Satellite-assisted pilot allocation method based on quantum convolutional neural network

Abstract

The application relates to the technical field of communication, in particular to a satellite-assisted pilot allocation method based on a quantum convolutional neural network, wherein the method comprises the following steps: determining all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user; determining a maximum value of the coefficients in all ground large-scale fading coefficients, and determining a target communication network architecture according to the maximum value; serving all single-antenna users by using the target communication network architecture, and inputting all large-scale fading coefficients corresponding to the target communication network architecture into a preset quantum convolutional neural network model to obtain pilot indexes of each single-antenna user, so as to perform pilot allocation according to the pilot indexes. Therefore, the pilot pollution problem caused by pilot reuse is solved, the pilot allocation can be optimized, and the pilot pollution is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a satellite-assisted pilot allocation method based on a quantum convolutional neural network. BACKGROUND

[0002] The sixth generation mobile communication system is conceived to meet the stringent requirements of Internet of Things applications such as intelligent transportation and industrial automation, aiming to achieve an air interface delay of less than 0.1 ms, a reliability of more than 99.9999%, a connection number density of more than 107 / km2, and global connectivity. In order to meet the stringent requirements of 6G for low delay, high reliability, high energy efficiency, large-scale connection and scalability, cell-free massive multiple-input and multiple-output (CF mMIMO) technology has become a research hotspot.

[0003] With the growing demand for global connectivity, low earth orbit (LEO) satellite networks have become a key technology for achieving global connectivity due to their low latency and high data transmission rates. In cases where ground communication links are poor and cannot support effective high-reliability low-latency communication, devices are provided with the necessary communication link connection. At the same time, with the intensification of space resource competition, satellites are deployed more and more densely, and it is becoming a reality that dozens of satellites will appear simultaneously above the user, therefore, inter-machine communication can use multiple satellites to provide services for users at the same time, improving communication quality.

[0004] However, inter-machine communication usually uses short packet communication, and with the access of a large number of devices, the number of orthogonal pilots that can be allocated under a limited block length is very limited, and when the pilot length is less than the number of users, pilot reuse will cause serious pilot pollution.

[0005] Related technologies formulate the pilot allocation problem as a convex optimization problem and try different objective functions, or formulate the pilot allocation problem as a graph partitioning problem and introduce different heuristic algorithms such as graph coloring algorithms and Hungarian algorithms, however, with the increase in the number of users and the increase in channel dimension, the complexity of traditional algorithms will increase sharply, in addition, these traditional algorithms usually rely on explicit mathematical models or rules, and have low generalization ability, making it difficult to cope with changing demands in practical applications; in addition, related technologies also propose a pilot allocation algorithm based on artificial intelligence, such as an algorithm based on K-means, an algorithm based on deep reinforcement learning, and an algorithm based on convolutional neural networks, however, when the number of access UEs is large, these algorithms are still limited by their explosive complexity. SUMMARY

[0006] The application provides a satellite-assisted pilot allocation method based on a quantum convolutional neural network to solve the pilot pollution problem caused by pilot reuse in harsh conditions, optimize pilot allocation, reduce pilot pollution, and improve the reliability of the system under ultra-low delay conditions.

[0007] The first aspect of the application provides a satellite-assisted pilot allocation method based on a quantum convolutional neural network, which is applied to a preset communication network architecture, wherein the preset communication network architecture includes a preset ground communication network architecture and a preset satellite communication network architecture, the preset ground communication network architecture includes at least one ground access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point, wherein the method comprises the following steps:

[0008] Determine all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and the single-antenna user;

[0009] Determine the maximum value of the coefficients in the all ground large-scale fading coefficients, and determine the target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value of the coefficients;

[0010] Use the target communication network architecture to serve all single-antenna users, and input all large-scale fading coefficients corresponding to the target communication network architecture into a preset quantum convolutional neural network model to obtain pilot indexes of the single-antenna users, so as to allocate pilots to each single-antenna user according to the pilot indexes of the single-antenna users.

[0011] Optionally, in some embodiments, the target communication network architecture is determined from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value of the coefficients, comprising:

[0012] Determine whether the maximum value of the coefficients is greater than a preset threshold;

[0013] If the maximum value of the coefficients is greater than the preset threshold, the preset ground communication network architecture is used as the target communication network architecture, otherwise, the preset satellite communication network architecture is used as the target communication network architecture.

[0014] Optionally, in some embodiments, before inputting all large-scale fading coefficients corresponding to the target communication network architecture into the preset quantum convolutional neural network model, comprising:

[0015] Obtain training set data;

[0016] inputting the training set data into a preset quantum circuit, wherein the training set data is encoded by using an RX gate in the preset quantum circuit to obtain encoded data;

[0017] performing quantum operation on the encoded data to obtain a quantum state, and performing quantum entanglement on the quantum state by using a CNOT gate to obtain a quantum-entangled quantum state;

[0018] observing the quantum-entangled quantum state by using a Pauli-Z to obtain a quantum computing result, inputting the quantum computing result into an initial neural network model to obtain a trained neural network model, calculating a cross-entropy loss function of the trained neural network model to obtain a loss value, and optimizing the trained neural network model by using the loss value to obtain a new trained neural network model;

[0019] taking the new trained neural network model as the initial neural network model, and re-executing the step of inputting the training set data into the preset quantum circuit until the cumulative training round number reaches a preset training round number, and outputting a final trained neural network model as the preset quantum convolutional neural network model.

