Geographical location-based high-speed rail mobile communication 6G fully separated network uplink transmission method and system

Abstract

The present invention discloses a geographically-based, fully separated network-uplink transmission method and system for 6G high-speed rail mobile communications. The method includes the steps of: performing pure line-of-sight channel mapping based on the geographical locations of the train and each uplink base station, and using the mapping as channel estimation to perform joint receiver processing for secondary merging of uplink signals; designing a deep learning neural network to input the mapped pure line-of-sight channel and output multi-user joint precoding that satisfies constraints; offline training the network using historical channel data for each location; and inputting the pure line-of-sight channel corresponding to the train location during the actual deployment phase and outputting joint precoding at the corresponding location through the network, thereby realizing feedback-free transmission. Compared with traditional precoding design methods, the present invention effectively solves the problems of untimely channel feedback and large resource overhead caused by high-speed movement in high-speed rail scenarios, improves the spectral efficiency of wireless communications in high-speed rail scenarios, and provides an effective uplink transmission method for future high-speed rail mobile communications.

Description

[Technical Field]

[0001] The present invention relates to the technical field of wireless mobile communications, and to multi-user multiple-input multiple-output transmission in deep learning and high-speed railway scenarios, particularly to a geographically-based high-speed railway mobile communication 6G fully separated network uplink transmission method and system. [Background technology]

[0002] With the rapid development of railway systems, high-speed train communications has attracted widespread attention. Fifth-generation mobile networks (5G) achieve significant performance improvements through massive multiple-input multiple-output (mMIMO) with joint precoding design, maximizing diversity and multipath gain. However, high mobility causes channel degradation in various aspects, such as Doppler frequency offset (DFO), reduced coherence time, and phase noise, significantly reducing the performance of high-speed train communications systems. Furthermore, because 5G uses spectrum in higher frequency bands, handovers between base stations become more frequent in high-speed mobile scenarios.

[0003] Conventional solutions for Long Term Evolution Technology for Railways (LTE-R) are planned to use more frequent pilot transmission and channel state information (CSI) feedback, which leads to the consumption of large amounts of time-frequency resources. Furthermore, feedback delay becomes more severe in high-speed mobile scenarios, making it difficult to obtain accurate CSI even when more frequent pilots are used for channel estimation.

[0004] Sixth-generation (6G) mobile communication networks are expected to provide more flexible resource allocation and improve user experience. In their paper, "A Fully-Decoupled RAN Architecture for 6G Inspired by Neurotransmission," Academician Yu Quan and his colleagues proposed a 6G fully decoupled radio access network (FD-RAN) architecture that separates traditional base stations into uplink base stations (UBS), downlink base stations, and control base stations. FD-RAN offers significant advantages in high-speed rail communication systems. On the one hand, because all control plane signals are transmitted by the control base station, the amount of control information data is small. This allows the control base station to cover a wider area using lower frequency bands, significantly reducing handover overhead. On the other hand, FD-RAN employs a feedback-free transmission solution based on geographic location. This solution not only mitigates the negative effects of channel changes and resource overhead, but also avoids feedback delays through location prediction. Since the train track is fixed, the train's location can be easily predicted in advance. However, determining the appropriate precoding from location information alone remains challenging. Furthermore, considering multiple mobile relays (MRs) installed on trains, a joint precoding design is required because a multi-user MIMO (MU-MIMO) scenario must be utilized. Furthermore, it is difficult to use the real-time channel estimated by the pilots for signal reception at the base station side to save more pilot resources.

[0005] As described above, the problems with the existing technology are as follows: (1) Traditional pilot-based channel estimation incurs large resource overhead in high-speed railway scenarios, and feedback delays make it difficult for the transmitter to obtain real-time channel information. (2) In 6G fully isolated networks, the large feedback delays caused by the hardware isolation and isolation architecture significantly degrade the performance of transmission methods that rely on existing feedback mechanisms. Summary of the Invention [Problem to be solved by the invention]

[0006] Objective of the invention: The objective of the present invention is to provide a geographically-based high-speed railway mobile communication 6G fully separated network uplink transmission method and system, which can directly select precoding based on the geographical location of the train by adapting the mapping relationship between the geographical location of the train and precoding through deep learning, instead of the traditional channel estimation and channel feedback process. [Means for solving the problem]

