parameters of the model-driven network module are updated based on the output results of its model-driven network module, including: obtaining the output results of the target model-driven network module after the target training sample is input into the initial neural network model; wherein the target training sample represents any sample in multiple groups of training samples; the target model-driven network module represents the model-driven network module in the nth layer neural network unit; based on the beam space channel matrix samples and active matrix samples in the target training sample, the equivalent channel matrix samples corresponding to the target training sample are calculated; based on the output results of the target model-driven network module corresponding to all training samples and the equivalent channel matrix samples, the first loss function value is calculated; based on the first loss function value, the network parameters of the target model-driven network module are updated.
[0008] In an optional embodiment, the network parameters of the n-th layer neural network unit are updated based on the output results of its data-driven network module, including: inputting the target training sample into the neural network model after the network parameters of the target model-driven network module are updated, and obtaining the output results of the target data-driven network module; wherein the target data-driven network module represents the data-driven network module in the n-th layer neural network unit; calculating the second loss function value based on the output results of the target data-driven network module corresponding to all training samples and the equivalent channel matrix samples; and updating the network parameters of the n-th layer neural network unit based on the second loss function value.
[0009] In an optional embodiment, the network parameters of the 1st to nth layer neural network units are updated based on the output results of the nth layer neural network units after the network parameters are updated, including: inputting the target training samples into the neural network model after the network parameters of the nth layer neural network units are updated, and obtaining the output results of the nth layer neural network units; calculating the third loss function value based on the output results of the nth layer neural network units corresponding to all training samples and the equivalent channel matrix samples; and updating the network parameters of the 1st to nth layer neural network units based on the third loss function value.
[0010] In an optional implementation, the dimension of the equivalent channel matrix is K×N RF , K represents the total number of ground users of multi-beam satellites, N RFrepresents the number of beams; based on the equivalent channel matrix, the number of beams and the preset active decision threshold, the beam space channel matrix and the active matrix of all ground users are determined, including: based on the maximum value of the elements in the equivalent channel matrix, the minimum value and the maximum value of the absolute value of the non-zero elements, the channel gain threshold is determined; based on the elements in the k-th row of the equivalent channel matrix, the channel gain threshold, the number of beams and the preset active decision threshold, the activity indicator of the k-th ground user of the multi-beam satellite is determined; if the value of the activity indicator of the k-th ground user is 1, it is determined that the k-th ground user is in an active state, and the elements in the k-th row of the equivalent channel matrix are used as the beam space channel estimation value of the k-th ground user; if the value of the activity indicator of the k-th ground user is 0, it is determined that the k-th ground user is in an inactive state, and the beam space channel estimation value of the k-th ground user is 0; constructing an active matrix based on the activity indicators of all ground users, and constructing a beam space channel matrix based on the beam space channel estimation values of all ground users.
[0011] In an optional embodiment, determining the activity indicator of the kth ground user of the multi-beam satellite based on the elements of the kth row in the equivalent channel matrix, the channel gain threshold, the number of beams, and the preset active decision threshold includes: using the formula Determine an activity indicator for the kth ground user of a multi-beam satellite; wherein, represents the activity indicator of the kth terrestrial user, represents the element in the kth row and rth column of the equivalent channel matrix, represents the channel gain threshold, υ represents the ratio of the minimum and maximum absolute values of non-zero elements in the equivalent channel matrix, λ th Indicates the preset active decision threshold.
[0012] In a second aspect, the present invention provides a joint channel estimation and detection device for unlicensed access of multi-beam satellites, including a first acquisition module for acquiring a predefined pilot matrix, a pilot sequence received by the multi-beam satellite, and the number of beams generated by the multi-beam satellite; wherein the predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite; a processing module for processing the predefined pilot matrix and the pilot sequence using a target neural network model to obtain an equivalent channel matrix for all ground users; wherein the target neural network model includes multiple layers of neural network units, each layer of the neural network unit includes: a model-driven network module and a data-driven network module, the input of the model-driven network module includes: the input of the target neural network model and the output of the data-driven network module in the neural network unit of the previous layer, and the output of the model-driven network module is the input of the data-driven network module of this layer; a determination module for determining the beam space channel matrix and the active matrix of all ground users based on the equivalent channel matrix, the number of beams, and a preset active decision threshold, and using the beam space channel matrix and the active matrix as the joint channel estimation and detection results for unlicensed access of multi-beam satellites.
[0013] In a third aspect, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the joint channel estimation and detection method for unauthorized access of multi-beam satellites in any of the aforementioned implementation modes are implemented.
[0014] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implements the joint channel estimation and detection method for unauthorized access of a multi-beam satellite according to any one of the aforementioned implementations.