[0020] Optionally, in some embodiments, the encoding of the training set data by using the RX gate to obtain the encoded data comprises:

[0021] encoding the training set data by using the RX gate according to a preset encoding formula to obtain the encoded data, wherein the preset encoding formula is:

[0022] ;

[0023] wherein, is a rotation gate around the X-axis, is an angle of rotation of a quantum state around the X-axis, is an imaginary unit.

[0024] Optionally, in some embodiments, the calculation formula of all terrestrial large-scale fading coefficients between each ground access point and each single-antenna user is:

[0025] ;

[0026] the calculation formula of all satellite large-scale fading coefficients between each satellite access point and each single-antenna user is:

[0027] ;

[0028] wherein, is the i-th The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first The standard deviation among individual antenna users is The shadow of decline, For the first The access point to the first Standard deviation between individual antenna users For the first The access point to the first A standard normally distributed random variable among single-antenna users For the first The access point to the first Large-scale satellite fading coefficients between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first The standard deviation among individual antenna users is The shadow of decline, For the first The access point to the first Standard deviation between individual antenna users For the first The access point to the first A standard normally distributed random variable among single-antenna users.

[0029] A second aspect of this application provides a satellite-assisted pilot allocation device based on a quantum convolutional neural network. The device is applied to a preset communication network architecture, wherein the preset communication network architecture includes a preset terrestrial communication network architecture and a preset satellite communication network architecture. The preset terrestrial communication network architecture includes at least one ground access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point. The device includes:

[0030] The first determining module is used to determine all ground-scale fading coefficients between each ground access point and each single-antenna user, and all satellite-scale fading coefficients between each satellite access point and each single-antenna user;

[0031] determining a maximum value of the coefficients in the all ground large-scale fading coefficients, and determining a target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value of the coefficients;

[0032] a pilot allocation module configured to serve all single-antenna users by using the target communication network architecture, and input all large-scale fading coefficients corresponding to the target communication network architecture into a preset quantum convolutional neural network model to obtain a pilot index of each single-antenna user, so as to allocate a pilot to each single-antenna user according to the pilot index of each single-antenna user.

[0033] Optionally, in some embodiments, the second determining module comprises:

[0034] a judging unit configured to judge whether the maximum value of the coefficients is greater than a preset threshold value;

[0035] a generating unit configured to, when the maximum value of the coefficients is greater than the preset threshold value, take the preset ground communication network architecture as the target communication network architecture, or otherwise take the preset satellite communication network architecture as the target communication network architecture.

[0036] Optionally, in some embodiments, before inputting all large-scale fading coefficients corresponding to the target communication network architecture into the preset quantum convolutional neural network model, the pilot allocation module comprises:

[0037] an obtaining unit configured to obtain training set data;

[0038] an encoding unit configured to input the training set data into a preset quantum circuit, and encode the training set data by using an RX gate in the preset quantum circuit to obtain encoded data;

[0039] an operation unit configured to perform quantum operation on the encoded data to obtain a quantum state, and perform quantum entanglement on the quantum state by using a CNOT gate to obtain a quantum-entangled quantum state;

[0040] a training unit configured to obtain a quantum calculation result by using a Pauli-Z to observe the quantum-entangled quantum state, input the quantum calculation result into an initial neural network model to perform training to obtain a trained neural network model, calculate a cross-entropy loss function of the trained neural network model to obtain a loss value, optimize the trained neural network model by using the loss value, and obtain a new trained neural network model;

[0041] The generation unit is used to take the new trained neural network model as the initial neural network model and re-execute the step of inputting the training set data into the preset quantum circuit until the cumulative number of training rounds reaches the preset number of training rounds, and output the final trained neural network model as the preset quantum convolutional neural network model.

[0042] Optionally, in some embodiments, the encoding unit includes:

[0043] An encoding subunit is used to encode the training set data using an RX gate according to a preset encoding formula to obtain encoded data, wherein the preset encoding formula is:

[0044] ;

[0045] in, It is a revolving door around the X-axis. The angle of rotation of the quantum state around the X-axis. It is the imaginary unit.

[0046] Optionally, in some embodiments, the formula for calculating the large-scale fading coefficient of all ground-based connections between each ground access point and each single-antenna user is as follows:

[0047] ;

[0048] The formula for calculating the large-scale fading coefficient of all satellites between each satellite access point and each single-antenna user is as follows:

[0049] ;