[0007] In order to achieve the above object of the invention, the technical solution adopted in the present invention is as follows: A geographically-based high-speed rail mobile communication 6G fully separated network uplink transmission method: Step 1: Performing pure line-of-sight channel mapping based on the geographical locations of the train and each uplink base station, and using this as channel estimation to perform joint receiving processing for secondary merging of uplink signals; Step 2: designing a deep neural network that takes pure line-of-sight channel information mapped between the train and the uplink base station as input and outputs multi-user joint precoding information that satisfies a certain modulus constraint, designs a loss function of the deep neural network according to the system optimization goal, and performs data transmission using the precoding output by the deep neural network; Step 3: offline training a deep neural network using historical channel data for samples at each location and updating parameters until convergence; Step 4 includes inputting the line-of-sight channel corresponding to the train position in the actual deployment stage, and outputting joint precoding at the corresponding position by a deep neural network, thereby realizing feedback-free transmission.

[0008] Furthermore, in this scenario, we assume that there are L base stations serving K train mobile relays, and let H kl Then the pure line-of-sight channel mapping relationship is:

[0009]

number

number

[0010] Furthermore, the precoding vector of the k-th relay at the top of the train is

number

number

[0011] Furthermore, the step of performing joint receive processing for secondary merging of uplink signals includes:

[0012] The base station first performs first-order distributed processing on the received signal, and calculates the merge vector of the kth relay by the lth base station as

number

number

number

number

[0013] The edge cloud centrally performs the secondary merging of the signals that were first processed at the base station, and the signal at the kth relay after merging is

number

[0014] Furthermore, it supports maximum signal-to-interference and noise ratio.

number

number

number

[0015] Furthermore, by substituting, the maximum spectral efficiency that the kth relay can achieve is:

number

[0016] where τ p and τ c represent the length of the pilot and coherence time, respectively, and |·| represents the determinant of the matrix.

[0017] Furthermore, all user precoding vectors are denoted by P=[P1, ,P K ] and the system optimization objective is to maximize the total relay spectral efficiency, the system optimization problem can be modeled as follows:

[0018]

number

number

[0019] Furthermore, the designed deep learning neural network takes the channels between all relays and the base station as inputs and calculates the output α of the neural network before outputting. k teeth,

number

[0020] In specific implementation, the deep learning neural network may use a fully connected deep learning neural network or a residual network.

[0021] The present invention also provides a computer system including a memory, a processor, and a computer program stored in the memory and operable on the processor, the computer program being loaded into the processor to realize the steps of the geographically-based high-speed rail mobile communication 6G fully isolated network uplink transmission method. [Effects of the Invention]

[0022] The beneficial effects are as follows: Compared with the prior art, the present invention has the following advantages: 1. The present invention provides a high-speed railway uplink communication system under FD-RAN architecture, which avoids problems related to feedback delay and resource overhead in high-speed railway scenarios. 2. The present invention realizes multi-user multiple-input multiple-output precoding selection based on the train's geographic location, saving pilot estimation and feedback overhead and further improving system spectral efficiency. 3. The present invention uses deep learning to adapt the mapping relationship between the train's geographic location information and precoding, and uses pure line-of-sight channel mapping before input, effectively mitigating the low-latitude problem of geographic location information. [Brief explanation of the drawings]

[0023] [Figure 1] 1 is a schematic diagram of a scenario in which an embodiment of the present invention is applied; [Figure 2] FIG. 2 is a schematic diagram of a parameter update process for a neural network according to an embodiment of the present invention. [Figure 3] 1 is a schematic diagram of a neural network training convergence process according to an embodiment of the present invention; [Figure 4] FIG. 10 is a comparative curve diagram of spectral efficiency under various architectures according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] In order to make the objectives, technical solutions and advantages of the present invention clearer, the embodiments of the present invention will be described in detail below with reference to the drawings. This embodiment is implemented on the premise of the technical solutions of the present invention, and detailed embodiments and specific operation procedures are provided. It should be understood that the specific examples described herein are only used to explain the present invention, and the protection scope of the present invention is not limited to the following examples.

[0025] The geographical location-based high-speed rail mobile communication 6G fully separated network uplink transmission method disclosed in the embodiment of the present invention mainly includes the following steps:

[0026] Step 1: Perform pure line-of-sight channel mapping based on the geographical locations of the train and each uplink base station, and use this as channel estimation to perform joint receiving processing for secondary merging of uplink signals.

