NAFD system parameter determination method and device, electronic equipment and storage medium

Abstract

The invention provides an NAFD system parameter determination method and device, electronic equipment and a storage medium, and relates to the technical field of communication, and the method comprises the steps: dividing a terminal in a first space, determining at least one terminal cluster, and determining an access point (AP) cluster corresponding to each terminal cluster in the at least one terminal cluster; determining channel information of each terminal in each terminal cluster and each AP in the corresponding AP cluster; inputting the channel information, the information of the terminals in the at least one terminal cluster and the position information of the APs in the AP cluster into a trained first network, and obtaining target system parameters output by the trained first network under the optimal balance of uplink accumulated spectrum efficiency and downlink accumulated spectrum efficiency; and sending the target system parameters to the APs in the AP cluster. According to the invention, the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency of the NAFD can be considered.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a method, device, electronic device and storage medium for determining NAFD system parameters. Background Art

[0002] The network-assisted full-duplex (NAFD) system based on the non-cellular architecture can realize the uplink and downlink service mode selection in the space domain within the same time-frequency resource. Its main working principle is: the NAFD system includes multiple access points (APs), each of which includes two service modes: uplink transmission and downlink transmission. In the same time-frequency resource, by selecting each AP to be in the uplink transmission service mode or the downlink transmission service mode, the uplink and downlink service mode selection in the space domain can be realized within the same time-frequency resource.

[0003] At present, in order to ensure the communication quality of users, the NAFD system parameters are usually optimized based on the maximum uplink network throughput or the maximum downlink network throughput. However, the uplink network throughput and the downlink network throughput restrict each other. Optimizing the NAFD system parameters based on the maximum uplink network throughput or the maximum downlink network throughput will result in the inability to take into account both the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency. Summary of the invention

[0004] The embodiments of the present application provide a method, device, electronic device and storage medium for determining NAFD system parameters to solve the problem in the prior art that the NAFD system parameters are optimized based on the maximum uplink network throughput or the maximum downlink network throughput, resulting in the inability to take into account both the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency.

[0005] In order to solve the above technical problems, the present application is implemented as follows;

[0006] In a first aspect, an embodiment of the present application provides a method for determining NAFD system parameters, which is applied to a core network device, and the method includes:

[0007] Divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0008] Determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0009] Inputting the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtaining the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0010] The target system parameters are sent to APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to terminals in a corresponding terminal cluster according to the target system parameters.

[0011] Optionally, after inputting the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, the method further includes:

[0012] Obtaining the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0013] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0014] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0015] Optionally, the target system parameters include:

[0016] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0017] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0018] Optionally, dividing the terminals in the first space to determine at least one terminal cluster includes:

[0019] Determining a three-dimensional position of each terminal in the first space;

[0020] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0021] Optionally, determining the three-dimensional position of each terminal in the first space includes:

[0022] Obtain a received signal strength RSS vector corresponding to the first terminal, where the RSS vector includes multiple RSSs, where the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, where the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and where the first terminal is any terminal in the first space;

[0023] Comparing the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0024] When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0025] Optionally, determining the access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster includes:

[0026] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0027] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0028] Optionally, the method further includes:

[0029] Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

[0030] Optionally, the first network is a deep Q network DQN.

[0031] In a second aspect, an embodiment of the present application provides a NAFD system parameter determination device, including:

[0032] A first determination module is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0033] A second determining module is used to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0034] A first obtaining module is used to input the channel information, the information of the terminals in the at least one terminal cluster and the position information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0035] The first sending module is configured to send the target system parameters to the APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

[0036] Optionally, the NAFD system parameter determination device further includes:

[0037] A second obtaining module, used to obtain the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0038] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0039] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0040] Optionally, the target system parameters include:

[0041] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0042] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0043] Optionally, the first determining module is specifically configured to:

[0044] Determining a three-dimensional position of each terminal in the first space;

[0045] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0046] Optionally, the first determining module is further configured to:

[0047] Obtain a received signal strength RSS vector corresponding to the first terminal, where the RSS vector includes multiple RSSs, where the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, where the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and where the first terminal is any terminal in the first space;

[0048] Comparing the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0049] When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0050] Optionally, the first determining module is further configured to:

[0051] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0052] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0053] Optionally, the NAFD system parameter determination device further includes:

[0054] Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

[0055] Optionally, the first network is a deep Q network DQN.

[0056] In a third aspect, an embodiment of the present application further provides an electronic device, including a transceiver and a processor.

[0057] The processor is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0058] The processor is further configured to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0059] The processor is further configured to input the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0060] The transceiver is used to send the target system parameters to the APs in the AP cluster, and the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

[0061] Optionally, the processor is further configured to:

[0062] Obtaining the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0063] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0064] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0065] Optionally, the target system parameters include:

[0066] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0067] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0068] Optionally, the processor is further configured to:

[0069] Determining a three-dimensional position of each terminal in the first space;

[0070] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0071] Optionally, the transceiver is further used to obtain a received signal strength RSS vector corresponding to the first terminal, the RSS vector includes multiple RSSs, the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and the first terminal is any terminal in the first space;

[0072] The processor is further configured to compare the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0073] When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0074] Optionally, the processor is further configured to:

[0075] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0076] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0077] Optionally, the processor is further configured to:

[0078] Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

[0079] Optionally, the first network is a deep Q network DQN.

[0080] In a fourth aspect, an embodiment of the present application further provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor; the processor is configured to read the program in the memory to implement the steps of the NAFD system parameter determination method as described in the first aspect.

[0081] In a fifth aspect, an embodiment of the present application further provides a storage medium for storing a computer program, characterized in that when the computer program is executed by a processor, the steps of the NAFD system parameter determination method as described in the first aspect are implemented.

[0082] In an embodiment of the present application, the terminals in the first space are divided, at least one terminal cluster is determined, and the access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster is determined, the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster. The channel information of each terminal in each terminal cluster and each AP in the corresponding AP cluster is determined, so that the channel information, the information of the terminals in at least one terminal cluster, and the location information of the APs in the AP cluster are input into the trained first network, and the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network are obtained. The target system parameters are sent to the APs in the AP cluster, and the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters. By sending the target system parameters to the APs in the AP cluster, the APs in the AP cluster can provide services to the terminals in the corresponding terminal cluster according to the target system parameters, so that the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency of the NAFD system deployed in the first space are balanced and optimal, thereby taking into account the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency of NAFD. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0084] Figure 1 A flowchart of a method for determining NAFD system parameters provided in an embodiment of the present application;

[0085] Figure 2 One of the simulation result schematic diagrams provided in the embodiment of the present application;

[0086] Figure 3 A second schematic diagram of a simulation result provided in an embodiment of the present application;

[0087] Figure 4 A third schematic diagram of a simulation result provided in an embodiment of the present application;