[0015] The joint channel estimation and detection method for unlicensed access of multi-beam satellites provided by the present invention uses a target neural network model to process a predefined pilot matrix and a pilot sequence received by a multi-beam satellite to obtain an equivalent channel matrix for all ground users, and then determines the joint channel estimation and detection result for unlicensed access of multi-beam satellites based on the equivalent channel matrix, the number of beams generated by the multi-beam satellite, and a preset active decision threshold. Based on the structure of the target neural network model, it can be seen that each neural network unit in the model further extracts features using a data-driven network based on the output of each model-driven network. Therefore, the target neural network model can realize the complementary advantages of model-driven and data-driven, which not only reduces the model's dependence on prior information, but also achieves better detection performance, thereby alleviating the technical problem of low accuracy of detection results in the existing joint channel estimation and detection method for unlicensed access of multi-beam satellites. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0017] Figure 1 A flowchart of a joint channel estimation and detection method for multi-beam satellite unlicensed access provided by an embodiment of the present invention;
[0018] Figure 2 A schematic diagram of a multi-beam satellite unlicensed access system provided by an embodiment of the present invention;
[0019] Figure 3 A schematic diagram of the structure of a target neural network model provided by an embodiment of the present invention;
[0020] Figure 4 A functional module diagram of a joint channel estimation and detection device for multi-beam satellite unlicensed access provided by an embodiment of the present invention;
[0021] Figure 5 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.
[0023] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0024] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.
[0025] Embodiment 1
[0026] Figure 1 A flowchart of a joint channel estimation and detection method for multi-beam satellite unlicensed access provided by an embodiment of the present invention, such as Figure 1 As shown, the method specifically comprises the following steps:
[0027] Step S102: obtaining a predefined pilot matrix, a pilot sequence received by the multi-beam satellite, and the number of beams generated by the multi-beam satellite.
[0028] The predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite.
[0029] Figure 2 Schematic diagram of a multi-beam satellite unlicensed access system. Figure 2 As shown, the multi-beam satellite unlicensed access system consists of a multi-beam satellite, K ground users and a gateway. Each ground user is equipped with a single antenna. Due to the large number of ground users and limited satellite spectrum resources, the embodiment of the present invention considers that all ground users of the multi-beam satellite occupy the same frequency band resources at the same time, and the access process is unlicensed. The antenna panel equipped with the multi-beam satellite has M antenna units, and the interval between the antenna units is 1 / 2 of the wavelength. The feeder link connects the satellite and the gateway, and the gateway is used to configure the multi-beam satellite.
[0030] In the downlink transmission process, the multi-beam satellite adopts the multi-beam forming technology with hybrid precoding architecture. Without loss of generality, according to the number of beams N generated by the multi-beam satellite RF Divide the satellite coverage area into N RF cells (ground users are evenly distributed in N RF The uplink access method for ground users considers the most common scenario, that is, each cell is allocated one RF chain.
[0031] Based on the predefined pilot length P, each ground user can send a unique pilot sequence to the satellite. The pilot sequence of the kth ground user is expressed as Among them, s P,k represents the Pth pilot sent by the kth user, represents a complex vector of length P. In satellite Internet, the length of the pilot sequence is usually much smaller than the number of ground users, that is, P<<K, and the pilots are independent between different ground users (that is, the pilots sent by ground users are different).
[0032] When multi-beam satellites are unlicensed accessed, the multi-user non-orthogonal signals received by the satellite are expressed in the spatial domain as follows: Where Y represents the pilot sequence received by the multi-beam satellite, That is, the satellite passes N RFThe RF link receives the pilot sequence (known parameter). S represents the matrix composed of the pilot vectors of all ground users, that is, the predefined pilot matrix (known parameter). X represents the equivalent channel matrix (unknown parameter) of all ground users, Represents the equivalent combined noise matrix (unknown parameter) superimposed on the received signal.
[0033] Among them, the equivalent channel matrix of all ground users is X = diag(α1,…,α k ,…α K )H, in uplink access, the activities of all ground users are independent, α=α1,…,α K represents the active matrix of all ground users, H represents the beam space channel matrix of all ground users, and the joint channel estimation and detection method for multi-beam satellite unlicensed access provided by the embodiment of the present invention is to recover the active matrix and beam space channel matrix of all ground users from the pilot sequence Y received by the satellite and the known predefined pilot matrix S. That is to say, the above two matrices are the joint channel estimation and detection results to be determined in the embodiment of the present invention. Wherein, α k represents the activity indicator of the kth ground user. If the kth ground user is active, then α k =1; if the kth ground user is inactive, then α k =0. represents the beam space channel estimation value of the kth ground user.
[0034] It is known that for satellite channels, W RF The conjugate transpose of W RF represents the downlink precoding matrix of the channel, h k (p,f) represents the channel model when the kth ground user sends the pth pilot at the communication frequency f. In the multi-beam satellite access architecture, the downlink precoding matrix of the channel is During uplink access, it can be regarded as a receiving combiner. According to the channel reciprocity, the uplink analog combiner also adopts the downlink precoding matrix of the channel.
[0035] in, represents the combiner vector of the RF chain corresponding to the rth cell, represents the azimuth from the center of the rth cell to the satellite, Represents the elevation angle from the center of the rth cell to the satellite.
