Scene-driven intelligent wireless communication system based on digital twinning

Abstract

The invention relates to the technical field of wireless communication, provides a scene-driven intelligent wireless communication system based on digital twinning, and defines an interactive interface of a system-level simulation layer, a link-level twinning layer and a physical transmission layer in order to solve the problems that an existing wireless communication system is fixed in transmission mode and solidified in performance boundary. Wireless communication is enabled and intelligent by means of the digital twin technology, the air interface transmission mode is flexibly and rapidly adjusted according to the communication environment or user service requirements under the condition that an actual physical wireless communication system is not affected, higher-quality communication services can be provided as required, and the utilization rate of air interface time-frequency resources is increased. According to the method, air interface transmission parameters can be dynamically adjusted according to cognition of a current communication scene, future user requirements and prediction of channel quality, time-frequency resource allocation is pre-optimized in a predictable manner, and simulation deduction and rapid verification are carried out by utilizing a twin environment. And finally, an optimal communication optimization strategy under the current communication condition or an air interface transmission pre-adjustment strategy under the future channel quality condition is obtained, and the flexibility and efficiency of air interface transmission are improved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a scene-driven intelligent wireless communication system based on digital twins. Background Art

[0002] In order to promote the digital transformation of the entire industry and support various business models, future wireless communication systems will face more diverse application scenarios and extreme performance challenges. For example, immersive multimedia interaction requires real-time interactive video experience, which puts higher requirements on the throughput and transmission delay of wireless communication systems; the automated production and transportation of equipment in smart industrial applications require extremely reliable and low-latency connections. In addition, compared with civilian cellular communications, military wireless communication scenarios are more prone to unpredictable changes and complex and diverse combat requirements. For example, reconnaissance drones require reliable transmission with large bandwidth, and the channel environment has little impact from multipath fading but large Doppler shift, while ground combat units require real-time voice services, and the channel environment is dominated by multipath fading of non-line-of-sight transmission. Therefore, future wireless communication systems require more flexible communication waveform configuration and resource scheduling to improve transmission effectiveness, reliability and resource utilization, and can meet the diverse needs of users through customized waveform services in different scenarios.

[0003] However, the traditional evolution of wireless communications has always been aimed at improving peak rates and system capacity. The air interface architecture of existing wireless communication systems is relatively simple, and most of them are general-purpose systems designed for a certain task. The transmission mode and parameter configurations adopted are often pre-set, resulting in communication capabilities being fixed within a limited range. Although the commercial 5GNR defines multiple transmission parameter sets, it does not implement flexible configuration of parameter sets, making it difficult to dynamically adjust the air interface parameter set according to changes in the communication environment. It has poor flexibility and is difficult to cope with complex and diverse communication environments and user business needs.

[0004] Therefore, it is necessary to design a scenario-driven flexible air interface wireless communication system to improve the flexibility and configurability of the air interface, be able to perceive the current link quality and communication scenario in real time, predict the user's business needs in the future, etc., and then flexibly and dynamically configure waveform transmission parameters and allocate time and frequency resources based on the cognitive results to adapt to the current scenario, thereby improving resource utilization and transmission efficiency, enhancing the user's service experience, and effectively responding to more abundant application scenarios in the future.

[0005] As an emerging technology, digital twin technology was first proposed by Professor Grieves of the University of Michigan and applied to the aerospace manufacturing industry. Later, it was gradually extended to the field of industrial manufacturing. After 2015, the concept of digital twins was widely recognized and entered a period of explosive application. From 2016 to 2018, Gartner, the world's most authoritative IT research and consulting company, listed data twin technology as one of the top ten strategic technology development trends for three consecutive years. In the past two years, digital twins have also been considered as one of the important applications of 6G. The 6G Vision White Paper released by ITU-R in June 2023 clearly defined digital twins as a network enhancement technology. By building a digital twin network, the self-optimization and self-evolution capabilities of the communication network can be improved, and new services can be derived.

[0006] Digital twin technology has the characteristics of bidirectional precise mapping, real-time, and full life cycle. It creates virtual entities of physical systems in a digital way. With the help of historical data, real-time data, and algorithm models, it can simulate, verify, predict, and control the full life cycle of physical entities. Therefore, with the help of digital twin technology, a high-fidelity virtual body of the wireless communication system can be created, providing a virtual environment for simulation and deduction for scenario-driven flexible air interface transmission. The "virtual domain" introduced by digital twins provides rich data and computing resources for cognitive algorithms, greatly increasing the freedom of physical entities in virtual space, and helping to realize the arrangement and deduction of a large number of transmission parameter combinations, thereby completing the simulation and effect verification of air interface transmission solutions in the virtual domain. Summary of the invention

[0007] In response to the problems of fixed transmission modes and solid performance boundaries in existing wireless communication systems, the present invention proposes a scenario-driven intelligent wireless communication system based on digital twins. It seeks to utilize the advantages of digital twin technology such as precise mapping and parallel deduction and simulation to improve the flexibility of wireless communication air interface transmission, and to achieve the ability to adaptively and dynamically adjust physical layer transmission parameters according to cognitive information of communication scenarios, generate communication optimization strategies, and improve resource utilization.