[0088] Figure 5 A fourth schematic diagram of a simulation result provided in an embodiment of the present application;

[0089] Figure 6 A structural diagram of a NAFD system parameter determination device provided in an embodiment of the present application;

[0090] Figure 7 A structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0091] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0092] In order to make the embodiments of the present application clearer, the relevant technical knowledge involved in the embodiments of the present application is first introduced as follows:

[0093] The NAFD system (hereinafter referred to as the NAFD system) based on the Cell-free radio access network (CF-RAN) architecture includes multiple APs, multiple edge distributed units (EDUs) and several user-centric distributed units (UCDUs), one UCDU can be connected to multiple EDUs, and each EDU can be connected to corresponding multiple APs. The introduction of EDU improves the data processing capability of the NAFD system, but as the number of APs connected to the EDU continues to increase, the amount of backhaul data has exploded, which leads to the continuous increase in the requirements for the throughput of the backhaul link. Since each AP in the NAFD system includes two service modes, uplink transmission and downlink transmission, in the same time-frequency resource, by selecting each AP to be in the uplink transmission service mode or the downlink transmission service mode, the uplink and downlink service mode selection in the spatial domain can be realized in the same time-frequency resource. It can be seen that compared with the traditional full-duplex system, there is no self-interference inside the AP in the NAFD system, which can reduce the complexity of the transceiver. However, although there is no self-interference within the AP in the NAFD system, there is still cross link interference (CLI), which results in the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency in the NAFD system being mutually restricted. Specifically, increasing the uplink power of the AP can improve the spectrum efficiency of the uplink user terminal, but it will also increase the interference of the uplink user terminal to the downlink user terminal, thereby reducing the spectrum efficiency of the downlink user terminal. Conversely, when the downlink power of the AP is increased, the spectrum efficiency of the downlink user terminal increases, but the interference of the downlink AP to the uplink AP increases, thereby limiting the spectrum efficiency of the uplink user terminal. In summary, there is a problem of mutual interference between the uplink cumulative spectrum efficiency (i.e., the cumulative value of the spectrum efficiency of all uplink user terminals in the system) and the downlink cumulative spectrum efficiency (i.e., the cumulative value of the spectrum efficiency of all downlink user terminals in the system). At present, in order to ensure the quality of user communication, the NAFD system parameters are usually optimized based on the maximum uplink network throughput or downlink network throughput to maximize the uplink network throughput or downlink network throughput of the NAFD system. However, similar to the mutual constraints between the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency, the uplink network throughput and the downlink network throughput also have mutual constraints. Specifically, an increase in the uplink network throughput can increase the uplink cumulative spectrum efficiency, while causing the downlink network throughput to be limited, which in turn limits the downlink cumulative spectrum efficiency; an increase in the downlink network throughput can increase the downlink cumulative spectrum efficiency, while causing the uplink network throughput to be limited, which in turn limits the uplink cumulative spectrum efficiency.It can be seen that in the prior art, optimizing NAFD system parameters based on the maximum uplink network throughput or downlink network throughput will result in the inability to take into account both the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency. Based on this, an embodiment of the present application provides a method for determining NAFD system parameters, which achieves multi-objective optimization of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency by determining the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency; by sending the target system parameters to each AP of the NAFD system, each AP can provide services to the corresponding user terminal according to the target system parameters, thereby making the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency of the NAFD system optimally balanced, and achieving the simultaneous consideration of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency.

[0094] The following is a detailed description of the NAFD system parameter determination method provided in the embodiment of the present application.

[0095] See also Figure 1 , Figure 1 is a flow chart of a method for determining NAFD system parameters provided in an embodiment of the present application. Figure 1 As shown, the following steps are included:

[0096] Step 101: divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0097] Step 102: Determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0098] Step 103: input the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0099] Step 104: Send the target system parameters to the APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

[0100] In step 101, the first space is a preset space, which may be a space where the system parameters of the NAFD system need to be adjusted so that the deployed NAFD system can achieve the best balance between the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency. The core network device divides the terminals in the first space and may determine at least one terminal cluster, each of which includes at least one terminal. An AP cluster corresponding to each terminal cluster in the at least one terminal cluster is determined, each of which includes at least one AP, and the APs in each AP cluster are used to provide services to the terminals in the corresponding terminal cluster. In this way, the APs in each AP cluster only provide services to the terminals in the corresponding terminal cluster, so that when determining the AP corresponding to a terminal in the first space, it is not necessary to select at least one AP serving the terminal from all the APs in the first space, but only to select at least one AP serving the terminal from the AP cluster corresponding to the terminal cluster based on the terminal cluster to which the terminal belongs, which is conducive to reducing the calculation burden when selecting the AP; and the APs in each AP cluster only provide services to the terminals in the corresponding terminal cluster, which is conducive to reducing the complexity of the AP processing different user data.

[0101] In step 102, the core network device determines the channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster, and the channel information may be a channel matrix. In the embodiment of the present application, the channel between each terminal in each terminal cluster and each AP in the corresponding AP cluster may be a Rice fading channel or a Rayleigh channel.

[0102] In order to more clearly understand the technical solution of the embodiment of the present application, the channel is defined as a Rice fading channel as an example, and the channel matrix is ​​exemplarily explained.

[0103] When the channel is modeled as a Rice fading channel, the channel matrix between the kth user UE and the nth AP can be expressed as:

[0104]

[0105] where β n,k represents the large-scale fading coefficient, represents the small-scale fast fading vector. n,k represents the Rice factor; It is the deterministic part of the line of sight (LoS), i.e., the direct distance between the kth user equipment (UE) and the nth AP; The non-line of sight (NLoS) part, Right now Satisfy CN(0,I L ) distribution, CN(0,I L) is the mean value, “0” is the mean value, and “I L " is the variance. Based on the above channel matrix, the LoS component can be defined as The semi-positive definite covariance matrix of the NLoS component is Therefore, the channel between the kth UE and the nth AP satisfies

[0106] In step 103, the core network device inputs the channel information, the information of the terminals in at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network. The information of the terminal includes the location information of the terminal and the mode of the terminal, etc. The mode of the terminal is that the terminal is a terminal in the uplink (hereinafter referred to as the uplink terminal) or a terminal in the downlink (hereinafter referred to as the downlink terminal). By inputting the above information into the trained first network, the trained first network can output the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency based on the above information. By determining the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency, the multi-objective optimization of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency can be achieved, and the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency can be taken into account at the same time. The target system parameters include at least the mode of the AP in the AP cluster, and the mode of each AP includes that the AP is an AP in the uplink service mode (hereinafter referred to as the uplink AP) or an AP in the downlink service mode (hereinafter referred to as the downlink AP). The uplink AP is used to provide services to the uplink terminals in the corresponding terminal cluster, and the downlink AP is used to provide services to the downlink terminals in the corresponding terminal cluster. The uplink cumulative spectrum efficiency refers to the sum of the spectrum efficiencies of all uplink user terminals in the first space; the downlink cumulative spectrum efficiency refers to the sum of the spectrum efficiencies of all downlink user terminals in the first space.