[0036] The embodiment of the present invention considers the Ricean channel of the satellite. The Ricean channel has a direct path and multiple non-line-of-sight paths. Therefore, the channel model when the k-th ground user sends the p-th pilot at the communication frequency f is expressed as: Among them, κ i represents the Ricean factor of the ith path, ξ k,i represents the channel gain of the i-th path of the k-th ground user, 1≤i≤L k , represents the line-of-sight path channel response of the satellite, represents the non-line-of-sight path channel response of the satellite, represents a complex vector of length M. v k,i represents the Doppler shift of the i-th path of the k-th terrestrial user, τ k,i represents the propagation delay of the i-th path of the k-th terrestrial user, a s represents the i-th path of the k-th user and the satellite receiving antenna steering vector, represents the azimuth angle from the i-th path of the k-th user to the satellite receiving antenna, It represents the elevation angle from the i-th path of the k-th user to the satellite receiving antenna.
[0037] Step S104, using the target neural network model to process the predefined pilot matrix and pilot sequence to obtain the equivalent channel matrix of all ground users.
[0038] After obtaining the predefined pilot matrix and pilot sequence, the embodiment of the present invention uses the target neural network model trained with a large number of samples to process the predefined pilot matrix and pilot sequence to obtain the equivalent channel matrix of all ground users. The target neural network model includes multiple layers of neural network units, each layer of neural network units includes: a model-driven network module and a data-driven network module, the input of the model-driven network module includes: the input of the target neural network model and the output of the data-driven network module in the previous layer of neural network units, and the output of the model-driven network module is the input of the data-driven network module of this layer.
[0039] Figure 3 A schematic diagram of the structure of a target neural network model provided by an embodiment of the present invention is shown in FIG. Figure 3As shown, the target neural network model includes multiple layers of neural network units connected in sequence, each layer of neural network units is provided with a model-driven network module and a data-driven network module, and the input data of each layer of neural network units is composed of the input data of the neural network model and the output data of the neural network units in the previous layer. Since the first layer of neural network units does not have a previous layer of neural network units, its input data is the input data of the neural network model. The input data of each layer of neural network units needs to be processed by the model-driven network module first, and then processed again by the data-driven network module. The output of each layer of neural network units is the output of the data-driven network module of this layer.
[0040] The target neural network model can select MMV-DOAMPNet (Multiple measurement vectorsmodel and data dual driven orthogonal approximate message passing network). Each layer of neural network unit is regarded as an iterative network. The network parameters of the model-driven network module in the iterative network are used to control the convergence speed and learn the user access sparsity, and then estimate the sparse channels in the beam space. The data-driven network module embedded in the network is used to further extract the characteristics of the user dimension and beam space based on each model-driven network output. The data-driven network can adopt the common DNN, CNN or Transformer in the prior art. Therefore, the equivalent channel matrix output by the target neural network model can have a higher accuracy.
[0041] Step S106, based on the equivalent channel matrix, the number of beams and the preset active decision threshold, the beam space channel matrix and the active matrix of all ground users are determined, and the beam space channel matrix and the active matrix are used as the joint channel estimation and detection results of the multi-beam satellite unlicensed access.
[0042] In an embodiment of the present invention, a preset active decision threshold represents the lower limit of the activity probability of ground users. After obtaining the equivalent channel matrix of all ground users, the active state of each ground user can be determined by combining the number of beams generated by the multi-beam satellite and the preset active decision threshold. If the kth user is in an active state, the element of the kth row in the equivalent channel matrix is used as the beam space channel estimation value of the user. If the kth user is in an inactive state, the beam space channel estimation value of the user is 0. The activity indicators of the active state of all ground users (active takes 1, inactive takes 0) constitute the active matrix, and the beam space channel estimation values of all ground users constitute the beam space channel matrix.
[0043] The joint channel estimation and detection method for unlicensed access of multi-beam satellites provided in an embodiment of the present invention uses a target neural network model to process a predefined pilot matrix and a pilot sequence received by a multi-beam satellite to obtain an equivalent channel matrix for all ground users, and then determines the joint channel estimation and detection result for unlicensed access of multi-beam satellites based on the equivalent channel matrix, the number of beams generated by the multi-beam satellite, and a preset active decision threshold. Based on the structure of the target neural network model, it can be seen that each neural network unit in the model further extracts features using a data-driven network based on the output of each model-driven network. Therefore, the target neural network model can realize the complementary advantages of model-driven and data-driven, which not only reduces the model's dependence on prior information, but also achieves better detection performance, thereby alleviating the technical problem of low accuracy of detection results in the existing joint channel estimation and detection method for unlicensed access of multi-beam satellites.
[0044] In an optional embodiment, the method of the present invention further comprises the steps of:
[0045] Step S201, obtaining multiple groups of training samples; wherein each group of training samples includes: a predefined pilot matrix, pilot sequence samples received by a multi-beam satellite, beam space channel matrix samples and active matrix samples of all ground users; the beam space channel matrix samples and active matrix samples of all ground users are sample labels.
[0046] Step S202, inputting multiple groups of training samples into the initial neural network model to perform iterative training on the initial neural network model module by module and layer by layer to obtain a target neural network model.