[0027] Step 2: Design a deep neural network that takes the pure line-of-sight channel information mapped between the train and the uplink base station as input and outputs multi-user joint precoding information that satisfies a certain modulus constraint. Design the loss function of the deep neural network according to the system optimization objective. Then, perform data transmission using the precoding output by the deep neural network.

[0028] Step 3: Using the historical channel data of samples at each location, train the deep neural network offline and update the parameters until convergence.

[0029] Step 4: In the actual deployment stage, the pure line-of-sight channel corresponding to the train position is input, and the deep neural network outputs the joint precoding at the corresponding position, thereby realizing feedback-free transmission.

[0030] The specific steps of the embodiment of the present invention will be described below with reference to the high-speed railway communication scenario shown in Figure 1. Assuming that the uplink base station and the mobile relay are located on the same plane and there are L base stations serving K train mobile relays, the actual channel between the mobile relay on top of the kth train and the lth uplink base station is denoted by H kl Then the pure line-of-sight channel mapping relationship is:

[0031]

number

number

[0032] The specific process of second-order joint reception of signals by the base station and the edge cloud is as follows:

[0033] The precoding vector of the kth relay is

number

number

number

number

number

[0034] The base station first performs first-order distributed processing on the received signal, and calculates the merge vector of the kth relay by the lth base station as

number

number

number

number

number

number

number

[0035] The edge cloud centrally performs the secondary merging of the signals that have been primarily processed at the base station, and calculates the secondary merging coefficient as a kl Then, the signal at the k-th relay after merging is expressed as follows:

[0036]

number

number

number

number

number

number

[0037]

number

[0038] Let all user precoding vectors be P=[P1, ,P K ] and considering the maximization of the sum of relay spectral efficiencies, the optimization problem of maximizing the sum of system spectral efficiencies may be modeled as follows:

[0039]

number

[0040] The deep learning neural network configuration employed will be described below. In an embodiment of the present invention, the deep neural network may use a fully connected deep learning neural network to avoid the labeling overhead of supervised learning, or may use a residual network (ResNet) to improve the convergence speed of training the deep neural network. The channels between all relays and the base station are taken as inputs. After being converted into vectors by a flattening layer, for example, the number of neurons passed through in sequence is set to 1024, 512, 256, and K*N. V Then, batch normalization is performed for each layer to prevent overfitting. A Lambda processing layer is designed before output, and the output of the neural network is

number

number

[0041] The network training process is shown in Figure 2. The process of minimizing the loss function finds the precoding that maximizes the total spectral efficiency of the system, i.e., the optimal solution to the optimization problem, thereby achieving the design of geographical location-based precoding without channel information feedback. The convergence curve during network training is shown in Figure 3, which indicates that the network can complete convergence within 20 iterations and requires few computing resources during the training process.

[0042] To make this example more intuitive and compare the advantages and disadvantages of the FD-RAN architecture with other architectures, Figure 4 shows a comparative schematic diagram of the simulation results for the spectral efficiency of high-speed train relays under different architectures. The results show that small cell performance is inferior to that of the FD-RAN architecture because individual signal reception cannot eliminate large inter-relay interference. Compared to cellular networks, higher spectral efficiency can be achieved when the train is close to the cellular base station. However, spectral efficiency is lower at the cellular edge. This demonstrates that the FD-RAN architecture in this example can provide stable and high spectral efficiency for high-speed train communications. In one embodiment of the present invention, a computer system is disclosed, which includes a memory, a processor, and a computer program stored in the memory and operable on the processor, the computer program being loaded into the processor to implement the steps of the 6G fully isolated network uplink transmission method for high-speed rail mobile communications based on geographical location information according to the aforementioned steps.

[0043] The above are only preferred embodiments of the present invention, and do not limit the present invention, and any modifications, equivalent replacements and improvements made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims

1. A geographically-based high-speed rail mobile communication 6G fully separated network uplink transmission method, comprising: Step 1: performing pure line-of-sight channel mapping based on the geographical locations of the train and each uplink base station, and using this as channel estimation to perform joint receiving processing for secondary merging of uplink signals; Step 2: designing a deep neural network that takes pure line-of-sight channel information mapped between the train and the uplink base station as input and outputs multi-user joint precoding information that satisfies a certain modulus constraint, designs a loss function of the deep neural network according to the system optimization goal, and performs data transmission using the precoding output by the deep neural network; Step 3: offline training a deep neural network using historical channel data for samples at each location and updating parameters until convergence; and step 4. inputting a pure line-of-sight channel corresponding to the train position in the actual deployment stage, and outputting joint precoding at the corresponding position through a deep neural network, thereby realizing feedback-free transmission.