[0107] It should be noted that the trained first network adjusts the system parameters through the core network device to obtain the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency under different system parameters. Among them, after each adjustment of the system parameters by the core network device, the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency shall be determined based on the current system parameters and the information input into the first network, and the next adjustment of the system parameters shall be based on the current system parameters. As an example, the initial system parameters of the core network device are randomly determined. After the core network device adjusts the initial system parameters for the Mth time (each adjustment may at least include switching a certain AP from an uplink mode to a downlink mode or from a downlink mode to an uplink mode), the Mth uplink cumulative spectral efficiency and the Mth downlink cumulative spectral efficiency are determined based on the initial system parameters adjusted for the Mth time and the information input to the first network; if the Mth uplink cumulative spectral efficiency and the Mth downlink cumulative spectral efficiency are not balanced and optimal, then on the basis of the initial system parameters adjusted for the Mth time, an (M+1)th adjustment is performed, and based on the initial system parameters adjusted for the (M+1)th time and the information input to the first network, the (M+1)th uplink cumulative spectral efficiency and the (M+1)th downlink cumulative spectral efficiency are determined. When the (M+1)th uplink cumulative spectral efficiency and the (M+1)th downlink cumulative spectral efficiency are balanced and optimal, the (M+1)th adjusted initial system parameters output by the first network (i.e., the target system parameters) can be obtained, where M is an integer greater than 1.

[0108] As an example, the optimization of uplink cumulative spectrum efficiency can be defined as optimization problem P1, and the optimization of downlink cumulative spectrum efficiency can be defined as P2. Since there are mutual constraints between the two optimization problems P1 and P2, it is impossible to achieve the optimal solution at the same time. Therefore, the trade-off between P1 and P2 can be achieved based on multi-objective optimization problems (MOOPs). F is used to represent the optimal vector combining the two optimization problems. MOOP can be expressed as:

[0109]

[0110] The goal of MOOP is to achieve a trade-off between P1 and P2 so that the uplink cumulative spectrum efficiency P1 and the downlink cumulative spectrum efficiency P2 are optimally balanced, where stC1~Cn represent 1 to n constraints that P1 and P2 must meet, and n is an integer greater than 1.

[0111] In step 104, the core network device sends target system parameters to the APs in the AP cluster so that the APs in the AP cluster can provide services to the terminals in the corresponding terminal cluster according to the target system parameters, so that the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency of the NAFD system deployed in the first space are optimally balanced.

[0112] In some embodiments, the adjustment of the system parameters in the above step 103 may be to adjust only the mode element of the AP each time, and the power transmitted by the AP to the terminal may always be a fixed power. In this way, the target system parameters output by the first network may include: the mode element of each AP under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency, the mode element may be "0" or "1", the mode element "1" indicates that the AP is an uplink AP in the uplink service mode, and the mode element "0" indicates that the AP is a downlink AP in the downlink service mode. After determining the mode element of each AP, each uplink AP can provide services to the uplink terminals of the corresponding terminal cluster according to the preset fixed power, and each downlink AP can provide services to the downlink terminals of the corresponding terminal cluster according to the preset fixed power. For example: if the target parameters include: the mode elements of the three APs in the first AP cluster are "1", then the three APs are used to allocate power to each uplink terminal in the first terminal cluster according to the preset fixed power, and the first terminal cluster is the terminal cluster corresponding to the first AP cluster.

[0113] In other embodiments, the adjustment of the system parameters in the above step 103 may be to adjust the mode elements of the AP and to adjust the allocated power of the user terminal. It should be noted that the above adjustment of the mode elements of the AP and the allocated power of the user terminal does not require the adjustment of the mode elements of a certain AP and the power allocated to different users by the AP at the same time each time, and may also be adjusted in batches, for example, only the mode elements of a certain AP are adjusted at a certain time, and the allocated power of a certain terminal user may be adjusted in the next adjustment. In this way, the adjustment of the system parameters is to adjust the mode elements of the AP and the allocated power of the user terminal, so that the target system parameters output by the first network may include: the mode elements of each AP and the allocated power of each user terminal under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency. It should be noted that the allocated power corresponding to the same user terminal is the same. For example: if the mode elements of the three APs in the first AP cluster are "1" and the allocated power of the first uplink terminal is 0.8w, then the three APs will allocate power to the first uplink terminal according to 0.8w, and the first uplink terminal is the uplink terminal in the terminal cluster corresponding to the first AP cluster.

[0114] In an embodiment of the present application, by obtaining the target system parameters under the optimal balance of the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency output by the trained first network, and sending the target system parameters to the APs in the AP cluster, the APs in the AP cluster can provide services to the terminals in the corresponding terminal cluster according to the target system parameters, thereby enabling the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency of the NAFD system deployed in the first space to achieve a balanced optimal state, taking into account the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency of the NAFD.

[0115] Optionally, after inputting the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, the method further includes:

[0116] Obtaining the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0117] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0118] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0119] In this implementation, the first model is a predetermined model, which is used to characterize the mapping relationship between the uplink cumulative spectrum efficiency and the system parameters, channel information, information of the terminals in at least one terminal cluster, and the location information of the APs in the AP cluster. Among them, the channel information, the information of the terminals in at least one terminal cluster, and the location information of the APs in the AP cluster are fixed parameters input into the trained first network, and the system parameters are adjustable parameters. Through the first model, the fixed parameters input into the first network, and the system parameters after each adjustment, the uplink cumulative spectrum efficiency corresponding to the current system parameters after each adjustment of the system parameters can be determined. Based on this, the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency corresponding to the target system parameters can be calculated according to the target system parameters, the first model, and the fixed parameters input into the first network.

[0120] The second model is also a predetermined model, which is used to characterize the mapping relationship between the downlink cumulative spectrum efficiency and system parameters, channel information, information of terminals in at least one terminal cluster, and location information of APs in the AP cluster.

[0121] In the embodiment of the present application, the parameters output by the first network are set to include the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency, so that the core network device can obtain the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency output by the first network. In this way, based on the characteristics of the uplink cumulative spectral efficiency and the downlink cumulative spectral efficiency, it is convenient to subsequently improve the NAFD system and the NAFD system parameters.

[0122] Optionally, the target system parameters include:

[0123] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0124] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0125] In the NAFD system in the CF-RAN architecture in the first space, it is assumed that the system consists of several UCDUs, M EDUs, and N APs equipped with L antennas; and it is assumed that there are K user terminals in the first space, including K UL uplink user terminals and K DL downlink user terminals. Each user is equipped with a single antenna, and all users and APs are evenly distributed in the scene. In a certain time slot, assume that the system parameters include N U The mode of AP is uplink service mode, N D The mode of each AP is downlink service working mode.