[0008] The technical solution adopted by the present invention is:

[0009] A scenario-driven intelligent wireless communication system based on digital twins, comprising: a system-level simulation layer, a link-level twin layer, and a physical transmission layer;

[0010] The system-level simulation layer is located at the top layer of the entire intelligent wireless communication system. It is used to initialize the system-level parameters, simulate the real cell environment according to the specified base station distribution method, generate the network topology structure, and then select the link between a base station and a user pair in a random or specified manner as agreed, and pass the system-level channel scenario information of the selected link to the physical transmission layer, where the system-level parameters include the number of cells and the number of base station transmitting and receiving antennas; it is also used to establish a service flow model according to different service characteristics, generate different types of service data to simulate real user services, and establish a user activity prediction model, predict the service or location status at future times according to the user's historical service or spatial location information, send the user service requirements and spatial location information corresponding to the selected link at the current and future times to the link-level twin layer, and send the user service requirement information at the current time to the physical transmission layer; it is also used to receive the link quality information uploaded by the link-level twin layer, and obtain the resource allocation effect between users with multiple heterogeneous service requirements through system-level simulation according to the service scenarios of multiple users within the topology range, thereby optimizing the multi-user time-frequency resource allocation method and informing the link-level twin layer of the time-frequency resources allocated to the selected link;

[0011] The link-level twin layer is located in the middle of the intelligent wireless communication system. It is used to receive the user service requirements and spatial location information of the current and future times corresponding to the selected link sent by the system-level simulation layer, as well as the parameter configuration information and channel status data of the transmitter and receiver uploaded by the physical transmission layer, to achieve a one-to-one twin mapping of the physical wireless transmission link to form a mirror simulator. The mirror simulator is used to perform link-level simulation to obtain the bit error rate of the wireless communication link under different signal-to-noise ratios, and the link quality information including the bit error rate and signal-to-noise ratio is uploaded to the system-level simulation layer; it is also used for intelligent scene recognition, using the stored historical channel status data to build a channel scene recognition model, and sending it to the physical transmission layer; it is also used to receive the scene recognized by the physical transmission layer The current channel scenario category is combined with the current and future user service requirements and spatial location information, as well as the channel quality prediction results, issued by the system-level simulation layer. The multi-modal parameter arrangement and communication optimization strategy are deduced in real time according to the time-frequency resource information corresponding to the selected link issued by the system-level simulation layer. Then, the link-level simulation and pre-verification are performed through the mirror simulator. Thus, through the iterative optimization within the link-level twin layer, an air interface transmission optimization strategy that is suitable for the current channel conditions and meets the user service requirements or matches the future channel quality to meet the user's future service requirements is obtained, and transmitted to the physical transmission layer. It is also used to receive the application effect of the optimization strategy fed back by the physical transmission layer. If the effect does not meet the set requirements, a new round of iterative optimization is performed.

[0012] The physical transmission layer is located at the bottom layer of the intelligent wireless communication system. It belongs to the actual point-to-point wireless communication link. It is an LTE, 5G base station-terminal or a semi-physical transceiver built based on software radio. It is used to receive the system-level channel scenario corresponding to the selected link sent by the system-level simulation layer, and initialize the channel category of the wireless channel simulator; it is also used to receive the channel scenario recognition model trained in the link-level twin layer for online reasoning, and pass the identified current channel scenario to the link-level twin layer; it is also used to receive the air interface transmission optimization strategy sent by the link-level twin layer, control the internal transmitter and receiver to adjust the corresponding parameters respectively, and feedback the application effect of the air interface optimization strategy on the physical link to the link-level twin layer according to the block error rate of the receiving end.

[0013] Furthermore, the system-level simulation layer includes a network topology generation module, a typical service flow modeling module, a user activity prediction module, and a system-level transmission resource allocation module;

[0014] Network topology generation module: used to establish the geographical locations of multiple base stations and users. After initializing the system-level parameters, the system user customizes the layout of walls and buildings to simulate the real cell environment, generate the network topology structure, simulate the business scenario of each user's location and pass it to the system-level transmission resource allocation module and the typical business flow modeling module. It is also used to select the link between a base station and a user pair from the network topology structure, and transmit the system-level channel scenario information corresponding to the selected link to the physical transmission layer; the system-level parameters include the number of cells and the number of base station transmitting and receiving antennas;

[0015] Typical business flow modeling module: used to establish business flow models. The actual business data parameters obey typical distribution. The distribution type obeyed by the business data is determined according to the business scenario of each user transmitted by the network topology generation module. The method of generating channel data by Jakes' Model is used to generate data that obeys the corresponding distribution and has temporal correlation to simulate the actual business. The generated business data will be passed to the user activity prediction module, and the user business demand at the current moment corresponding to the selected link will be passed to the link-level twin layer.