[0047] Among them, in the module-by-module and layer-by-layer iterative training process, for the nth layer neural network unit in the initial neural network model, the network parameters of the model-driven network module are first updated based on the output results of its model-driven network module, and then the network parameters of the nth layer neural network unit are updated based on the output results of its data-driven network module, and finally the network parameters of the 1st to nth layer neural network units are updated based on the output results of the nth layer neural network unit after the network parameters are updated.
[0048] Specifically, in order to avoid the disappearance or explosion of training gradients, ensure that each module of each layer of neural network units can be fully trained, and avoid falling into a local optimal solution, the embodiment of the present invention proposes a method for training the initial neural network model in a module-by-module-by-layer manner. tr Group training samples, where the i-th group training samples can be expressed as: The subscript tr represents the training sample. represents the pilot sequence samples received by the multi-beam satellite, S represents the predefined pilot matrix, represents the active matrix samples of all ground users, Represents the beamspace channel matrix samples for all terrestrial users.
[0049] The model parameters of the nth layer of neural network units in the initial neural network model are expressed as: Ω n ={Ω m,n ,Ω d,n}; where Ω m,n Represents the network parameters of the model-driven network module in the nth layer of neural network units, Ω d,n Represents the network parameters of the data-driven network module in the nth layer of neural network units. Before starting training, the principles of iterative algorithms and common neural network initialization methods (such as Xavier initialization) are used to set all Ω in the initial application network model. m,n and Ω d,n The initial value of .
[0050] After starting to train the initial neural network model, the training samples need to be input into the initial neural network model in sequence. For each group of training samples, each layer of the neural network model can output an estimated equivalent channel matrix. Based on the description above, it can be seen that the embodiment of the present invention adopts a module-by-module, layer-by-layer iterative training method, that is, first use multiple groups of samples to train the model-driven network module in the first layer of neural network units, and then use the samples to simultaneously train the model-driven network module and data-driven network module of the first layer, and finally input the samples into the model, and use the output results of the first layer of neural network units to train the network parameters of this layer again. Next, start training the model-driven network module in the second layer of neural network units, and then train the model-driven network module and data-driven network module of the second layer at the same time, and finally use the output results of the second layer of neural network units to update the network parameters of the first and second layers of neural network units, and so on. After completing the network parameter update of all neural network units in the network model, the target neural network model is obtained.
[0051] In an optional implementation, updating the network parameters of the model-driven network module based on the output result of the model-driven network module specifically includes the following steps:
[0052] Step S301, after obtaining the target training sample and inputting it into the initial neural network model, the output result of the target model driving network module; wherein the target training sample represents any sample in multiple groups of training samples; the target model driving network module represents the model driving network module in the nth layer neural network unit.
[0053] Step S302: Calculate an equivalent channel matrix sample corresponding to the target training sample based on the beam space channel matrix sample and the active matrix sample in the target training sample.
[0054] Step S303, calculating a first loss function value based on the output results of the target model driving network module corresponding to all training samples and the equivalent channel matrix samples.
[0055] Step S304: Update the network parameters of the target model-driven network module based on the first loss function value.
[0056] Specifically, if f m,n represents the model-driven network module in the nth layer of neural network unit, then the target training sample After inputting the initial neural network model, the output result of the target model driving network module is expressed as: in, Represents the output result of the n-1th layer neural network unit, Represents the output result of the model-driven network module in the nth layer of neural network unit. and Then according to the formula The equivalent channel matrix sample corresponding to the target training sample can be calculated
[0057] In order to optimize the model drive, the embodiment of the present invention defines an equivalent channel matrix sample And the output result of the target model driving network module in the nth layer neural network unit The NMSE (Normalized mean square error) between the estimated equivalent channel matrix is used as the loss function of the n-th layer model-driven network module, recorded as the first loss function, and expressed as:
[0058]
[0059] Among them, J m,n (Ω m,n ) represents the first loss function, and the Ω in the brackets m,n represents the parameters that need to be optimized for the loss function, ||·|| F The F-norm of the matrix is represented, and the network parameters of the target model driving network module can be updated through the first loss function value.
[0060] In an optional implementation, updating the network parameters of the n-th layer neural network unit based on the output result of its data-driven network module specifically includes the following steps:
[0061] Step S401, input the target training sample into the neural network model after the network parameters of the target model-driven network module are updated, and obtain the output result of the target data-driven network module; wherein the target data-driven network module represents the data-driven network module in the nth layer of the neural network unit.
[0062] Step S402, calculating a second loss function value based on the output results of the target data-driven network module corresponding to all training samples and the equivalent channel matrix samples.
[0063] Step S403, updating the network parameters of the n-th layer neural network unit based on the second loss function value.
[0064] After updating the model-driven network module in the nth layer of neural network unit, its network parameters are expressed as: Ω′ m,n If f d,n Represents the data-driven network module in the nth layer of neural network units, and converts the target training samples After inputting the current neural network model, the data-driven network module further learns the sparse characteristics and the channel state responses of the user dimension and the beam dimension in the output result of the model-driven network module. In the embodiment of the present invention, the output result of the target data-driven network module is expressed as: Represents the output result of the data-driven network module in the n-th layer of neural network unit.