[0126] The first model and the second model are determined as follows:

[0127] Step 1: NAFD system channel modeling.

[0128] According to the channel information of the first network after input training (which can be determined by formula 1) and the information of the terminals in the at least one terminal cluster and the position information of the APs in the AP cluster:

[0129] The uplink channel matrix of the lth uplink UE is defined as

[0130] The downlink channel matrix of the lth downlink UE is g dl,l ;

[0131] The interference matrix between the lth downlink AP and all uplink APs is defined as G I,l , the interference matrix between all downlink APs and all uplink APs is defined as

[0132] The interference matrix between the lth uplink UE and all downlink UEs is defined as g t,l .

[0133] Among them, the above channels are all modeled as Rician fading channels.

[0134] Among them, in order to make the final determined uplink cumulative spectral efficiency and downlink cumulative spectral efficiency more accurate, the embodiment of the present application adopts a channel estimation scheme to estimate the channel state information of the uplink access link and the downlink access link in the scenario. As an example, the MMSE algorithm can be used for channel estimation, and the estimated channel matrix is ​​obtained by the MMSE algorithm.

[0135] Step 2: Based on the channel matrix in step 1, an uplink signal model in the first model is established, and a downlink signal model in the second model is established.

[0136] For ease of understanding, the following takes the cth terminal cluster and the cth AP cluster corresponding to the cth terminal cluster as an example to establish an uplink signal of the uth uplink user terminal in the cth terminal cluster received at the UCDU:

[0137] Assume that the fixed parameters of the first network after input training are determined to include K in the cth terminal cluster U,c uplink user terminals and K D,c downlink user terminals.

[0138] It is assumed that, in a certain time slot, the current system parameter is the first system parameter, and the c-th AP cluster contains N APs according to the first system parameter. U,c Uplink AP and N D,c A downlink AP.

[0139] Step 21: Establish an uplink signal model in the first model.

[0140] For uplink transmission, in each time slot, all uplink APs receive the signal from the uplink UE. Then the signal received by the nth uplink AP in the cth AP cluster is expressed as:

[0141]

[0142] Among them, g ul,n,u represents the uplink channel matrix of the u-th uplink user terminal (obtained through step 1), p U,u and p D,r represent the transmission power of the u-th uplink user terminal and the r-th downlink user terminal respectively (obtained according to the first system parameter). I represents the interference matrix between APs, w r represents the downlink precoding matrix of downlink user terminal k, s u Indicates the data signal of the uplink user terminal, which satisfies represents the data signal after conjugate transposition, E is the expectation, x r Indicates the data signal of the downlink user terminal, satisfying Represents the data signal after conjugate transposition. represents a Gaussian white noise vector.

[0143] Combine the uplink signals received by all uplink APs to determine the signal received at the EDU: in, Represents the association vector between EDU and AP. If the EDU and the uplink AP are wired connected, the corresponding vector element is set to 1, otherwise it is 0 (i.e., the signal of the uth uplink user terminal received at each EDU is determined). u represents the receiver vector of the u-th uplink user terminal. This scheme adopts the MRC receiver, then in represents the uplink channel matrix estimated by using the MMSE channel estimation scheme (obtained through step 1).

[0144] After determining the signals received at multiple EDUs, the signals received at the multiple EDUs are integrated to determine the uplink received signal at the UCDU connected to the multiple EDUs. The uplink received signal after weighted calculation of the user-centric distribution unit UCDU is expressed as: ξ=[ξ1,...,ξ m ] represents the weighted sum matrix calculated by EDU in UCDU. The weighted sum matrix can be set according to actual needs. The weight corresponding to each EDU can be set to be different or the weight corresponding to each EDU can be set to be the same.

[0145] Step 22: Establish a downlink signal model in the second model.

[0146] For downlink transmission, the baseband signal is generated in the core network and transmitted to the downlink user terminal in the cth terminal cluster through the downlink AP in the cth AP cluster. The signal received by the uth downlink user terminal in the cth terminal cluster is expressed as:

[0147]

[0148] Among them, g dl,u represents the downlink channel matrix of the u-th downlink user terminal, The downlink channel matrix of the u-th downlink user terminal after conjugate transposition is represented by g t,u,r is the inter-user interference channel between the u-th uplink user terminal in the cluster and the r-th downlink user terminal in the cluster, p D,u and p U,r denote the transmission power of the u-th downlink user and the r-th uplink user, respectively, W l represents the downlink precoding matrix of the lth downlink user in the cth terminal cluster, s r Represents the data signal of the rth downlink user and uplink user in the cth terminal cluster, satisfying x l represents the data signal of the lth downlink user in the cth terminal cluster, satisfying represents the downstream additive white Gaussian noise vector.

[0149] Step 3: Based on step 1 and step 2, a spectrum efficiency model of each uplink user terminal in the first model is established, and a spectrum efficiency model of each downlink user terminal in the second model is established.

[0150] It should be noted that the uplink or downlink spectrum efficiency of the system can be expressed as: Since the pilot length τ and the coherent time block length T are fixed parameters, when the pilot length τ and the coherent time block length T used for channel estimation are determined and the signal reachable rate R of the user terminal is determined, the spectrum efficiency SE can be determined based on the formula. Moreover, since the spectrum efficiency SE is positively correlated with the signal reachable rate R, when the uplink cumulative reachable rate and the downlink cumulative reachable rate in the embodiment of the present application are optimally balanced, the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency are also optimally balanced. Based on this, the present application can directly replace the user's spectrum efficiency with the signal reachable rate R, that is, the uplink signal reachable rate model of each uplink terminal can be used to replace the spectrum efficiency model of each uplink terminal; the downlink signal reachable rate model of each downlink terminal can be used to replace the spectrum efficiency model of each downlink terminal. The following specifically explains how to determine the uplink signal reachable rate model of each uplink terminal and how to determine the downlink signal reachable rate model of each downlink terminal.

[0151] Step 31: Establish an uplink signal achievable rate model for each uplink user in the first model

[0152] According to the uplink signal transmission model of the NAFD system, the uplink signal achievable rate of the u-th uplink user terminal in the c-th terminal cluster is The model is expressed as:

[0153]

[0154] in, represents the estimated uplink channel matrix of the u-th uplink user terminal, The rest of the part is expressed as:

[0155]

[0156]

[0157]

[0158]

[0159] in, represents the channel estimation error matrix of the uplink user terminal, Γ int,u , Γ est,u , Γ CLI,u and Γ Noi,uThey represent the interference from other uplink user terminals, channel estimation error, cross-link interference and noise respectively. Among them, the cross-link interference can be eliminated by using the beamforming training mechanism to improve the uplink spectrum efficiency of the NAFD system.