[0050] in, For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first The standard deviation among individual antenna users is The shadow of decline, For the first The access point to the first Standard deviation between individual antenna users For the first The access point to the first A standard normally distributed random variable among single-antenna users For the first The access point to the first a satellite large-scale fading coefficient between the i-th ground access point and the j-th single-antenna user, a path loss between the i-th ground access point and the j-th single-antenna user, a shadow fading with a standard deviation of between the i-th ground access point and the j-th single-antenna user, a standard deviation between the i-th ground access point and the j-th single-antenna user, a standard normal distribution random variable between the i-th ground access point and the j-th single-antenna user. The third aspect of the present application provides an electronic device, comprising: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the satellite-aided pilot allocation method based on the quantum convolutional neural network as described in the above embodiments. The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the satellite-aided pilot allocation method based on the quantum convolutional neural network as described in the above embodiments. Thus, by determining all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user, and determining the maximum value of the coefficients among all ground large-scale fading coefficients, and determining the target communication network architecture from the pre-set ground communication network architecture and the pre-set satellite communication network architecture according to the maximum value of the coefficients, all single-antenna users are served by using the target communication network architecture, and all large-scale fading coefficients corresponding to the target communication network architecture are input into the pre-set quantum convolutional neural network model to obtain the pilot index of each single-antenna user, so as to allocate pilots to each single-antenna user according to the pilot index of each single-antenna user. Thus, the pilot pollution problem caused by pilot reuse under harsh conditions is solved, the pilot allocation can be optimized, the pilot pollution is reduced, and the reliability of the system under the ultra-low delay condition is improved. Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0051] The third aspect of the present application provides an electronic device, comprising: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the satellite-aided pilot allocation method based on the quantum convolutional neural network as described in the above embodiments.

[0052] The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the satellite-aided pilot allocation method based on the quantum convolutional neural network as described in the above embodiments.

[0053] Thus, by determining all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user, and determining the maximum value of the coefficients among all ground large-scale fading coefficients, and determining the target communication network architecture from the pre-set ground communication network architecture and the pre-set satellite communication network architecture according to the maximum value of the coefficients, all single-antenna users are served by using the target communication network architecture, and all large-scale fading coefficients corresponding to the target communication network architecture are input into the pre-set quantum convolutional neural network model to obtain the pilot index of each single-antenna user, so as to allocate pilots to each single-antenna user according to the pilot index of each single-antenna user. Thus, the pilot pollution problem caused by pilot reuse under harsh conditions is solved, the pilot allocation can be optimized, the pilot pollution is reduced, and the reliability of the system under the ultra-low delay condition is improved.

[0054] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0055] ​​​​The above and / or additional aspects and advantages of the present application will become apparent and more readily appreciated from the following description of embodiments, taken in conjunction with the references to the following drawings, wherein:

[0056] Figure 1 A schematic diagram of a preset communication network architecture according to an embodiment of the present application;

[0057] Figure 2 A flowchart of a satellite-assisted pilot allocation method based on a quantum convolutional neural network according to an embodiment of the present application;

[0058] Figure 3 A flowchart of a satellite-assisted pilot allocation method based on a quantum convolutional neural network according to an embodiment of the present application;

[0059] Figure 4 A flowchart of constructing a preset quantum convolutional neural network model according to an embodiment of the present application;

[0060] Figure 5 A block schematic diagram of a satellite-assisted pilot allocation apparatus based on a quantum convolutional neural network according to an embodiment of the present application;

[0061] Figure 6 A structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0062] Embodiments of the present application are described in detail below with reference to the attached drawing figures, wherein the same or like component(s) have the same or similar designations throughout the drawing figures. The embodiments described below are presented by way of example only and are not intended to limit the present application as defined by the appended claims and their equivalents.

[0063] The following describes a satellite-assisted pilot allocation method based on a quantum convolutional neural network, according to an embodiment of this application, with reference to the accompanying drawings. Addressing the pilot pollution problem caused by pilot reuse under harsh conditions mentioned in the background art, this application provides a satellite-assisted pilot allocation method based on a quantum convolutional neural network. In this method, all large-scale fading coefficients of the ground between each ground access point and each single-antenna user, and all large-scale fading coefficients of the satellite between each satellite access point and each single-antenna user are determined. The maximum value of all large-scale fading coefficients of the ground access point is determined, and a target communication network architecture is determined from a preset ground communication network architecture and a preset satellite communication network architecture based on the maximum value of the coefficients. The target communication network architecture serves all single-antenna users, and all large-scale fading coefficients corresponding to the target communication network architecture are input into a preset quantum convolutional neural network model to obtain the pilot index of each single-antenna user. Pilot allocation is then performed on each single-antenna user based on its pilot index. This solves the pilot pollution problem caused by pilot reuse under harsh conditions, optimizes pilot allocation, reduces pilot pollution, and improves the reliability of the system under ultra-low latency conditions.

[0064] Before introducing the satellite-assisted pilot allocation method based on quantum convolutional neural networks in the embodiments of this application, let's first introduce the preset communication network architecture, such as... Figure 1 As shown, the preset communication network architecture includes a preset terrestrial communication network architecture and a preset satellite communication network architecture. The preset terrestrial communication network architecture includes at least one terrestrial access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point.

[0065] Specifically, the pre-defined terrestrial communication network architecture includes Individual with The ground access point (GAP) of the antenna, and Single-antenna user (User Equipment, UE) GAP services simultaneously One UE, Indicates the first The gap to the first Large-scale fading coefficients between individual UEs; the pre-defined satellite communication network architecture includes Individual with The satellite access point (SAP) of the root antenna, if necessary. One SAP service simultaneously One UE, Indicates the first SAP to the first Large-scale fading coefficients between individual UEs.