2. The pure line-of-sight channel mapping relationship between the mobile relay on top of the kth train and the lth uplink base station is as follows: [Equation 41] where: [Equation 42] represents the phase offset between the antennas, and θ ｋｌ represents the angle between the mobile relay on top of the kth train and the lth uplink base station, f represents the carrier frequency, λ represents the carrier wavelength, v represents the train's running speed, c represents the speed of light, and d Ｂ and d Ｖ represent the antenna spacing between the base station and the relay, respectively, and N Ｂ and N Ｖ represents the number of antennas at the base station and relay, respectively, and β ｋｌ The method for geographically-based high-speed rail mobile communication 6G fully isolated network uplink transmission according to claim 1, wherein: is a fading coefficient.

3. The precoding vector of the k-th relay at the top of the train is [Equation 43] and the power constraint [0.0000] where N Ｖ represents the number of relay antennas, and p ｍａｘ The geographical location-based high-speed rail mobile communication 6G fully isolated network uplink transmission method according to claim 1, wherein: is the maximum power limit.

4. The step of performing joint receive processing for secondary merging of uplink signals includes: The base station first performs first-order distributed processing on the received signal, and calculates the merge vector of the kth relay by the lth base station as [Equation 45] The received signal at the lth base station is expressed as y ｌ and the k-th relay signal estimate is [Equation 46] and pure line-of-sight channel mapping [Equation 47] By the local minimum mean square error merging method based on [Number 48] and calculating The edge cloud centrally performs the secondary merging of the signals that were initially processed at the base station, and the signal of the kth relay after merging is [Number 49] All the secondary merging coefficients are A ｋ [a ｋｌ , ..., a ｋＬ ] Ｔ 2. The method for high-speed rail mobile communication 6G fully separated network uplink transmission based on geographical location according to claim 1, further comprising:

5. Supports maximum signal-to-interference and noise ratio [Number 50] Calculate where: [Equation 51] and H ｎｌ represents the actual channel between the mobile relay on top of the nth train and the lth uplink base station, [Number 52] and P ｎ represents the precoding vector of the n-th relay, and P ｋ represents the precoding vector of the k-th relay, and N ｋ represents the interference term between relays, and σ ２ represents the noise power, [Number 53] and H ｋｌ The geographical location-based high-speed rail mobile communication 6G fully separated network uplink transmission method according to claim 4, characterized in that k represents the actual channel between the mobile relay on top of the kth train and the lth uplink base station, and K represents the number of relays on the train.

6. The maximum spectral efficiency that the kth relay can achieve is [Number 54] where τ ｐ and τ ｃ represent the length of the pilot and the coherence time, respectively, |·| represents the determinant of the matrix, [Number 55] The geographical location-based high-speed rail mobile communication 6G fully separated network uplink transmission method according to claim 5, characterized in that:

7. All user precoding vectors are denoted by P = [P １ , ..., P Ｋ ] and the system optimization goal is to maximize the total relay spectrum efficiency, the system optimization problem can be modeled as follows: [Number 56] Here, SE ｋ represents the maximum spectral efficiency that the kth relay can achieve, and N Ｖ represents the number of antennas in the relay, K represents the number of relays in the train, and the loss function of the deep neural network is defined as the negative of the sum of the system spectral efficiencies, [Number 57] 2. The method for fully separated network uplink transmission for 6G high-speed rail mobile communication based on geographical location as claimed in claim 1, wherein N represents the number of samples.

8. The designed deep learning neural network takes the pure line-of-sight channels between all relays and the base station as input, and before outputting, it multiplies the output α ｋ teeth, [Number 58] 2. The method for fully separated network uplink transmission of 6G high-speed rail mobile communication based on geographical location according to claim 1, wherein the precoding vector is converted into a precoding vector that satisfies a certain modulus constraint by:

9. The method for fully separated network uplink transmission for 6G high-speed rail mobile communication based on geographical location according to claim 1, characterized in that the deep learning neural network adopts a fully connected deep learning neural network or a residual network.

10. A computer system including a memory, a processor, and a computer program stored in the memory and operable on the processor, The computer program is loaded into a processor to implement the steps of the geographically-based high-speed rail mobile communication 6G fully isolated network uplink transmission method according to any one of claims 1 to 9. A computer system.

Images