[0016] User activity prediction module: used to establish a user activity prediction model, receive business data transmitted by the typical business flow modeling module, use machine learning models to learn user business behaviors at multiple historical moments, predict business flow changes at multiple future moments, and combine neural network algorithms to predict user movement trajectories. The predicted user business needs and spatial location information at future moments are passed to the link-level twin layer;

[0017] System-level transmission resource allocation module: used for optimizing communication resources among multiple users with heterogeneous service requirements; based on the business scenario information of different users transmitted by the network topology generation module and combined with the link quality information uploaded by the link-level twin layer, the resource allocation algorithm is used to optimize resource allocation among multiple users with heterogeneous service requirements, and the effects of different algorithms are evaluated through system-level simulation to find the optimal resource allocation method, and the time-frequency resource information allocated to the selected link is informed to the link-level twin layer.

[0018] Furthermore, the link-level twin layer includes a data warehouse, a mirror simulator, and an intelligent decision maker;

[0019] Data warehouse: used to collect and store the configuration information of the physical transceiver, channel status data and the identified current channel scenario category collected from the physical transmission layer. As an intermediate station for data transmission, it provides data resources from the physical transmission layer for the construction of the mirror simulator and intelligent decision maker.

[0020] Mirror simulator: used to receive parameter configuration information, channel status data and the identified current channel scenario category of the physical transceiver transmitted from the data warehouse; based on the received parameter configuration information and channel status data of the physical transceiver, the "white box" method is used to simulate the baseband processing process of the transceiver signal, and the generative adversarial network method is used to establish a twin channel to achieve modeling of the physical channel; it is also used to receive the air interface transmission optimization strategy issued by the intelligent decision maker, simulate the air interface transmission optimization strategy and pre-verify the effect. If the effect does not meet the expectations, it will be fed back to the intelligent decision maker to re-arrange and deduce the multi-modal parameters, and the air interface transmission optimization strategy that meets the expectations will be directly passed to the physical transmission layer; after generating the air interface transmission optimization strategy, the mirror simulator passes the link quality information obtained from the link-level simulation results, that is, the corresponding relationship between the bit error rate and the signal-to-noise ratio, to the system-level simulation layer;

[0021] Intelligent decision maker: Receives channel status data from the physical receiver end collected by the data warehouse, completes the training and verification of channel scenario identification and channel quality prediction, and sends the trained channel scenario identification model to the physical transmission layer for online reasoning; it is also used to receive the current channel scenario category detected in real time by the physical transmission layer, and then combines the current and future user business needs and spatial location information sent by the system-level simulation layer, as well as the channel quality prediction results. According to the time-frequency resource information corresponding to the selected link sent by the system-level simulation layer, it arranges the link-level communication parameter combination, deduces and matches the air interface transmission optimization strategy that meets the user business needs under the current communication environment or meets the user business needs under future channel quality conditions, passes it to the mirror simulator for pre-verification, and then receives the effect information fed back by the mirror simulator for iterative optimization.

[0022] Further, the physical transmission layer includes a control module, a transmitter, a receiver and a wireless channel emulator;

[0023] Control module: used to receive the air interface transmission optimization strategy after iterative optimization sent by the link-level twin layer, send it to the transmitter and receiver respectively to adjust the transmission parameters, and receive the parameter configuration and channel status data of the transmitter and receiver, and pass it to the link-level twin layer;

[0024] Wireless channel emulator: used to receive the system-level channel scenario information corresponding to the selected link transmitted by the system-level simulation layer, and complete the initialization configuration of the wireless channel scenario;

[0025] Transmitter: includes the transmission signal baseband processing software and RF front end installed on the general computer. The transmitter is used for the baseband signal processing and RF transmission of the transmitter in the actual communication process.

[0026] Receiver: includes the RF front end, the receiving signal baseband processing software installed on the general computer, and the channel scene recognition module, which is used to receive the channel scene recognition model trained by the link-level twin layer, and pass the channel status data to the channel scene recognition module, identify the current channel scene type in real time, and pass the current channel scene type to the link-level twin layer through the control module.

[0027] The advantages of the present invention compared with the prior art are:

[0028] The present invention proposes a scenario-driven intelligent wireless communication system based on digital twins, which can dynamically adjust air interface transmission parameters based on the understanding of the current communication scenario and the prediction of future user needs and channel quality, and can predictably optimize the allocation of time and frequency resources in advance, and use the twin environment for simulation and rapid verification. Ultimately, the best communication optimization strategy under the current communication conditions is obtained, the flexibility and resource utilization of air interface transmission are improved, and on-demand services are provided to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a principle block diagram of a scene-driven intelligent wireless communication system based on digital twins in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The present invention proposes a scenario-driven intelligent wireless communication system based on digital twins, constructs a twin of end-to-end link-level wireless communication, defines the interactive interface of system-level simulation layer-link-level twin layer-physical transmission layer, and empowers intelligent wireless communication with the help of digital twin technology. Without affecting the actual physical system, it realizes flexible and rapid adjustment of air interface transmission mode according to the communication environment or user business needs, can provide better quality communication services on demand, and improve the utilization rate of air interface time and frequency resources.