[0160] Step 32: Establish a downlink signal achievable rate model for each downlink user terminal in the second model

[0161] In the downlink data transmission phase, all downlink APs transmit downlink signals to the corresponding downlink UEs. The effective downlink channel matrix is ​​estimated by using the beamforming training mechanism at the downlink user terminal. Then in step 22, the signal received by the lth downlink UE in the cth terminal cluster can be rewritten as:

[0162]

[0163] in μ l,i The channel estimation error is Independent of each other. Let K dl is the set of downlink user terminals in the cth terminal cluster, Right now Including the downlink terminals in the cth terminal cluster except the lth downlink terminal, K ul is the set of uplink user terminals in the c terminal clusters, σ dl That is, in step 22 The achievable rate of the lth downlink user terminal is The model is expressed as:

[0164]

[0165] Step 4, after determining the model of the achievable rate of the uplink user terminal in the first model, the achievable rate of each uplink user terminal under any system parameters can be determined based on the fixed parameters and adjustable system parameters input into the first network (the achievable rate can replace the spectrum efficiency), and the cumulative uplink spectrum efficiency can be determined by adding up the achievable rates of all uplink user terminals.

[0166] Similarly, after determining the model of the achievable rate of the downlink user terminal in the second model, the achievable rate of each downlink user terminal under any system parameters can be determined based on the fixed parameters and adjustable system parameters input into the first network (the achievable rate can replace the spectrum efficiency), and the cumulative downlink spectrum efficiency can be determined by adding up the achievable rates of all downlink user terminals.

[0167] It should be noted that P1 in the above formula 2 is the optimization problem of the uplink cumulative spectrum efficiency (which can be replaced by the optimization problem of the cumulative value of the reachable rate of the uplink user terminal in step 4), and P2 in the above formula 2 is the optimization problem of the downlink cumulative spectrum efficiency (which can be replaced by the optimization problem of the cumulative value of the reachable rate of the downlink user terminal in step 4).

[0168] Based on this, the optimization problem P1 can be described as:

[0169]

[0170] stC1:

[0171] C2:

[0172] C3:

[0173] C4:

[0174] C5:

[0175] Where B represents the number of terminal clusters of the first network after input training, K U,b and K D,b They respectively represent the number of uplink user terminals and the number of downlink user terminals in the b-th terminal cluster. represents the power vector allocated to the uplink user terminal in the b-th terminal cluster, D represents the node mode selection vector, and when the vector element d i When it is 1, it indicates that the i-th AP is in the uplink service mode, and when it is 0, it indicates that the i-th AP is in the downlink service mode. The constraint C1 indicates the mode selection restriction. C2 and C3 indicate the uplink and downlink power restrictions. Indicates the power threshold of the uplink user. Indicates the downlink data power threshold. C4 and C5 are the Quality of Service (QoS) limits for downlink and uplink users, satisfying and

[0176] The optimization problem P2 can be described as:

[0177]

[0178] stC1~C5

[0179] The constraints of optimization problem P2 are the same as those of optimization problem P1. D,b Indicates the number of downlink users in the bth user cluster. Represents the power vector allocated to downlink users in the cluster.

[0180] The constraints of the optimization problem P2 are the same as those of the optimization problem P1, so “C1-Cn” in the above formula 2 is also “C1-C5” in the embodiment of the present application.

[0181] Based on the above, the cumulative uplink spectrum efficiency corresponding to any system parameter can be determined according to the first model, the fixed parameters of the first network after input training, and any system parameter; the cumulative downlink spectrum efficiency and the cumulative downlink spectrum efficiency corresponding to any system parameter can be determined according to the second model, the fixed parameters of the first network after input training, and any system parameter. Moreover, any system parameter may include the service mode of each AP and the allocated power corresponding to each terminal user. Based on this, the target system parameters in the embodiments of the present application may include: the service mode of each AP in the AP cluster; the power allocated by the uplink AP in each AP cluster to each uplink terminal in the corresponding terminal cluster, and the power allocated by the downlink AP in each AP cluster to the downlink terminal in the corresponding terminal cluster.

[0182] In the related art, the system spectrum efficiency can be improved by selecting the working mode of the AP in the spatial domain within the same time-frequency resource block or by optimizing the power allocated to each user. Specifically, the allocated power corresponding to each user terminal is first determined according to the power allocation method, and then the service mode of each AP is determined by the Q-learning algorithm. It can be seen that the related art does not consider the solution of jointly optimizing multiple system parameters, and the allocation efficiency corresponding to the user and the service mode of the AP are determined in turn, resulting in limited optimization efficiency of system parameters.

[0183] In the embodiment of the present application, by obtaining the target system parameters, the service mode of each AP in the AP cluster can be obtained at the same time; the power of the uplink AP in each AP cluster allocated to each uplink terminal in the corresponding terminal cluster, and the power of the downlink AP in each AP cluster allocated to the downlink terminal in the corresponding terminal cluster. It can be seen that the embodiment of the present application can improve the optimization efficiency of system parameters.

[0184] In related technologies, NAFD performance analysis and optimization are mainly targeted at two-dimensional planes, without considering the three-dimensional position of users in low-altitude stereoscopic coverage scenarios. It can be seen that the existing positioning solutions cannot be applied in low-altitude stereoscopic coverage scenarios to achieve NAFD system performance analysis and optimization and improve system performance.

[0185] Based on this, in some implementations, dividing the terminals in the first space to determine at least one terminal cluster includes:

[0186] Determining a three-dimensional position of each terminal in the first space;

[0187] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0188] In this implementation, the three-dimensional position of each terminal in the first space is determined, and after the three-dimensional position of each terminal in the first space is determined, the terminals in the first space may be divided. Specifically, terminals whose three-dimensional positions are in the same position range are divided into the same terminal cluster.

[0189] As an example, in the embodiment of the present application, the K-means algorithm may be used to divide the terminals. Specifically, the K-means algorithm is used to cluster the user terminals, and the user terminals with close geographical locations (i.e., in the same location range) are regarded as a whole and are divided into the same terminal cluster. Thus, a total of B terminal clusters can be formed in the first space, and the user terminal set in the first space can be expressed as B = {B1,...,B B}, where B b Represents the set of user terminals in the bth terminal cluster.

[0190] In the embodiment of the present application, the three-dimensional position of each user terminal is taken into account, so that the embodiment of the present application can be applied in low-altitude three-dimensional coverage scenarios to achieve NAFD system performance analysis and optimization and improve system performance. In addition, in the embodiment of the present application, the geographical locations of the terminals in the same terminal cluster are close, so the effect of each AP in the AP cluster serving each terminal in the corresponding terminal cluster is roughly the same, which is conducive to balancing the communication quality between each terminal in the terminal cluster and the AP.