[0065] Since the data-driven network module is embedded in each iteration, in order to ensure that the parameters corresponding to the model-driven network module and the data-driven network module in the nth layer can be fully trained, the embodiment of the present invention defines an equivalent channel matrix sample and the equivalent channel matrix estimated by the n-th layer data-driven network module The NMSE between is used as the loss function of the n-th layer data-driven network, recorded as the second loss function, and the second loss function is expressed as:
[0066]
[0067] Among them, J d,n (Ω m,n ,Ω d,n ) represents the second loss function, and the Ω in the brackets m,n ,Ω d,n Representing the parameters that need to be optimized by the loss function, the network parameters of the n-th layer neural network unit can be updated through the second loss function value, that is, the network parameters of the model-driven network module and the data-driven network module in the n-th layer neural network unit are updated.
[0068] In an optional implementation, updating the network parameters of the neural network units of the first to n layers based on the output result of the neural network unit of the nth layer after the network parameters are updated specifically includes the following steps:
[0069] Step S501, input the target training sample into the neural network model after the network parameters of the n-th layer neural network unit are updated, and obtain the output result of the n-th layer neural network unit.
[0070] Step S502, calculate the third loss function value based on the output results of the nth layer neural network units corresponding to all training samples and the equivalent channel matrix samples.
[0071] Step S503, updating the network parameters of the neural network units from the 1st to the nth layers based on the third loss function value.
[0072] In order to prevent the network parameters of each layer of neural network units from falling into the local optimal solution, the embodiment of the present invention trains the neural network units from the 1st to the nth layer together once after the first training of the neural network unit in the nth layer (the model-driven network module is trained twice and the data-driven network module is trained once). After the first training of the neural network unit in the nth layer, its updated network parameters are expressed as: (Ω″ m,n ,Ω′ d,n ). Therefore, after the target training sample is input into the neural network model after the network parameters of the n-th layer neural network unit are updated, the output result of the n-th layer neural network unit, that is, the output result of its data-driven network module can be expressed as:
[0073] According to the definition of the third loss function above, the third loss function can be expressed as:
[0074]
[0075] Among them, J n (Ω1,Ω2…Ω n ) represents the third loss function, and the Ω1, Ω2…Ω in the brackets n The network parameters of the first n layers of neural network units that need to be optimized by the loss function can be updated through the third loss function value. After completing the network parameter update of all N layers of neural network units in the model, the target neural network model is obtained, and its model parameters can be expressed as: That is, Figure 3 The network model shown in .
[0076] In an optional implementation, the dimension of the equivalent channel matrix is K×N RF , K represents the total number of ground users of multi-beam satellites, NRF Indicates the number of beams; the above step S106, based on the equivalent channel matrix, the number of beams and the preset active decision threshold, determines the beam space channel matrix and the active matrix of all ground users, specifically comprising the following steps:
[0077] Step S1061 : determining a channel gain threshold based on the maximum value of the elements in the equivalent channel matrix, and the minimum and maximum values of the absolute values of the non-zero elements.
[0078] Specifically, the channel gain threshold is express, represents the equivalent channel matrix The element in row k and column r in represents the maximum value of the elements in the equivalent channel matrix, and υ represents the ratio of the minimum and maximum absolute values of the non-zero elements in the equivalent channel matrix.
[0079] Step S1062, determining the activity indicator of the kth ground user of the multi-beam satellite based on the elements of the kth row in the equivalent channel matrix, the channel gain threshold, the number of beams and the preset active decision threshold.
[0080] In an optional implementation, determining the activity indicator of the kth ground user of the multi-beam satellite based on the elements of the kth row in the equivalent channel matrix, the channel gain threshold, the number of beams, and the preset active decision threshold specifically includes the following contents:
[0081] Using the formula Determine an activity indicator for the kth ground user of a multi-beam satellite; wherein, represents the activity indicator of the kth terrestrial user, represents the element in the kth row and rth column of the equivalent channel matrix, represents the channel gain threshold, υ represents the ratio of the minimum and maximum absolute values of non-zero elements in the equivalent channel matrix, λ th Indicates the preset active decision threshold.
[0082] Step S1063: If the value of the activity indicator of the kth terrestrial user is 1, it is determined that the kth terrestrial user is in an active state, and the element of the kth row in the equivalent channel matrix is used as the beam space channel estimation value of the kth terrestrial user. That is,
[0083] Step S1064: if the value of the activity indicator of the kth terrestrial user is 0, it is determined that the kth terrestrial user is in an inactive state, and the beam space channel estimation value of the kth terrestrial user is 0.
[0084] Step S1065: construct an activity matrix based on the activity indicators of all terrestrial users, and construct a beam space channel matrix based on the beam space channel estimation values of all terrestrial users.
[0085] In summary, the embodiment of the present invention proposes a multi-beam satellite unlicensed access architecture, which makes full use of existing user communication resources and supports large-scale user access. The embodiment of the present invention also proposes a model-data dual-driven joint channel estimation and detection method suitable for this architecture, and designs a corresponding module-by-module-by-layer training method to fully train parameters. The sparse recovery performance of the model-driven network and the adaptive learning characteristics of the data-driven network are utilized to improve detection accuracy and reduce access overhead. The target neural network model can effectively reduce channel estimation errors and inaccurate channel or user prior information, improve the accuracy of data recovery, reduce user pilot overhead, and achieve fast and seamless access for large-scale users in multi-beam satellite scenarios.