[0031] like Figure 1 As shown, a scenario-driven intelligent wireless communication system based on digital twins includes a system-level simulation layer, a link-level twin layer, and a physical transmission layer;

[0032] The system-level simulation layer is located at the top layer of the entire intelligent wireless communication system. It is used to initialize the system-level parameters, simulate the real cell environment according to the specified base station distribution method, generate the network topology structure, and then select the link between a base station and a user pair in a random or specified manner as agreed, and pass the system-level channel scenario information of the selected link to the physical transmission layer, where the system-level parameters include the number of cells and the number of base station transmitting and receiving antennas; it is also used to establish a service flow model according to different service characteristics, generate different types of service data to simulate real user services, and establish a user activity prediction model, predict the service or location status at future times according to the user's historical service or spatial location information, send the user service requirements and spatial location information corresponding to the selected link at the current and future times to the link-level twin layer, and send the user service requirement information at the current time to the physical transmission layer; it is also used to receive the link quality information uploaded by the link-level twin layer, and obtain the resource allocation effect between users with multiple heterogeneous service requirements through system-level simulation according to the service scenarios of multiple users within the topology range, thereby optimizing the multi-user time-frequency resource allocation method and informing the link-level twin layer of the time-frequency resources allocated to the selected link;

[0033] The link-level twin layer is located in the middle of the intelligent wireless communication system. It is used to receive the user service requirements and spatial location information of the current and future times corresponding to the selected link sent by the system-level simulation layer, as well as the parameter configuration information and channel status data of the transmitter and receiver uploaded by the physical transmission layer, to achieve a one-to-one twin mapping of the physical wireless transmission link to form a mirror simulator. The mirror simulator is used to perform link-level simulation to obtain the bit error rate of the wireless communication link under different signal-to-noise ratios, and the link quality information including the bit error rate and signal-to-noise ratio is uploaded to the system-level simulation layer; it is also used for intelligent scene recognition, using the stored historical channel status data to build a channel scene recognition model, and sending it to the physical transmission layer; it is also used to receive the current channel scene type identified by the physical transmission layer. In particular, the channel quality information at future moments and the time-frequency resource block information corresponding to the selected link sent by the system-level simulation layer, the user's business needs and spatial location information at current and future moments are combined, and the air interface parameter arrangement and communication optimization strategy are deduced in real time according to the time-frequency resource information corresponding to the selected link sent by the system-level simulation layer, and then the link-level simulation and pre-verification are performed through the mirror simulator, so as to obtain the air interface transmission optimization strategy suitable for the current channel conditions to meet the user's current business needs or match the future channel quality and user business needs through the iterative optimization within the link-level twin layer, and transmit it to the physical transmission layer. It is also used to receive the application effect of the optimization strategy fed back by the physical transmission layer. If the effect does not meet the set requirements, a new round of iterative optimization is carried out;

[0034] The physical transmission layer is located at the bottom layer of the intelligent wireless communication system. It belongs to the actual point-to-point wireless communication link. It is an LTE, 5G base station-terminal or a semi-physical transceiver built based on software radio. It is used to receive the system-level channel scenario corresponding to the selected link sent by the system-level simulation layer, and initialize the channel category of the wireless channel simulator; it is also used to receive the channel scenario recognition model trained in the link-level twin layer for online reasoning, and pass the identified current channel scenario to the link-level twin layer; it is also used to receive the air interface transmission optimization strategy sent by the link-level twin layer, control the internal transmitter and receiver to adjust the corresponding parameters respectively, and feedback the application effect of the air interface optimization strategy on the physical link to the link-level twin layer according to the block error rate of the receiving end.

[0035] Among them, the system-level simulation layer includes a network topology generation module, a typical business flow modeling module, a user activity prediction module, and a system-level transmission resource allocation module;

[0036] Network topology generation module: used to establish the geographical locations of multiple base stations and users. After initializing the system-level parameters, the system user customizes the layout of walls and buildings to simulate the real cell environment, generate the network topology structure, simulate the business scenario of each user's location and pass it to the system-level transmission resource allocation module and the typical business flow modeling module. It is also used to select the link between a base station and a user pair from the network topology structure, and transmit the system-level channel scenario information corresponding to the selected link to the physical transmission layer; the system-level parameters include the number of cells and the number of base station transmitting and receiving antennas;

[0037] Typical business flow modeling module: used to establish business flow models. The actual business data parameters obey typical distribution. The distribution type obeyed by the business data is determined according to the business scenario of each user transmitted by the network topology generation module. The method of generating channel data by Jakes' Model is used to generate data that obeys the corresponding distribution and has temporal correlation to simulate the actual business. The generated business data will be passed to the user activity prediction module, and the user business demand at the current moment corresponding to the selected link will be passed to the link-level twin layer.