[0191] Optionally, determining the three-dimensional position of each terminal in the first space includes:

[0192] Obtain a received signal strength (RSS) vector corresponding to the first terminal, where the RSS vector includes multiple RSSs, where the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, where the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and where the first terminal is any terminal in the first space;

[0193] Comparing the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0194] When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0195] In this implementation, the first terminal sends a reference signal to multiple APs in the first space, and the first terminal may be an active terminal in the scene. After the multiple APs in the first space receive the first reference signal sent by the first terminal, the signal strength (RSS) of the received first reference signal may be determined. The core network device may obtain the RSS of the first reference signal in multiple APs (i.e., obtain the RSS vector corresponding to the first terminal).

[0196] It should be understood that the distance between the terminal and the base station can be determined based on the RSS, and the specific three-dimensional position of the terminal can be determined based on multiple RSSs. In the embodiment of the present application, a fingerprint library is preset, and each RSS vector (also referred to as a channel fingerprint) in the fingerprint library is associated with a three-dimensional position. Based on this, after the core network device obtains the RSS vector corresponding to the first terminal, the RSS vector can be used as a fingerprint to compare the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library. In this way, when the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal can be quickly determined based on the fingerprint library.

[0197] In the embodiment of the present application, the RSS vector corresponding to the terminal is determined, and the RSS vector is used as a fingerprint to compare with the RSS vector in the fingerprint library, so that the three-dimensional position of the terminal can be determined. The embodiment of the present application determines the three-dimensional position of the terminal based on fingerprint positioning, which can overcome the multipath effect and the problem of inaccurate positioning in complex environments, and can achieve high-precision real-time positioning with low complexity.

[0198] Optionally, determining the access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster includes:

[0199] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0200] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0201] In this implementation, the channel quality between each terminal in the first terminal cluster and each AP in the first space is determined, and when the channel quality between the second terminal and the first AP meets the preset condition, the first AP is determined to be an AP in the first AP cluster, and the first AP is used to provide services to the terminals in the first terminal cluster. As an example, the preset condition may be that among the channel quality between the second terminal and each AP in the first space, the AP with the best channel quality is selected to serve the terminals in the first terminal cluster (that is, when the channel quality between the second terminal and the first AP is the best, the first AP is determined to be an AP in the first AP cluster); the preset condition may also be that among the channel quality between the second terminal and each AP in the first space, the APs ranked in the top N in terms of channel quality are selected to serve the terminals in the first terminal cluster, where N is an integer greater than or equal to 1; the preset condition may also be that among the channel quality between the second terminal and each AP in the first space, the AP with a channel quality exceeding a preset threshold is selected to serve the terminals in the first terminal cluster.

[0202] As an example, the channel quality in the embodiment of the present application can be determined based on RSS. RSS is positively correlated with the channel quality. When the RSS of the reference signal received by the first AP from the second terminal is higher, it indicates that the channel quality between the first AP and the second terminal is better.

[0203] The terminals in each terminal cluster select several APs to serve them according to the channel quality and preset conditions. The APs selected by the terminals in the same terminal cluster are divided into one AP cluster. As an example, suppose the AP set selected by the kth user in the bth terminal cluster is S k,b ,but represents the AP cluster serving the bth terminal cluster, where K b Indicates the number of terminals contained in terminal cluster b.

[0204] In the embodiment of the present application, the AP is determined to be usable to serve the terminals in the first terminal cluster only when the channel quality between the terminals in the first terminal cluster and the AP meets the preset conditions. In this way, the service effect of the AP can be improved, and then the communication quality of the users can be improved.

[0205] In some embodiments, the core network device may allocate different time-frequency resources to geographically adjacent AP clusters, and the APs in the AP cluster provide services to the corresponding terminals according to the corresponding time-frequency resources. By allocating different time-frequency resources to geographically adjacent AP clusters, inter-cluster interference can be reduced and communication efficiency can be improved.

[0206] In other embodiments, the core network device may also allocate different time-frequency resources to each AP cluster, and the time-frequency resources are used by the APs in each AP cluster to provide services to the corresponding terminals according to the corresponding time-frequency resources. By allocating different time-frequency resources to each AP cluster, inter-cluster interference can be further reduced and communication efficiency can be improved.

[0207] Optionally, the first network is a deep Q network DQN.

[0208] In the prior art, the node mode selection of AP is usually determined by Q-learning algorithm. However, the Q-learning algorithm has a high computational complexity and is difficult to apply to the actual situation where there are many APs and terminals. In the embodiment of the present application, a DQN network is used. After the DQN network is trained, the required fixed parameters are input into the trained DQN network. The trained DQN network will quickly output the target system parameters that can achieve the best balance between the cumulative uplink spectrum efficiency and the cumulative downlink spectrum efficiency under the current fixed parameters. The target system parameters may include: the mode of each AP in the AP cluster, the power of the uplink AP in each AP cluster allocated to each uplink terminal in the corresponding terminal cluster, and the power of the downlink AP in each AP cluster allocated to the downlink terminal in the corresponding terminal cluster (i.e., the AP uplink and downlink mode allocation vector D opt and the optimal user power allocation vector and )

[0209] The following is a detailed description of the training of the DQN network.

[0210] The NAFD system in the second space is regarded as a large environment, wherein the second space includes the first space, and the second space may be the entire large area space where the NAFD system is deployed;

[0211] The core network device is used as the intelligent agent of the reinforcement learning framework. The core network device can observe the system to obtain environmental information (including terminal clustering information, AP clustering information, channel information of each terminal in each terminal cluster and each AP in the corresponding AP cluster, information of each terminal and location information of AP in the AP cluster, etc.). The core network device makes intelligent agent decisions to adjust system parameters, and determines the system parameters as including AP uplink and downlink mode allocation vector D opt and the user power allocation vector and Each action of the core network device is to adjust the system parameters according to the preset rules. Before the core network device starts to act, the core network device will first randomly determine a system parameter. The first action of the core network device is to adjust the randomly determined system parameter, and the second action of the core network device is to adjust the system parameter after the first action.

[0212] The state of the environment at step t is modeled as a vector Where D(t), P UL (t) and P DL (t) represent the node mode vector and uplink and downlink user power vector corresponding to the environment state at step t. Since each AP has only two working modes, uplink and downlink, and the user power is limited, when the core network makes a decision to change the working mode of the AP and the user power, the action should be to switch a certain AP from uplink mode to downlink mode or from downlink mode to uplink mode, while increasing or decreasing or not modifying the power allocated to different users. The action at step t is defined as Indicates switching the working mode of the AP. and Indicates the power of the AP in the AP cluster modified in step t allocated to the uplink or downlink user in the corresponding terminal. Each step only modifies the allocated power of one user and only adjusts it by 0.1W. After each action taken by the agent, the rates of uplink and downlink users will be affected accordingly, thereby changing the cumulative spectrum efficiency of uplink and downlink. After each action, the agent will receive a corresponding reward r(t), which is defined as:

[0213]

[0214] Wherein, P1 is the uplink cumulative spectrum efficiency, and P2 is the downlink cumulative spectrum efficiency. The specific expressions of P1 and P2 can be determined according to the above embodiment; α represents a coefficient to ensure network convergence, which can be set according to actual needs. When more attention is paid to the downlink cumulative spectrum efficiency, α can be set to a value greater than 1, and when more attention is paid to the uplink cumulative spectrum efficiency, α can be set to a value less than 1; It represents the function used to ensure that users can achieve the minimum service quality. u0(x) is a step function, which is 0 when x≤0 and 1 otherwise. λ is a constant. The reward function is used to calculate the loss function to ensure that the algorithm can gradually converge to the optimal value.