[0066] Specifically, Figure 2 This is a flowchart illustrating a satellite-assisted pilot allocation method based on a quantum convolutional neural network, provided in an embodiment of this application.

[0067] like Figure 2 As shown, the satellite-assisted pilot allocation method based on quantum convolutional neural networks includes the following steps:

[0068] In step S101, all ground-scale fading coefficients between each ground access point and each single-antenna user are determined, as well as all satellite-scale fading coefficients between each satellite access point and each single-antenna user.

[0069] It is understood that the embodiments of this application define and analyze the pilot pollution term for the k-th user. Specifically, when providing services using terrestrial GAP, the pilot pollution term can be represented as:

[0070] ;

[0071] When using SAP SkyService to provide services, pilot pollution terms can be represented as:

[0072] ;

[0073] in, For the first Pilot pollution term for a single antenna user The number of ground access points. For ground access point, For different Single-antenna users, For different and Single-antenna users, For reuse, not with users k The set of users with the same pilot signal In order to connect with users User sets that reuse the same pilot signal For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the nth access point to the nth access point Large-scale satellite fading coefficients between individual antenna users For the first Access point to the Large-scale satellite fading coefficients between individual antenna users For the first The access point to the first Large-scale satellite fading coefficients between individual antenna users The number of antennas available for each ground access point. The number of antennas available for each satellite access point.

[0074] As can be seen from the above formula, Indicates with UE The set of users who reuse pilot signals, as can be seen from the formula, when the UE When fixed, and The smaller the value, the more pilot pollution. The smaller it is. Furthermore, if , and (or , and If the difference between them is large enough, then The smaller the value, the fewer users reuse the pilot signal, resulting in less pilot pollution. In contrast, UEs that are geographically far apart can effectively mitigate pilot pollution and obtain a higher SINR by reusing pilot signals. In other words, the pilot pollution term is closely related to the large-scale fading coefficient.

[0075] Specifically, the formula for calculating the large-scale fading coefficient of all ground-based connections between each ground access point and each single-antenna user is as follows:

[0076] ;

[0077] The formula for calculating the large-scale fading coefficient of all satellites between each satellite access point and each single-antenna user is as follows:

[0078] ;

[0079] in, For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first the standard deviation between the first access point and the first single-antenna user, shadow fading with a standard deviation of the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, the first access point and the first single-antenna user, . All ground large-scale fading coefficients and all satellite large-scale fading coefficients

[0080] are obtained through the above calculation formulas. In step S102, the maximum value of the coefficients in all ground large-scale fading coefficients is determined, and the target communication network architecture is determined from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value.

[0081] Further, in some embodiments, determining the target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value includes: judging whether the maximum value is greater than a preset threshold; if the maximum value is greater than the preset threshold, taking the preset ground communication network architecture as the target communication network architecture, otherwise, taking the preset satellite communication network architecture as the target communication network architecture.

[0082] The preset threshold can be preset by a user, can be obtained through a limited number of experiments, or can be obtained through a limited number of computer simulations, and is not limited here.

[0083] The preset threshold can be preset by a user, can be obtained through a limited number of experiments, or can be obtained through a limited number of computer simulations, and is not limited here.

[0084] ​Specifically, a maximum value of all ground large-scale fading coefficients is determined, and when the maximum value is greater than a preset threshold , a preset ground communication network architecture is taken as a target communication network architecture, and when the maximum value is less than the preset threshold , a preset satellite communication network architecture is taken as the target communication network architecture.

[0085] In step S103, all single-antenna users are served by using the target communication network architecture, and all large-scale fading coefficients corresponding to the target communication network architecture are input into a preset quantum convolutional neural network model to obtain a pilot index of each single-antenna user, so as to perform pilot allocation on each single-antenna user according to the pilot index of each single-antenna user.

[0086] Specifically, when the maximum value is greater than the preset threshold, all single-antenna users are served by using the preset ground communication network architecture, and all ground large-scale fading coefficients are input into the preset quantum convolutional neural network model, the preset quantum convolutional neural network model outputs a first pilot index of each single-antenna user, and pilot allocation is performed on each single-antenna user according to the first pilot index of each single-antenna user; when the maximum value is greater than the preset threshold, all single-antenna users are served by using the preset satellite communication network architecture, and all satellite large-scale fading coefficients are input into the preset quantum convolutional neural network model, the preset quantum convolutional neural network model outputs a second pilot index of each single-antenna user, and pilot allocation is performed on each single-antenna user according to the second pilot index of each single-antenna user.

[0087] That is, when the maximum value of the large-scale fading coefficient in the GAP serving the user is greater than the threshold , the CF mMIMO (cell-free massive multiple-input multiple-output) architecture on the ground is used to serve these users. When the maximum value of the large-scale fading coefficient in the GAP serving the user is less than the threshold , the CF mMIMO architecture in the sky is used to serve the user.