[0038] User activity prediction module: used to establish a user activity prediction model, receive business data transmitted by the typical business flow modeling module, use machine learning models to learn user business behaviors at multiple historical moments, predict business flow changes at multiple future moments, and combine neural network algorithms to predict user movement trajectories. The predicted user business needs and spatial location information at future moments are passed to the link-level twin layer;

[0039] System-level transmission resource allocation module: used for optimizing communication resources among multiple users with heterogeneous service requirements; based on the business scenario information of different users transmitted by the network topology generation module and combined with the link quality information uploaded by the link-level twin layer, the resource allocation algorithm is used to optimize resource allocation among multiple users with heterogeneous service requirements, and the effects of different algorithms are evaluated through system-level simulation to find the optimal resource allocation method, and the time-frequency resource information allocated to the selected link is informed to the link-level twin layer.

[0040] Among them, the link-level twin layer includes a data warehouse, a mirror simulator, and an intelligent decision maker;

[0041] Data warehouse: used to collect and store the configuration information of the physical transceiver, channel status data and the identified current channel scenario category collected from the physical transmission layer. As an intermediate station for data transmission, it provides data resources from the physical transmission layer for the construction of the mirror simulator and intelligent decision maker.

[0042] Mirror simulator: used to receive parameter configuration information, channel status data and the identified current channel scenario category of the physical transceiver transmitted from the data warehouse; based on the received parameter configuration information and channel status data of the physical transceiver, the "white box" method is used to simulate the baseband processing process of the transceiver signal, and the generative adversarial network method is used to establish a twin channel to achieve modeling of the physical channel; it is also used to receive the air interface transmission optimization strategy issued by the intelligent decision maker, simulate the air interface transmission optimization strategy and pre-verify the effect. If the effect does not meet the expectations, it will be fed back to the intelligent decision maker to re-arrange and deduce the multi-modal parameters, and the air interface transmission optimization strategy that meets the expectations will be directly passed to the physical transmission layer; after generating the air interface transmission optimization strategy, the mirror simulator passes the link quality information obtained from the link-level simulation results, that is, the corresponding relationship between the bit error rate and the signal-to-noise ratio, to the system-level simulation layer;

[0043] Intelligent decision maker: Receives channel status data from the physical receiver end collected by the data warehouse, completes the training and verification of channel scenario identification and channel quality prediction, and sends the trained channel scenario identification model to the physical transmission layer for online reasoning; it is also used to receive the current channel scenario category detected in real time by the physical transmission layer, and then combines the current and future user business needs and spatial location information sent by the system-level simulation layer, as well as the channel quality prediction results, to arrange the link-level communication parameter combination according to the time-frequency resource information corresponding to the selected link sent by the system-level simulation layer, deduce and match the air interface transmission optimization strategy that meets the user business needs under the current communication environment or meets the user business needs under future channel quality conditions, pass it to the mirror simulator for pre-verification, and then receive the effect information fed back by the mirror simulator for iterative optimization.

[0044] Among them, the physical transmission layer includes a control module, a transmitter, a receiver and a wireless channel emulator;

[0045] Control module: used to receive the air interface transmission optimization strategy after iterative optimization sent by the link-level twin layer, send it to the transmitter and receiver respectively to adjust the transmission parameters, and receive the parameter configuration and channel status data of the transmitter and receiver, and pass it to the link-level twin layer;

[0046] Wireless channel emulator: used to receive the system-level channel scenario information corresponding to the selected link transmitted by the system-level simulation layer, and complete the initialization configuration of the wireless channel scenario;

[0047] Transmitter: includes the transmission signal baseband processing software and RF front end installed on the general computer. The transmitter is used for the baseband signal processing and RF transmission of the transmitter in the actual communication process.

[0048] Receiver: includes the RF front end, the receiving signal baseband processing software installed on the general computer, and the channel scene recognition module, which is used to receive the channel scene recognition model trained by the link-level twin layer, and pass the channel status data to the channel scene recognition module, identify the current channel scene type in real time, and pass the current channel scene type to the link-level twin layer through the control module.

[0049] The internal and external interfaces are defined as follows:

[0050] 1. The system-level simulation layer, link-level twin layer, and physical transmission layer interact with each other to complete the external interface of the outer loop:

[0051] Interface 1-1: The link-level twin layer collects the parameter configuration information and status data of the transmitter and receiver uploaded by the physical transmission layer in real time, as well as the channel scenario type identified in real time;

[0052] Interface 1-2: The link-level twin layer sends the trained channel scenario recognition model to the physical transport layer for online reasoning;

[0053] Interface 1-3: The link-level twin layer sends the air interface transmission optimization strategy verified by the mirror simulator to the physical transmission layer;

[0054] Interface 1-4: The link-level twin layer uploads the calculated link quality information to the system-level simulation layer for system-level performance simulation;

[0055] Interface 1-5: The system-level simulation layer selects the link between a base station and a user pair as the object of link-level mirror simulation based on the generated network topology and large-scale channel fading, and sends the user service requirements and spatial location information corresponding to the link at the current and future times to the link-level twin layer, and informs the link-level twin layer of the result of system-level transmission resource allocation, that is, the time-frequency resource information corresponding to the selected link.