[0215] Based on the state-action Q value and the ε greedy strategy, the consecutive actions with the highest Q value are given a fixed selection probability ε t , when the Q value is the highest, the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency reach a balanced and optimal state. The DQN network can output the highest Q value and the continuous action corresponding to the highest Q value, and can output s at the highest Q value. The output s is the target system parameter under which the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency reach a balanced and optimal state.

[0216] When exploring states with higher reward values, the agent will randomly select an action and calculate its Q value, and then put it into the experience pool to avoid repeated calculations. In each step, after calculating the reward, the agent will merge the new state information and record the current state information, reward value, next state information and current action into the experience revisit pool, represented by the vector d(t) = [s(t), a(t), r(t), s(t+1)]. Randomly transfer the minimum batch samples from the experience pool to calculate the relative training error of the model in this training round, and obtain the corresponding loss function The state s′ with the maximum value, A represents the action set, and Q(s(t), a(t)) represents the Q value corresponding to the action a(t) selected when the state is s(t).

[0217] Through multiple rounds of learning and training, the trained model tends to be stable. For any NAFD system deployed in the third space (the third space is the second space, or a subspace of the second space), the intelligent agent can find a set of continuous actions leading to the optimal trade-off state of maximizing the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency, and determine the target parameters for the lowest balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency. By obtaining the target system parameters output by the DQN network, the dynamic selection of AP uplink and downlink modes and the dynamic allocation of user power can be determined.

[0218] In the embodiments of the present application, deep reinforcement learning DQN is used to solve multi-objective optimization problems. Compared with the Q-learning algorithm, the use of DQN can quickly realize the dynamic configuration of AP node resources and allocated power resources according to real scenarios with lower complexity, and can be effectively applied to low-altitude three-dimensional coverage scenarios with limited power and large coverage range.

[0219] To further illustrate the technical effect of the embodiment of the present application, the NAFD system parameter determination method of the embodiment of the present application is used for simulation below, and the simulation results are specifically analyzed.

[0220] Assume a multi-layer indoor scenario, where user terminals and base stations are randomly distributed in a 1000*1000m scenario. Assume that the system contains 2 EDUs, 12 APs, 10 uplink user terminals, and 10 downlink user terminals.

[0221] like Figure 2 and Figure 3 As shown, Figure 2 It shows the distribution of users and APs in the scene. Figure 3The situation of AP clustering and terminal clustering in the xy plane is shown. The simulation results show that users tend to choose APs that are close to them for service, that is, the geographical location of the terminal cluster is close to the geographical location of the corresponding AP cluster. It is easy to understand that the closer the geographical location of the AP is to the user terminal, the better the service effect of the AP to the user terminal, and thus the better the communication quality of the user terminal. By adopting the technical solution of the embodiment of the present application, the geographical location of the AP cluster is close to the geographical location of the corresponding terminal cluster, which is conducive to providing communication quality for the terminals in the cluster, that is, verifying the effectiveness of the AP selection solution based on fingerprint positioning technology and channel quality in the embodiment of the present application.

[0222] See also Figure 4 , Figure 4 The uplink spectrum efficiency and downlink spectrum efficiency of AP clustering based on the fingerprint positioning algorithm and those without AP clustering are compared. The power of uplink and downlink users is set to 500mW, and 6 APs are set in each cluster. UAV is used to refer to users in a higher dimension. From the simulation diagram, it can be determined that the average uplink spectrum efficiency and the average downlink spectrum efficiency are higher without AP clustering. This is because all APs in the space can provide services to terminals in the space, while APs in non-AP clusters provide services to terminals in the corresponding terminal clusters, so that more APs can be used for data processing. However, it should be noted that when there are a large number of APs and user terminals in the scene, although the spectrum efficiency will be appropriately improved without AP clustering, the high fronthaul traffic overhead and power consumption will greatly increase the system burden as the number of APs increases. The technical solution of the embodiment of the present application can reduce the system computing burden while improving the spectrum efficiency.

[0223] See also Figure 5 , Figure 5 The downlink spectrum efficiency of the scheme with power allocation and the scheme without power allocation is compared. It is assumed that there are 6 APs, 8 uplink users and 8 downlink users in the scenario. UAV represents high-rise users whose height is higher than that of ordinary ground users, and Fixed represents the fixed transmission power of users, that is, the transmission power of uplink and downlink users is set to a power upper limit of 1W. It can be seen from the simulation diagram that by applying the DQN algorithm to optimize user power and AP node mode, the proportion of users with low achievable rates (<2 [b / s / Hz]) is greatly reduced, and the overall downlink spectrum efficiency is improved, which proves the effectiveness of the scheme of jointly optimizing user power and node mode using the DQN algorithm. Similar conclusions are drawn for uplink spectrum efficiency.

[0224] See also Figure 6 , Figure 6 is a structural diagram of a NAFD system parameter determination device provided in an embodiment of the present application, such as Figure 6 As shown, the NAFD system parameter determination device 200 includes:

[0225] A first determining module 201 is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0226] A second determining module 202 is used to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0227] The first obtaining module 203 is used to input the channel information, the information of the terminals in the at least one terminal cluster and the location information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0228] The first sending module 204 is configured to send the target system parameters to the APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

[0229] Optionally, the NAFD system parameter determination device further includes:

[0230] A second obtaining module, used to obtain the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0231] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0232] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0233] Optionally, the target system parameters include:

[0234] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0235] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0236] Optionally, the first determining module is specifically configured to:

[0237] Determining a three-dimensional position of each terminal in the first space;

[0238] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0239] Optionally, the first determining module is further configured to:

[0240] Acquire a received signal strength RSS vector of the first terminal, where the RSS vector includes multiple RSSs, where the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, and the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space;

[0241] Comparing the RSS vector of the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0242] In a case where the RSS vector of the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0243] Optionally, the first determining module is further configured to:

[0244] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0245] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0246] Optionally, the NAFD system parameter determination device further includes:

[0247] Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

[0248] Optionally, the first network is a deep Q network DQN.

[0249] It should be noted that the NAFD system parameter determination device provided in the embodiment of the present application can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated in this embodiment.