[0086] Embodiment 2
[0087] An embodiment of the present invention also provides a joint channel estimation and detection device for unlicensed access of multi-beam satellites. The device is mainly used to execute the joint channel estimation and detection method for unlicensed access of multi-beam satellites provided in the above-mentioned embodiment one. The following is a specific introduction to the joint channel estimation and detection device for unlicensed access of multi-beam satellites provided in an embodiment of the present invention.
[0088] Figure 4 A functional module diagram of a joint channel estimation and detection device for multi-beam satellite unlicensed access provided by an embodiment of the present invention, such as Figure 4 As shown, the device mainly includes: a first acquisition module 10, a processing module 20, and a determination module 30, wherein:
[0089] The first acquisition module 10 is used to acquire a predefined pilot matrix, a pilot sequence received by the multi-beam satellite and the number of beams generated by the multi-beam satellite; wherein the predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite.
[0090] The processing module 20 is used to process the predefined pilot matrix and pilot sequence using the target neural network model to obtain the equivalent channel matrix of all ground users; wherein the target neural network model includes multiple layers of neural network units, each layer of neural network units includes: a model-driven network module and a data-driven network module, the input of the model-driven network module includes: the input of the target neural network model and the output of the data-driven network module in the previous layer of neural network units, and the output of the model-driven network module is the input of the data-driven network module of this layer.
[0091] The determination module 30 is used to determine the beam space channel matrix and the active matrix of all ground users based on the equivalent channel matrix, the number of beams and the preset active decision threshold, and use the beam space channel matrix and the active matrix as the joint channel estimation and detection results of the multi-beam satellite unlicensed access.
[0092] The joint channel estimation and detection device for unlicensed access of multi-beam satellites provided in an embodiment of the present invention uses a target neural network model to process a predefined pilot matrix and a pilot sequence received by a multi-beam satellite to obtain an equivalent channel matrix for all ground users, and then determines the joint channel estimation and detection result for unlicensed access of multi-beam satellites based on the equivalent channel matrix, the number of beams generated by the multi-beam satellite, and a preset active decision threshold. Based on the structure of the target neural network model, it can be seen that each neural network unit in the model further extracts features using a data-driven network based on the output of each model-driven network. Therefore, the target neural network model can realize the complementary advantages of model-driven and data-driven, which not only reduces the model's dependence on prior information, but also achieves better detection performance, thereby alleviating the technical problem of low accuracy of detection results in the existing joint channel estimation and detection method for unlicensed access of multi-beam satellites.
[0093] Optionally, the device further comprises:
[0094] The second acquisition module is used to obtain multiple groups of training samples; each group of training samples includes: a predefined pilot matrix, pilot sequence samples received by the multi-beam satellite, beam space channel matrix samples and active matrix samples of all ground users; the beam space channel matrix samples and active matrix samples of all ground users are sample labels.
[0095] The training module is used to input multiple groups of training samples into the initial neural network model to iteratively train the initial neural network model module by module and layer by layer to obtain the target neural network model.
[0096] Among them, in the module-by-module and layer-by-layer iterative training process, for the nth layer neural network unit in the initial neural network model, the network parameters of the model-driven network module are first updated based on the output results of its model-driven network module, and then the network parameters of the nth layer neural network unit are updated based on the output results of its data-driven network module, and finally the network parameters of the 1st to nth layer neural network units are updated based on the output results of the nth layer neural network unit after the network parameters are updated.
[0097] Optionally, the training module includes:
[0098] The first acquisition unit is used to obtain the output result of the target model driving network module after the target training sample is input into the initial neural network model; wherein the target training sample represents any sample in multiple groups of training samples; the target model driving network module represents the model driving network module in the nth layer neural network unit.
[0099] The first calculation unit is used to calculate an equivalent channel matrix sample corresponding to the target training sample based on the beam space channel matrix sample and the active matrix sample in the target training sample.
[0100] The second calculation unit is used to calculate the first loss function value based on the output results of the target model driving network module corresponding to all training samples and the equivalent channel matrix samples.
[0101] The first updating unit is used to update the network parameters of the target model-driven network module based on the first loss function value.
[0102] Optionally, the training module also includes:
[0103] The second acquisition unit is used to input the target training sample into the neural network model after the network parameters of the target model-driven network module are updated, and obtain the output result of the target data-driven network module; wherein the target data-driven network module represents the data-driven network module in the nth layer of the neural network unit.
[0104] The third calculation unit is used to calculate the second loss function value based on the output results of the target data-driven network module corresponding to all training samples and the equivalent channel matrix samples.
[0105] The second updating unit is used to update the network parameters of the n-th layer neural network unit based on the second loss function value.
[0106] Optionally, the training module also includes:
[0107] The third acquisition unit is used to input the target training sample into the neural network model after the network parameters of the n-th layer neural network unit are updated, and obtain the output result of the n-th layer neural network unit.