[0056] Interface 1-6: The system-level simulation layer selects one of the base station-user links according to the generated network topology, passes the system-level channel scenario information corresponding to the link to the wireless channel emulator in the physical transmission layer, completes the initialization configuration of the channel emulator, and informs the physical transmission layer of the system-level channel scenario information of the selected link.

[0057] 2. The internal interface of the inner loop is completed by the interaction between different modules within the link-level twin layer:

[0058] Interface 2-1: The data warehouse provides the collected parameter configuration and status information of the physical transmission layer to the mirror simulator for modeling and simulation of the twin, and transmits the channel scenario identified by the physical receiver;

[0059] Interface 2-2: The data warehouse provides the collected channel state information to the intelligent decision maker for channel quality prediction, multi-modal parameter arrangement and deduction;

[0060] Interface 2-3: The multi-modal parameter arrangement and deduction module sends the deduced optimization strategy of the air interface transmission parameter configuration to the mirror simulator for pre-verification;

[0061] Interface 2-4: The simulation control module will control the mirror simulator to simulate and test the air interface transmission optimization strategy obtained by the multimodal parameter arrangement and deduction module, and simulate and verify its application effect in the real-time physical transmission layer. If it does not meet the standards, it will feedback to the intelligent decision maker and trigger the multimodal module to re-arrange and deduce.

[0062] Ways and means to implement the present invention:

[0063] Example 1: Adaptive optimization of the designed system for harsh environments. The scenario of the wireless channel emulator is changed from a suburban type to a harsh urban macrocell. Assuming that the user service demand remains unchanged, the physical transmission layer detects that the block error rate has deteriorated. First, the receiver end of the physical transmission layer identifies online that the current channel scenario has changed to a harsh urban macrocell, and uploads the information to the link-level twin layer through the control module via interface 1-1. At the same time, the link-level twin layer receives the user service demand information sent down from the upper layer, and the data warehouse distributes the current channel state information to the mirror simulator and the intelligent decision maker via interfaces 2-1 and 2-2 respectively. The mirror simulator realizes the twin modeling of the actual physical channel and establishes a one-to-one mapping link-level twin system. The multimodal parameter orchestration and deduction module in the intelligent decision maker deduce the physical layer parameter configuration and transmission optimization strategy suitable for the current scenario and in line with user needs, and passes the parameter adjustment strategy to the mirror simulator through 2-3 for real-time pre-verification. If the system's block error rate is improved to meet the preset requirements, the strategy is passed and sent to the physical transmission layer for execution; if it does not meet the preset requirements, it is fed back to the multimodal module through 2-4 to continue deduction, and through a limited number of iterations until a suitable optimization strategy is found to reduce the block error rate of the physical system, and finally complete the air interface adaptive optimization of the wireless communication system when the communication environment deteriorates.

[0064] Example 2: Adaptive optimization of the designed system for a good environment. The scenario of the wireless channel emulator is changed from a mountainous area to a suburban area. Assuming that the user service scenario is a high-traffic service with a service data volume of 10MB and a maximum transmission delay limited to 20s, the physical transmission layer detects that the throughput rate has increased but is still lower than the user's requirement of 4Mbps. The receiver end of the physical transmission layer online identifies that the current channel scenario has changed to the suburbs, and uploads the information to the link-level twin layer through the control module through interface 1-1. At the same time, the link-level twin layer receives the user service demand information sent down by the upper layer. The data warehouse distributes the current channel state information to the mirror simulator and the intelligent decision maker through interfaces 2-1 and 2-2 respectively. The mirror simulator implements twin modeling of the actual physical channel and establishes a one-to-one mapping link-level twin system. The multimodal parameter arrangement and deduction module in the intelligent decision maker deduce the physical layer parameter configuration and transmission optimization strategy that is suitable for the current scenario and can meet user needs, and pass the parameter adjustment strategy to the mirror simulator through 2-3 for real-time pre-verification. If the system throughput is improved and meets user needs, the strategy is passed and sent to the physical transmission layer for execution; if it does not meet the preset requirements, it is fed back to the multimodal module through 2-4 to continue deduction, and a limited number of iterations are made until a suitable optimization strategy is found to achieve an increase in the physical system throughput, and finally complete the air interface adaptive optimization of the wireless communication system when the communication environment improves.