[0088] Optionally, in some embodiments, before inputting all large-scale fading coefficients corresponding to the target communication network architecture into a preset quantum convolutional neural network model, the method includes: acquiring training set data; inputting the training set data into a preset quantum circuit, where the training set data is encoded using an RX (rotation around the X-axis) gate to obtain encoded data; performing quantum operations on the encoded data to obtain quantum states, and performing quantum entanglement on the quantum states using a CNOT (controlled NOT gate or controlled inverse gate) gate to obtain entangled quantum states; observing the entangled quantum states using a Pauli-Z gate to obtain quantum computation results, and inputting the quantum computation results into an initial neural network model for training to obtain a trained neural network model, calculating the cross-entropy loss function of the trained neural network model to obtain a loss value, optimizing the trained neural network model using the loss value to obtain a new trained neural network model; using the new trained neural network model as the initial neural network model, and re-executing the step of inputting the training set data into the preset quantum circuit until the cumulative number of training rounds reaches a preset number of training rounds, and outputting the final trained neural network model as the preset quantum convolutional neural network model.

[0089] This application embodiment can construct a pre-defined quantum convolutional neural network model, aiming to leverage the parallelism of quantum computing to improve the system's computational speed and reliability through quantum superposition and quantum interference. An HQCNN (Hybrid Quantum-Classical Neural Network) model is constructed, comprising quantum circuit design (including data encoding, quantum entanglement, and quantum observation design), quantum convolutional layers, and fully connected layers. In the quantum circuit, RX gates are used to encode the large-scale fading coefficient information of the channel into quantum states, and CNOT gates are used to establish quantum correlations between qubits. Pauli-Z measurements convert quantum information into classical data. Through training and optimization, the pre-defined quantum convolutional neural network model outputs the pilot index for each user, achieving pilot allocation.

[0090] To enable those skilled in the art to further understand the satellite-assisted pilot allocation method based on quantum convolutional neural networks in the embodiments of this application, the following detailed description is provided in conjunction with specific embodiments.

[0091] like Figure 3 As shown, Figure 3 This is a flowchart of a satellite-assisted pilot allocation method based on a quantum convolutional neural network according to an embodiment of this application.

[0092] Step 201: Initialize the number of users K, the number of access points M, the number of satellites N, and the pilot length. Large-scale fading coefficient of the ground The preset communication network architecture includes One AP, one LEO satellite and one user. Assuming all APs are connected to the CPU through ideal backhaul connections with no transmission errors and bandwidth limitations, the CPU is provided with unlimited capacity.

[0093] Step 202: Determine whether the maximum value of the ground large-scale fading coefficient of the AP serving each user is greater than a threshold value . If greater than , the GAP is used to serve the user, and step 203 is performed; if less than , the SAP is used to serve the user, and step 204 is performed. Step 203: Use the pre-set ground communication network architecture GAP as the target communication network architecture.

[0094] Step 204: Use the pre-set satellite communication network architecture SAP as the target communication network architecture.

[0095] Step 205: Input all large-scale fading coefficients corresponding to the target communication network architecture into the pre-set quantum neural network model.

[0096] Step 206: Output the pilot sequence number corresponding to each user.

[0097] Before using the satellite-assisted pilot allocation method based on the quantum convolutional neural network according to the embodiments of the present application, a pre-set quantum convolutional neural network model needs to be constructed, and the specific steps of constructing the pre-set quantum convolutional neural network model are as shown in Figure 4 :

[0098] Step 301: Input the large-scale fading coefficient matrix B of the training data set into the quantum circuit.

[0099] Step 302: Encode the input data using the RX gate:

[0100] ;

[0101] wherein, is a rotation gate around the X-axis, is the angle of rotation of the quantum state around the X-axis, is the imaginary unit.

[0102] After four quantum bits, the operation quantum state is: ;

[0103] wherein, is a four-qubit quantum state, is the quantum state of the first quantum bit, is the quantum state of the second quantum bit, is the quantum state of the third quantum bit, is the quantum state of the fourth quantum bit.

[0104] Step 303: quantum entanglement is performed using a CNOT gate, and sufficient learning of features is performed: wherein,

[0105]

[0106] Step 304: using a Pauli-Z observation, the quantum bit state is extracted, noise is reduced, and the output data variance is reduced, and a traditional fully connected layer and a ReLU activation function are used for feature learning.

[0107] Step 305: the loss function is calculated.

[0108] Step 306: the weights of the quantum convolution layer and the fully connected layer are updated.

[0109] Step 307: it is judged whether the training round epoch reaches the maximum training round R, if yes, step 308 is executed, otherwise, step 301 is executed.

[0110] Step 308: a preset quantum convolutional neural network model is output.

[0111] In summary, the embodiment of the present application makes full use of the coverage advantage of LEO satellites, provides a reliable option when the ground cell-free architecture cannot provide a reliable communication link, and enhances the continuity of coverage. In addition, compared with the traditional neural network, which usually needs a long time and higher computational complexity to process high-dimensional data, the quantum convolutional neural network can process multiple signals in parallel by using the quantum superposition state, accelerate the data processing process, and improve the learning efficiency of the model. Compared with the traditional neural network which relies on gradient descent method to find local optimal solution, the quantum convolutional neural network can effectively avoid falling into local optimum through the superposition and interference effect of quantum state.