[0250] An embodiment of the present invention further provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, the various processes of the above-mentioned embodiment of the method for determining parameters of the NAFD system are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0251] For details, see Figure 7 As shown, an embodiment of the present invention further provides an auxiliary device, including a bus 301 , a transceiver 302 , an antenna 303 , a bus interface 304 , a processor 305 and a memory 306 .

[0252] The processor 305 is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, where the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster;

[0253] The processor 305 is further configured to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster;

[0254] The processor 305 is further configured to input the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0255] The transceiver 302 is used to send the target system parameters to the APs in the AP cluster, and the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

[0256] Optionally, the processor 305 is further configured to:

[0257] Obtaining the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network;

[0258] The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model;

[0259] The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

[0260] Optionally, the target system parameters include:

[0261] A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode;

[0262] The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

[0263] Optionally, the processor 305 is further configured to:

[0264] Determining a three-dimensional position of each terminal in the first space;

[0265] The terminals in the first space are divided according to the three-dimensional position of each terminal to determine at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

[0266] Optionally, the transceiver 302 is further used to obtain a received signal strength RSS vector corresponding to the first terminal, where the RSS vector includes multiple RSSs, and the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and the first terminal is any terminal in the first space;

[0267] The processor 305 is further configured to compare the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position;

[0268] When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

[0269] Optionally, the processor 305 is further configured to:

[0270] Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster;

[0271] When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

[0272] Optionally, the processor 305 is further configured to:

[0273] Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

[0274] Optionally, the first network is a deep Q network DQN.

[0275] exist Figure 7 In the embodiment, the bus architecture (represented by bus 301) is shown, and bus 301 may include any number of interconnected buses and bridges, and bus 301 links various circuits including one or more processors represented by processor 305 and memory represented by memory 306. Bus 301 may also link various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. Bus interface 304 provides an interface between bus 301 and transceiver 302. Transceiver 302 may be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. Data processed by processor 305 is transmitted on a wireless medium via antenna 303, and further, antenna 303 also receives data and transmits the data to processor 305.

[0276] The processor 305 is responsible for managing the bus 301 and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management and other control functions. The memory 306 can be used to store data used by the processor 305 when performing operations.

[0277] Optionally, the processor 305 may be a CPU, an ASIC, an FPGA or a CPLD.

[0278] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above-mentioned NAFD system parameter determination method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0279] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0280] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0281] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A network-assisted full-duplex NAFD system parameter determination method, applied to core network equipment, characterized in that: The method comprises: Divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster; Determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster; Inputting the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtaining the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network; The target system parameters are sent to APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to terminals in a corresponding terminal cluster according to the target system parameters.

2. The method according to claim 1, characterized in that After inputting the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, the method further includes: Obtaining the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network; The uplink cumulative spectrum efficiency is calculated based on the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and the first model; The downlink cumulative spectrum efficiency is calculated according to the target system parameters, the channel information, the information of the terminals in the at least one terminal cluster, the location information of the APs in the AP cluster, and a second model.

3. The method according to claim 2, characterized in that The target system parameters include: A mode of each AP in the AP cluster, wherein the mode of each AP includes an uplink service mode or a downlink service mode; The uplink AP in each AP cluster allocates power to each uplink terminal in the corresponding terminal cluster, and the downlink AP in each AP cluster allocates power to the downlink terminals in the corresponding terminal cluster. The uplink AP is the AP in the AP cluster that is in uplink service mode, and the downlink AP is the AP in the AP cluster that is in downlink service mode.

4. The method according to claim 1, characterized in that: The dividing the terminals in the first space to determine at least one terminal cluster includes: Determining a three-dimensional position of each terminal in the first space; The terminals in the first space are divided according to the three-dimensional position of each terminal to determine the at least one terminal cluster, wherein the terminals in the same terminal cluster in the at least one terminal cluster are in the same position range.

5. The method according to claim 4, characterized in that The determining the three-dimensional position of each terminal in the first space includes: Obtain a received signal strength RSS vector corresponding to the first terminal, where the RSS vector includes multiple RSSs, where the multiple RSSs are signal strengths of first reference signals respectively received by multiple APs in the first space, where the first reference signal is a reference signal sent by the first terminal to multiple APs in the first space, and where the first terminal is any terminal in the first space; Comparing the RSS vector corresponding to the first terminal with the RSS vector in the fingerprint library, each RSS vector in the fingerprint library being associated with a three-dimensional position; When the RSS vector corresponding to the first terminal matches the RSS vector in the fingerprint library, the three-dimensional position of the first terminal is determined based on the fingerprint library.

6. The method according to claim 1, characterized in that The determining the access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster includes: Determine a channel quality between each terminal in a first terminal cluster and each AP in the first space, the first terminal cluster being any terminal cluster in the at least one terminal cluster; When the channel quality between the second terminal and the first AP meets the preset conditions, it is determined that the first AP is an AP within the first AP cluster, the second terminal is any terminal within the first terminal cluster, the first AP cluster is the AP cluster corresponding to the first terminal cluster, and the first AP is the AP in the first space.

7. The method according to claim 1, characterized in that The method further comprises: Different time-frequency resources are allocated to each AP cluster, and the time-frequency resources are used by APs in each AP cluster to provide services to corresponding terminals according to corresponding time-frequency resources.

8. The method according to claim 1, characterized in that: The first network is a deep Q network DQN.

9. A network-assisted full-duplex NAFD system parameter determination device, characterized in that: The device comprises: A first determination module is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster; A second determining module is used to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster; A first obtaining module is used to input the channel information, the information of the terminals in the at least one terminal cluster and the position information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network; The first sending module is configured to send the target system parameters to the APs in the AP cluster, where the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

10. An electronic device, characterized in that: Including transceiver and processor, The processor is used to divide the terminals in the first space, determine at least one terminal cluster, and determine an access point AP cluster corresponding to each terminal cluster in the at least one terminal cluster, wherein the APs in the AP cluster are APs of the NAFD system, and the APs in the AP cluster are used to provide services to the terminals in the corresponding terminal cluster; The processor is further configured to determine channel information between each terminal in each terminal cluster and each AP in the corresponding AP cluster; The processor is further configured to input the channel information, the information of the terminals in the at least one terminal cluster, and the location information of the APs in the AP cluster into the trained first network, and obtain the target system parameters under the optimal balance of the uplink cumulative spectrum efficiency and the downlink cumulative spectrum efficiency output by the trained first network; The transceiver is used to send the target system parameters to the APs in the AP cluster, and the target system parameters are used by the APs in the AP cluster to provide services to the terminals in the corresponding terminal cluster according to the target system parameters.

11. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor; wherein the processor is used to read the program in the memory to implement the steps of the method for determining the parameters of the NAFD system as described in any one of claims 1 to 8.

12. A storage medium for storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for determining NAFD system parameters according to any one of claims 1 to 8 are implemented.

Images