[0065] Embodiment 3: Adaptive optimization of the designed system to changes in user service requirements. Assuming that the channel scenario remains unchanged, the link-level twin layer collects and obtains the current channel state information estimated by the physical transmission layer in real time. The user activity prediction module in the system-level simulation layer predicts the size of the user's service group at the next moment, and sends the result to the link-level twin layer. The multimodal parameter arrangement and deduction module in the intelligent decision maker combines the current channel conditions to deduce the physical layer parameter configuration and transmission optimization strategy that can meet the user's service requirements at the next moment, and passes the strategy to the mirror simulator for pre-verification through 2-3. If the user's service volume demand at the next moment increases, then verify whether the transmission rate index of the mirror simulator increases; if the user's service volume demand at the next moment decreases, then verify whether the block error rate index of the mirror simulator decreases, which can meet the user's higher requirements for communication reliability. The pre-verified strategy will be sent to the physical transmission layer for execution. Otherwise, if the communication optimization strategy fails the pre-verification, it needs to be fed back to the multimodal module for iterative deduction, and finally complete the air interface adaptive optimization of the wireless communication system to changes in user service requirements.

Claims

1. A scenario-driven intelligent wireless communication system based on digital twins, characterized in that: include: System-level simulation layer, link-level twin layer, and physical transmission layer; The system-level simulation layer is located at the top layer of the entire intelligent wireless communication system. It is used to initialize the system-level parameters, simulate the real cell environment according to the specified base station distribution method, generate the network topology structure, and then select the link between a base station and a user pair in a random or specified manner as agreed, and pass the system-level channel scenario information of the selected link to the physical transmission layer. The system-level parameters include the number of cells and the number of base station transmitting and receiving antennas; it is also used to establish a business flow model according to different business characteristics, generate different types of business data to simulate real user business, and establish a user activity prediction model. According to the user's historical business or spatial location information, the business or location status at future times is predicted, and the user business requirements and spatial location information corresponding to the selected link at current and future times are sent to the link-level twin layer; it is also used to receive the link quality information uploaded by the link-level twin layer, and according to the business scenarios of multiple users within the topology range, the resource allocation effect between users with multiple heterogeneous service requirements is obtained through system-level simulation, thereby optimizing the multi-user time-frequency resource allocation method, and informing the link-level twin layer of the time-frequency resources allocated to the selected link; The link-level twin layer is located in the middle of the intelligent wireless communication system. It is used to receive the user service requirements and spatial location information of the current and future times corresponding to the selected link sent by the system-level simulation layer, as well as the parameter configuration information and channel status data of the transmitter and receiver uploaded by the physical transmission layer, to achieve a one-to-one twin mapping of the physical wireless transmission link to form a mirror simulator. The mirror simulator is used to perform link-level simulation to obtain the bit error rate of the wireless communication link under different signal-to-noise ratios, and the link quality information including the bit error rate and signal-to-noise ratio is uploaded to the system-level simulation layer; it is also used for intelligent scene recognition, using the stored historical channel status data to build a channel scene recognition model, and sending it to the physical transmission layer; it is also used to receive the scene recognized by the physical transmission layer The current channel scenario category is combined with the current and future user service requirements and spatial location information, as well as the channel quality prediction results, issued by the system-level simulation layer. The multi-modal parameter arrangement and communication optimization strategy are deduced in real time according to the time-frequency resource information corresponding to the selected link issued by the system-level simulation layer. Then, the link-level simulation and pre-verification are performed through the mirror simulator. Thus, through the iterative optimization within the link-level twin layer, an air interface transmission optimization strategy that is suitable for the current channel conditions and meets the user service requirements or matches the future channel quality to meet the user's future service requirements is obtained, and transmitted to the physical transmission layer. It is also used to receive the application effect of the optimization strategy fed back by the physical transmission layer. If the effect does not meet the set requirements, a new round of iterative optimization is performed. The physical transmission layer is located at the bottom layer of the intelligent wireless communication system. It belongs to the actual point-to-point wireless communication link. It is an LTE, 5G base station-terminal or a semi-physical transceiver built based on software radio. It is used to receive the system-level channel scenario corresponding to the selected link sent by the system-level simulation layer, and initialize the channel category of the wireless channel simulator; it is also used to receive the channel scenario recognition model trained in the link-level twin layer for online reasoning, and pass the identified current channel scenario to the link-level twin layer; it is also used to receive the air interface transmission optimization strategy sent by the link-level twin layer, control the internal transmitter and receiver to adjust the corresponding parameters respectively, and feedback the application effect of the air interface optimization strategy on the physical link to the link-level twin layer according to the block error rate of the receiving end.