[0112] According to the satellite-assisted pilot allocation method based on the quantum convolutional neural network provided in the embodiment of the present application, all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user are determined, the maximum value of the coefficients in all ground large-scale fading coefficients is determined, and the target communication network architecture is determined from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value of the coefficients. The target communication network architecture is used to serve all single-antenna users, and all large-scale fading coefficients corresponding to the target communication network architecture are input into the preset quantum convolutional neural network model to obtain the pilot index of each single-antenna user, so as to allocate pilots to each single-antenna user according to the pilot index of each single-antenna user. Thus, the pilot pollution problem caused by pilot reuse under harsh conditions is solved, the pilot allocation can be optimized, the pilot pollution is reduced, and the reliability of the system under the ultra-low delay condition is improved.

[0113] With reference to the accompanying drawings, the satellite-assisted pilot allocation device based on quantum convolutional neural network according to the embodiments of the present application is described.

[0114] It should be noted that the satellite-assisted pilot allocation based on quantum convolutional neural network is applied to a preset communication network architecture, wherein the preset communication network architecture includes a preset ground communication network architecture and a preset satellite communication network architecture, the preset ground communication network architecture includes at least one ground access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point.

[0115] Figure 5 is a block schematic diagram of the satellite-assisted pilot allocation device based on quantum convolutional neural network according to the embodiments of the present application.

[0116] As shown in Figure 5 The satellite-assisted pilot allocation device 10 based on quantum convolutional neural network includes a first determination module 100, a second determination module 200, and a pilot allocation module 300.

[0117] The first determination module 100 is configured to determine all ground large-scale fading coefficients between each ground access point and each single-antenna user, and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user.

[0118] The second determination module 200 is configured to determine a maximum value of the coefficients in the all ground large-scale fading coefficients, and determine a target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture according to the maximum value.

[0119] The pilot allocation module 300 is configured to serve all single-antenna users by using the target communication network architecture, and input all large-scale fading coefficients corresponding to the target communication network architecture to a preset quantum convolutional neural network model to obtain a pilot index of each single-antenna user, so as to allocate pilots to each single-antenna user according to the pilot index of each single-antenna user.

[0120] Optionally, in some embodiments, the second determination module includes a judgment unit and a generation unit.

[0121] The judgment unit is configured to judge whether the maximum value is greater than a preset threshold.

[0122] The generation unit is configured to take the preset ground communication network architecture as the target communication network architecture when the maximum value is greater than the preset threshold, and otherwise, take the preset satellite communication network architecture as the target communication network architecture.

[0123] Optionally, in some embodiments, before inputting all large-scale fading coefficients corresponding to the target communication network architecture into the preset quantum convolutional neural network model, the pilot allocation module 300 comprises an acquisition unit, an encoding unit, an operation unit, a training unit and a generation unit.

[0124] The acquisition unit is configured to acquire training set data.

[0125] The encoding unit is configured to input the training set data into a preset quantum circuit, and encode the training set data by using an RX gate in the preset quantum circuit to obtain encoded data.

[0126] The operation unit is configured to perform quantum operation on the encoded data to obtain a quantum state, and perform quantum entanglement on the quantum state by using a CNOT gate to obtain a quantum-entangled quantum state.

[0127] The training unit is configured to obtain quantum calculation results by using a Pauli-Z observation quantum-entangled quantum state, input the quantum calculation results into an initial neural network model for training to obtain a trained neural network model, calculate a cross-entropy loss function of the trained neural network model to obtain a loss value, optimize the trained neural network model by using the loss value, and obtain a new trained neural network model.

[0128] The generation unit is configured to use the new trained neural network model as the initial neural network model, and re-perform the step of inputting the training set data into the preset quantum circuit until the cumulative training round number reaches a preset training round number, and output a final trained neural network model as the preset quantum convolutional neural network model.

[0129] Optionally, in some embodiments, the encoding unit comprises an encoding subunit.

[0130] The encoding subunit is configured to encode the training set data by using an RX gate according to a preset encoding formula to obtain encoded data, wherein the preset encoding formula is:

[0131] ;

[0132] wherein, is a rotation gate around the X-axis, is an angle of rotation of the quantum state around the X-axis, is an imaginary unit.

[0133] Optionally, in some embodiments, the second determination unit comprises a determination subunit.

[0134] The determining sub-unit is configured to determine all ground large-scale fading coefficients between each ground access point and each single-antenna user and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user by using a preset large-scale fading coefficient calculation formula.

[0135] ;

[0136] wherein, is a large-scale fading coefficient between the i th access point and the j th single-antenna user, is a path loss between the i th access point and the j th single-antenna user, is a shadow fading with a standard deviation of between the i th access point and the j th single-antenna user, is a standard deviation between the i th access point and the j th single-antenna user, is a standard normal distribution random variable between the i th access point and the j th single-antenna user, that is, .

[0137] It should be noted that the foregoing explanation and description of the embodiment of the satellite-aided pilot allocation method based on the quantum convolutional neural network also applies to the embodiment of the satellite-aided pilot allocation device based on the quantum convolutional neural network, which will not be described here.