[0108] The fourth calculation unit is used to calculate the third loss function value based on the output results of the nth layer neural network unit corresponding to all training samples and the equivalent channel matrix samples.
[0109] The third updating unit is used to update the network parameters of the 1st to nth layer neural network units based on the third loss function value.
[0110] Optionally, the dimension of the equivalent channel matrix is K×N RF , K represents the total number of ground users of multi-beam satellites, N RFIndicates the number of beams; the determination module 30 includes:
[0111] The first determining unit is used to determine the channel gain threshold based on the maximum value of the elements in the equivalent channel matrix, the minimum value and the maximum value of the absolute values of the non-zero elements.
[0112] The second determining unit is used to determine the activity indicator of the kth ground user of the multi-beam satellite based on the elements of the kth row in the equivalent channel matrix, the channel gain threshold, the number of beams and the preset active decision threshold.
[0113] The third determination unit is used to determine that the kth terrestrial user is in an active state when the value of the activity indicator of the kth terrestrial user is 1, and use the elements of the kth row in the equivalent channel matrix as the beam space channel estimation value of the kth terrestrial user.
[0114] The fourth determination unit is used to determine that the kth terrestrial user is in an inactive state and the beam space channel estimation value of the kth terrestrial user is 0 when the value of the activity indicator of the kth terrestrial user is 0.
[0115] The construction unit is used to construct an activity matrix based on the activity indicators of all terrestrial users, and to construct a beam space channel matrix based on the beam space channel estimation values of all terrestrial users.
[0116] Optionally, the second determining unit is specifically configured to:
[0117] Using the formula Determine an activity indicator for the kth ground user of a multi-beam satellite; wherein, represents the activity indicator of the kth terrestrial user, represents the element in the kth row and rth column of the equivalent channel matrix, represents the channel gain threshold, υ represents the ratio of the minimum and maximum absolute values of non-zero elements in the equivalent channel matrix, λ th Indicates the preset active decision threshold.
[0118] Embodiment 3
[0119] See also Figure 5 An embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62 and a communication interface 63, the processor 60, the communication interface 63 and the memory 61 are connected through the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.
[0120] The memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 63 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.
[0121] The bus 62 may be an ISA bus, a PCI bus, or an EISA bus. The bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0122] Among them, the memory 61 is used to store programs, and the processor 60 executes the program after receiving the execution instruction. The method executed by the device defined by the process disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 60 or implemented by the processor 60.
[0123] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 60. The above processor 60 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present invention can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and completes the steps of the above method in combination with its hardware.
[0124] A computer program product of a method and apparatus for joint channel estimation and detection of multi-beam satellite unlicensed access provided in an embodiment of the present invention includes a computer-readable storage medium storing a non-volatile program code executable by a processor. The instructions included in the program code can be used to execute the method in the previous method embodiment. For specific implementation, please refer to the method embodiment, which will not be repeated here.
[0125] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0126] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0127] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0128] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0129] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.
[0130] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A joint channel estimation and detection method for multi-beam satellite unlicensed access, characterized in that: include: Acquire a predefined pilot matrix, a pilot sequence received by a multi-beam satellite, and the number of beams generated by the multi-beam satellite; wherein the predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite; The predefined pilot matrix and the pilot sequence are processed by using a target neural network model to obtain an equivalent channel matrix of all ground users; wherein the target neural network model includes multiple layers of neural network units, each layer of neural network units includes: a model-driven network module and a data-driven network module, the input of the model-driven network module includes: the input of the target neural network model and the output of the data-driven network module in the neural network unit of the previous layer, and the output of the model-driven network module is the input of the data-driven network module of this layer; Based on the equivalent channel matrix, the number of beams and a preset active decision threshold, the beam space channel matrix and the active matrix of all ground users are determined, and the beam space channel matrix and the active matrix are used as the joint channel estimation and detection results of the multi-beam satellite unlicensed access.
2. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 1, characterized in that: Also includes: Acquire multiple groups of training samples; wherein each group of the training samples includes: the predefined pilot matrix, the pilot sequence samples received by the multi-beam satellite, the beam space channel matrix samples and active matrix samples of all ground users; the beam space channel matrix samples and active matrix samples of all ground users are sample labels; Inputting the multiple groups of training samples into an initial neural network model to perform iterative training on the initial neural network model module by module and layer by layer to obtain the target neural network model; Among them, in the module-by-module and layer-by-layer iterative training process, for the nth layer neural network unit in the initial neural network model, the network parameters of the model-driven network module are first updated based on the output result of its model-driven network module, and then the network parameters of the nth layer neural network unit are updated based on the output result of its data-driven network module, and finally the network parameters of the 1st to nth layer neural network units are updated based on the output result of the nth layer neural network unit after the network parameters are updated.
3. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 2, characterized in that: The network parameters of the model-driven network module are updated based on the output results of the model-driven network module, including: After obtaining the target training sample and inputting it into the initial neural network model, the output result of the target model driving network module; wherein the target training sample represents any sample in the plurality of groups of training samples; the target model driving network module represents the model driving network module in the nth layer neural network unit; Calculating an equivalent channel matrix sample corresponding to the target training sample based on the beam space channel matrix sample and the active matrix sample in the target training sample; Calculate a first loss function value based on the output results of the target model driving network module corresponding to all training samples and the equivalent channel matrix samples; The network parameters of the target model-driven network module are updated based on the first loss function value.
4. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 3, characterized in that: The network parameters of the nth layer neural network unit are updated based on the output results of its data-driven network module, including: Input the target training sample into the neural network model after the network parameters of the target model-driven network module are updated, and obtain the output result of the target data-driven network module; wherein the target data-driven network module represents the data-driven network module in the n-th layer neural network unit; Calculate the second loss function value based on the output results of the target data-driven network module corresponding to all training samples and the equivalent channel matrix samples; The network parameters of the nth layer neural network unit are updated based on the second loss function value.
5. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 4, characterized in that: The network parameters of the neural network units from the first to the nth layers are updated based on the output results of the neural network units in the nth layer after the network parameters are updated, including: Input the target training sample into the neural network model after the network parameters of the n-th layer neural network unit are updated, and obtain the output result of the n-th layer neural network unit; Calculate the third loss function value based on the output results of the nth layer neural network units corresponding to all training samples and the equivalent channel matrix samples; The network parameters of the 1st to nth layer neural network units are updated based on the third loss function value.
6. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 1, characterized in that: The dimension of the equivalent channel matrix is K×N RF , K represents the total number of ground users of the multi-beam satellite, N RF represents the number of beams; Determining beam space channel matrices and activity matrices of all ground users based on the equivalent channel matrix, the number of beams, and a preset active decision threshold includes: Determine a channel gain threshold based on a maximum value of elements in the equivalent channel matrix, a minimum value and a maximum value of absolute values of non-zero elements; Determining an activity indicator of a kth ground user of the multi-beam satellite based on the elements of the kth row in the equivalent channel matrix, the channel gain threshold, the number of beams, and the preset active decision threshold; If the value of the activity indicator of the kth terrestrial user is 1, it is determined that the kth terrestrial user is in an active state, and the element of the kth row in the equivalent channel matrix is used as the beam space channel estimation value of the kth terrestrial user; If the value of the activity indicator of the kth terrestrial user is 0, it is determined that the kth terrestrial user is in an inactive state, and the beam space channel estimation value of the kth terrestrial user is 0; The activity matrix is constructed based on the activity indicators of all terrestrial users, and the beam space channel matrix is constructed based on the beam space channel estimation values of all terrestrial users.
7. The joint channel estimation and detection method for multi-beam satellite unlicensed access according to claim 6, characterized in that: Determining an activity indicator of a k-th ground user of the multi-beam satellite based on elements of the k-th row in the equivalent channel matrix, the channel gain threshold, the number of beams, and a preset active decision threshold, comprising: Using the formula Determining an activity indicator of a k-th ground user of the multi-beam satellite; wherein, represents the activity indicator of the kth terrestrial user, represents the element in the kth row and the rth column in the equivalent channel matrix, represents the channel gain threshold, υ represents the ratio of the minimum and maximum absolute values of the non-zero elements in the equivalent channel matrix, λ th represents the preset active decision threshold.
8. A joint channel estimation and detection device for multi-beam satellite unlicensed access, characterized in that: A first acquisition module is used to acquire a predefined pilot matrix, a pilot sequence received by a multi-beam satellite, and the number of beams generated by the multi-beam satellite; wherein the predefined pilot matrix represents a matrix composed of pilot vectors of all ground users of the multi-beam satellite; A processing module, used to process the predefined pilot matrix and the pilot sequence using a target neural network model to obtain an equivalent channel matrix of all ground users; wherein the target neural network model includes multiple layers of neural network units, each layer of neural network units includes: a model-driven network module and a data-driven network module, the input of the model-driven network module includes: the input of the target neural network model and the output of the data-driven network module in the neural network unit of the previous layer, and the output of the model-driven network module is the input of the data-driven network module of this layer; A determination module is used to determine the beam space channel matrix and the active matrix of all ground users based on the equivalent channel matrix, the number of beams and a preset active decision threshold, and use the beam space channel matrix and the active matrix as the joint channel estimation and detection result of the multi-beam satellite unlicensed access.
9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein: When the processor executes the computer program, the steps of the joint channel estimation and detection method for multi-beam satellite unlicensed access according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the joint channel estimation and detection method for unlicensed access of a multi-beam satellite according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Hybrid model-driven and data-driven downlink channel estimation method and system
CN117834349A
Activated user detection method based on intelligent auxiliary multi-satellite combination
CN117880015A
Nonlinear MIMO channel estimation method based on Inception-ResNet network and dense residual network
CN118118301A
Enhancement of channel estimation in wireless communication based on supervised learning
US20220337455A1
Cited By
Channel estimation method and device for asynchronous access of large-scale users
CN120223475A
A channel estimation method and device for large-scale user asynchronous access
CN120223475B
Satellite auxiliary pilot frequency distribution method based on quantum convolutional neural network
CN120263373A