2. According to claim 1, a scenario-driven intelligent wireless communication system based on digital twins is characterized in that: The system-level simulation layer includes a network topology generation module, a typical service flow modeling module, a user activity prediction module, and a system-level transmission resource allocation module; Network topology generation module: used to establish the geographical locations of multiple base stations and users. After initializing the system-level parameters, the system user customizes the layout of walls and buildings to simulate the real cell environment, generate the network topology structure, simulate the business scenario of each user's location and pass it to the system-level transmission resource allocation module and the typical business flow modeling module. It is also used to select the link between a base station and a user pair from the network topology structure, and transmit the system-level channel scenario information corresponding to the selected link to the physical transmission layer; The system-level parameters include the number of cells and the number of base station transmit and receive antennas; Typical business flow modeling module: used to establish business flow models. The actual business data parameters obey typical distribution. The distribution type obeyed by the business data is determined according to the business scenario of each user transmitted by the network topology generation module. The method of generating channel data by Jakes' Model is used to generate data that obeys the corresponding distribution and has temporal correlation to simulate the actual business. The generated business data will be passed to the user activity prediction module, and the user business demand at the current moment corresponding to the selected link will be passed to the link-level twin layer. User activity prediction module: used to establish a user activity prediction model, receive business data transmitted by the typical business flow modeling module, use machine learning models to learn user business behaviors at multiple historical moments, predict business flow changes at multiple future moments, and combine neural network algorithms to predict user movement trajectories. The predicted user business needs and spatial location information at future moments are passed to the link-level twin layer; System-level transmission resource allocation module: used for optimizing communication resources among multiple users with heterogeneous service requirements; According to the business scenario information of different users transmitted by the network topology generation module and combined with the link quality information uploaded by the link-level twin layer, the resource allocation algorithm is used to optimize the resource allocation among users with various heterogeneous service requirements. The effects of different algorithms are evaluated through system-level simulation to find the optimal resource allocation method, and the time-frequency resource information allocated to the selected link is informed to the link-level twin layer.

3. The scenario-driven intelligent wireless communication system based on digital twin according to claim 1, characterized in that: The link-level twin layer includes a data warehouse, a mirror simulator, and an intelligent decision maker; Data warehouse: used to collect and store the configuration information of the physical transceiver, channel status data and the identified current channel scenario category collected from the physical transmission layer. As an intermediate station for data transmission, it provides data resources from the physical transmission layer for the construction of the mirror simulator and intelligent decision maker. Mirror simulator: used to receive the parameter configuration information, channel status data and the identified current channel scenario category of the physical transceiver transmitted from the data warehouse; based on the received parameter configuration information and channel status data of the physical transceiver, the "white box" method is used to simulate the baseband processing process of the transceiver signal, and the twin channel is established by the generative adversarial network method to achieve modeling of the physical channel; it is also used to receive the air interface transmission optimization strategy issued by the intelligent decision maker, simulate and pre-verify the air interface transmission optimization strategy, and if the effect does not meet the expectations, it will be fed back to the intelligent decision maker to re-arrange and deduce the multi-modal parameters, and the air interface transmission optimization strategy that meets the expectations will be directly passed to the physical transmission layer; after generating the air interface transmission optimization strategy, the mirror simulator passes the link quality information obtained from the link-level simulation results, that is, the corresponding relationship between the bit error rate and the signal-to-noise ratio, to the system-level simulation layer; Intelligent decision maker: Receives channel status data from the physical receiver end collected by the data warehouse, completes the training and verification of channel scenario identification and channel quality prediction, and sends the trained channel scenario identification model to the physical transmission layer for online reasoning; it is also used to receive the current channel scenario category detected in real time by the physical transmission layer, and then combines the current and future user business needs and spatial location information sent by the system-level simulation layer, as well as the channel quality prediction results, to arrange the link-level communication parameter combination according to the time-frequency resource information corresponding to the selected link sent by the system-level simulation layer, deduce and match the air interface transmission optimization strategy that meets the user business needs under the current communication environment or meets the user business needs under future channel quality conditions, pass it to the mirror simulator for pre-verification, and then receive the effect information fed back by the mirror simulator for iterative optimization.

4. The scenario-driven intelligent wireless communication system based on digital twin according to claim 1, characterized in that: The physical transmission layer includes a control module, a transmitter, a receiver and a wireless channel emulator; Control module: used to receive the air interface transmission optimization strategy after iterative optimization sent by the link-level twin layer, send it to the transmitter and receiver respectively to adjust the transmission parameters, and receive the parameter configuration and channel status data of the transmitter and receiver, and pass it to the link-level twin layer; Wireless channel emulator: used to receive the system-level channel scenario information corresponding to the selected link transmitted by the system-level simulation layer, and complete the initialization configuration of the wireless channel scenario; Transmitter: includes the transmission signal baseband processing software and RF front end installed on the general computer. The transmitter is used for the baseband signal processing and RF transmission of the transmitter in the actual communication process. Receiver: It includes a radio frequency front end, a receiving signal baseband processing software installed on a general computer, and a channel scene recognition module, which is used to receive the channel scene recognition model trained by the link-level twin layer, and pass the channel status data to the channel scene recognition module, identify the current channel scene type in real time, and pass the current channel scene type to the link-level twin layer through the control module.

Images