[0138] The satellite-aided pilot allocation device based on the quantum convolutional neural network according to the embodiment of the present application determines all ground large-scale fading coefficients between each ground access point and each single-antenna user and all satellite large-scale fading coefficients between each satellite access point and each single-antenna user, determines a maximum value of the coefficients in all ground large-scale fading coefficients, and determines a target communication network architecture from a preset ground communication network architecture and a preset satellite communication network architecture according to the maximum value. The target communication network architecture is used to serve all single-antenna users, and all large-scale fading coefficients corresponding to the target communication network architecture are input into a preset quantum convolutional neural network model to obtain a pilot index of each single-antenna user, so as to allocate pilots to each single-antenna user according to the pilot index of each single-antenna user. Thus, the pilot pollution problem caused by pilot reuse under harsh conditions is solved, the pilot allocation is optimized, the pilot pollution is reduced, and the reliability of the system under the ultra-low delay condition is improved. ​​​​​​​​​​​

[0139] Figure 6 A structural schematic diagram of an electronic device is provided for the embodiments of the present application. The electronic device can include:

[0140] The memory 601, the processor 602 and the computer program stored in the memory 601 and executable on the processor 602.

[0141] The processor 602 implements the satellite-assisted pilot allocation method based on the quantum convolutional neural network provided in the above embodiments when executing the program.

[0142] Further, the electronic device further includes:

[0143] The communication interface 603 is used for communication between the memory 601 and the processor 602.

[0144] The memory 601 is used to store the computer program executable on the processor 602.

[0145] The memory 601 can include a high-speed RAM (Random Access Memory) memory, and can also include a non-volatile memory, such as at least one disk memory.

[0146] If the memory 601, the processor 602 and the communication interface 603 are independently implemented, the communication interface 603, the memory 601 and the processor 602 can be connected to each other through a bus and complete communication between each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 6 In the figure, only one thick line is used to represent, but it does not mean that there is only one bus or one type of bus.

[0147] Optionally, in specific implementation, if the memory 601, the processor 602 and the communication interface 603 are integrated on a chip, the memory 601, the processor 602 and the communication interface 603 can complete communication between each other through an internal interface.

[0148] The processor 602 can be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or be configured as one or more integrated circuits implementing embodiments of the present application.

[0149] The embodiments of the present application further provide a computer readable storage medium, which has stored thereon a computer program, and the computer program is executed by a processor to implement the above satellite-aided pilot allocation method based on a quantum convolutional neural network.

[0150] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms is not necessarily directed to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or N embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0151] In addition, the terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, for example, two, three, etc., unless otherwise explicitly specified.

[0152] Any process or method descriptions in flow charts or otherwise described herein can be understood as representing code modules, segments, or portions of code which include one or more executable instructions for performing specified logic functions or steps, and the various embodiments of the preferred implementation of the present application include additional or fewer functions performed in the same order or in a different order, as appropriate, and with additional functions, in which the functions can be performed at least in part concurrently, by separate or integrated circuits, and in a variety of orders, as appropriate, as would be understood by those skilled in the art.

[0153] It should be understood that portions of the application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. As such, if implemented in hardware, and in another embodiment, any of the following technologies, known in the art, or their combinations, can be used to implement: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays, field programmable gate arrays, etc.

[0154] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment methods can be completed by programs instructing related hardware, and the programs can be stored in a computer readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0155] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A satellite-assisted pilot allocation method based on quantum convolutional neural networks, characterized in that, The satellite-assisted pilot allocation method based on quantum convolutional neural networks is applied to a preset communication network architecture, wherein the preset communication network architecture includes a preset terrestrial communication network architecture and a preset satellite communication network architecture. The preset terrestrial communication network architecture includes at least one terrestrial access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point. The method includes the following steps: Determine all ground-scale fading coefficients between each ground access point and each single-antenna user, and all satellite-scale fading coefficients between each satellite access point and each single-antenna user; Determine the maximum value of all ground-based large-scale fading coefficients, and determine the target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture based on the maximum value of the coefficients; The target communication network architecture is used to serve all single-antenna users. All large-scale fading coefficients corresponding to the target communication network architecture are input into a preset quantum convolutional neural network model to obtain the pilot index for each single-antenna user. Pilot allocation is then performed for each single-antenna user based on its pilot index. Before inputting all large-scale fading coefficients corresponding to the target communication network architecture into the preset quantum convolutional neural network model, the process includes: acquiring training set data; inputting the training set data into a preset quantum circuit, where the training set data is encoded using an RX gate to obtain encoded data; performing quantum operations on the encoded data to obtain quantum states, and performing quantum entanglement on the quantum states using a CNOT gate to obtain entangled quantum states; observing the entangled quantum states using Pauli-Z to obtain quantum computation results, and inputting the quantum computation results into an initial neural network model for training to obtain a trained neural network model, calculating the cross-entropy loss function of the trained neural network model to obtain a loss value, optimizing the trained neural network model using the loss value to obtain a new trained neural network model; using the new trained neural network model as the initial neural network model, and re-executing the step of inputting the training set data into the preset quantum circuit until the cumulative number of training rounds reaches a preset number of training rounds, and outputting the final trained neural network model as the preset quantum convolutional neural network model; The step of determining the target communication network architecture from the preset terrestrial communication network architecture and the preset satellite communication network architecture based on the maximum value of the coefficient includes: when the maximum value of the coefficient is greater than a preset threshold, using the preset terrestrial communication network architecture to serve all single-antenna users, and inputting all large-scale fading coefficients of the ground into a preset quantum convolutional neural network model, outputting a first pilot index for each single-antenna user through the preset quantum convolutional neural network model, so as to perform pilot allocation for each single-antenna user according to the first pilot index of each single-antenna user; when the maximum value of the coefficient is less than or equal to the preset threshold, using the preset satellite communication network architecture to serve all single-antenna users, and inputting all large-scale fading coefficients of the satellite into the preset quantum convolutional neural network model, outputting a second pilot index for each single-antenna user through the preset quantum convolutional neural network model, so as to perform pilot allocation for each single-antenna user according to the second pilot index of each single-antenna user.

2. The method according to claim 1, characterized in that, The process of encoding the training set data using the RX gate to obtain the encoded data includes: According to a preset encoding formula, the training set data is encoded using an RX gate to obtain the encoded data, wherein the preset encoding formula is: ； in, It is a revolving door around the X-axis. The angle of rotation of the quantum state around the X-axis. It is the imaginary unit.

3. The method according to claim 1, characterized in that, The formula for calculating the large-scale fading coefficient of all ground-based connections between each ground access point and each single-antenna user is as follows: ； The formula for calculating the large-scale fading coefficient of all satellites between each satellite access point and each single-antenna user is as follows: ； in, For the first The access point to the first Large-scale ground fading coefficient between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first The standard deviation among individual antenna users is The shadow of decline, For the first The access point to the first Standard deviation between individual antenna users For the first The access point to the first A standard normally distributed random variable among single-antenna users For the first The access point to the first Large-scale satellite fading coefficients between individual antenna users For the first The access point to the first Path loss between individual antenna users It is the first The access point to the first The standard deviation among individual antenna users is The shadow of decline, For the first The access point to the first Standard deviation between individual antenna users For the first The access point to the first A standard normally distributed random variable among single-antenna users.

4. A satellite-assisted pilot allocation device based on a quantum convolutional neural network, characterized in that, The satellite-assisted pilot allocation device based on quantum convolutional neural networks is applied to a preset communication network architecture, wherein the preset communication network architecture includes a preset terrestrial communication network architecture and a preset satellite communication network architecture. The preset terrestrial communication network architecture includes at least one terrestrial access point and at least one single-antenna user, and the preset satellite communication network architecture includes at least one satellite access point. The device includes: The first determining module is used to determine all ground-scale fading coefficients between each ground access point and each single-antenna user, and all satellite-scale fading coefficients between each satellite access point and each single-antenna user; The second determining module is used to determine the maximum value of the coefficients among all the ground-based large-scale fading coefficients, and to determine the target communication network architecture from the preset ground communication network architecture and the preset satellite communication network architecture based on the maximum value of the coefficients. A pilot allocation module is used to serve all single-antenna users using a target communication network architecture. It inputs all large-scale fading coefficients corresponding to the target communication network architecture into a preset quantum convolutional neural network model to obtain the pilot index for each single-antenna user. Pilot allocation is then performed on each single-antenna user based on its pilot index. Before inputting all large-scale fading coefficients corresponding to the target communication network architecture into the preset quantum convolutional neural network model, the pilot allocation module includes: an acquisition unit for acquiring training set data; an encoding unit for inputting the training set data into a preset quantum circuit, where the training set data is encoded using an RX gate to obtain encoded data; and a computation unit for performing quantum operations on the encoded data to obtain quantum states. The process involves: a CNOT gate to entangle the quantum state to obtain an entangled quantum state; a training unit to obtain quantum computation results by observing the entangled quantum state using Pauli-Z, inputting the quantum computation results into an initial neural network model for training to obtain a trained neural network model, calculating the cross-entropy loss function of the trained neural network model to obtain a loss value, optimizing the trained neural network model using the loss value to obtain a new trained neural network model; and a generation unit to use the new trained neural network model as the initial neural network model, and re-execute the step of inputting the training set data into a preset quantum circuit until the cumulative number of training rounds reaches a preset number of training rounds, and outputting the final trained neural network model as the preset quantum convolutional neural network model. The step of determining the target communication network architecture from the preset terrestrial communication network architecture and the preset satellite communication network architecture based on the maximum value of the coefficient includes: when the maximum value of the coefficient is greater than a preset threshold, using the preset terrestrial communication network architecture to serve all single-antenna users, and inputting all large-scale fading coefficients of the ground into a preset quantum convolutional neural network model, outputting a first pilot index for each single-antenna user through the preset quantum convolutional neural network model, so as to perform pilot allocation for each single-antenna user according to the first pilot index of each single-antenna user; when the maximum value of the coefficient is less than or equal to the preset threshold, using the preset satellite communication network architecture to serve all single-antenna users, and inputting all large-scale fading coefficients of the satellite into the preset quantum convolutional neural network model, outputting a second pilot index for each single-antenna user through the preset quantum convolutional neural network model, so as to perform pilot allocation for each single-antenna user according to the second pilot index of each single-antenna user.

5. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the satellite-assisted pilot allocation method based on a quantum convolutional neural network as described in any one of claims 1-3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the satellite-assisted pilot allocation method based on quantum convolutional neural networks as described in any one of claims 1-3.

Images