Low earth orbit satellite network digital twin drive switching optimization method and system

Abstract

The invention discloses a low earth orbit satellite network digital twin drive switching optimization method and system, and relates to the technical field of inter-satellite communication, and the method comprises the steps: building a digital twin system which is synchronously evolved with a physical LEO satellite network; predicting the link quality between each candidate satellite and the user terminal in a future time window based on a digital twin system; dividing a time window into a plurality of continuous time slices, constructing a directed weighted dynamic graph model which takes a satellite-time slice as a node and takes link comprehensive utility cost as an edge weight, and searching an optimal path in the directed weighted dynamic graph model by utilizing a shortest path algorithm so as to determine an optimal switching opportunity and a target satellite sequence; the optimal switching strategy is issued to a user terminal to be executed, and a link quality prediction result is corrected in real time through a feedback mechanism of the digital twin system to form closed-loop optimization; the method can effectively meet differentiated service quality requirements in a multi-service scene, and is suitable for a high-dynamic LEO satellite communication system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of inter-satellite communication, and particularly relates to a low-orbit satellite network digital twin driven switching optimization method and system. BACKGROUND

[0002] Low Earth Orbit (LEO) satellite networks have become a key foundation for supporting diversified services such as emergency communication, unmanned driving, and remote medical treatment. Specifically, in order to ensure task safety, real-time control type services require extremely low propagation delay and packet loss rate; in order to ensure efficient transmission, data-intensive services rely more on stable data rate and bandwidth.

[0003] Due to the high speed of LEO satellites in orbit and the short visible window, ground user terminals must complete rapid link switching within a very short time to maintain the continuity of communication; at the same time, the diversified demand for Quality of Service (QoS) in different application scenarios presents a high degree of dynamicity and unbalance. Therefore, it has become an important topic to carry out diversified service demand and satellite switching optimization for high dynamic topology.

[0004] Existing researches have made certain progress in user preference perception, service continuity and multi-attribute decision-making in LEO network switching strategies, but they mostly rely on static modeling, which is difficult to cope with the dynamic evolution and differentiated QoS demand of LEO networks. Dynamic graph model can depict the time-varying topology of LEO network and jointly optimize the switching target and timing, thereby improving the stability and response accuracy of switching. At the same time, the fusion of digital twin technology to enhance state perception and strategy prediction ability has become the focus of recent research. Although emerging technologies including large models have made significant progress in generalization modeling, they still have problems in adapting to the deployment cost, response efficiency and other aspects in the face of the stringent demand for low delay and high frequency switching in LEO networks.

[0005] At present, when modeling the LEO network switching model to optimize the switching path, the influence of the above factors on meeting differentiated QoS demand is often ignored, resulting in that the satellite switching optimization path is not accurate and comprehensive enough. SUMMARY

[0006] To address the shortcomings of existing technologies that rely heavily on static modeling and are ill-suited to the dynamic evolution and differentiated QoS requirements of LEO networks, this invention proposes a digital twin-driven handover optimization method, system, and device for low-Earth orbit satellite networks. This method utilizes a digital twin system to perceive the physical network status in real time, constructs a directed weighted graph model incorporating a time dimension, employs a shortest path algorithm to search for the optimal handover sequence, and introduces a QoS-aware dynamic weight adjustment mechanism. It adaptively optimizes the weights of various indicators based on service type to generate personalized handover targets, thereby resolving the problems inherent in existing technologies.

[0007] A digital twin-driven handover optimization method for low-Earth orbit satellite networks includes the following steps: Real-time acquisition of satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network; establishment of a digital twin system that evolves synchronously with the physical LEO satellite network. Based on a digital twin system, the link quality between each candidate satellite and the user terminal within a future time window is predicted. The time window is divided into multiple consecutive time slices, and a directed weighted dynamic graph model is constructed with satellite-time slice as nodes and link comprehensive utility cost as edge weights. The link comprehensive utility cost is obtained by normalizing and weighting the data rate, propagation delay, and packet loss rate indicators. Based on the user's current service type, dynamically configure the weight coefficients of various QoS indicators in the comprehensive utility cost of the link; based on the configured weight coefficients, use the shortest path algorithm to search for the optimal path from the virtual start node to the virtual end node in the directed weighted dynamic graph model, so as to determine the optimal handover timing and target satellite sequence; The optimal switching strategy is sent to the user terminal for execution, and the link quality prediction results are corrected in real time through the feedback mechanism of the digital twin system, forming a closed-loop optimization.

[0008] Furthermore, the utility function calculation process in the directed weighted dynamic graph model is expressed as follows: ; in, R sj-i , T sj-i and P sj-i They represent time respectively i Time Satellite s j Data rate, propagation latency, and packet loss rate; ω 1 , ω 2 , ω 3 The weighting coefficients for the three types of indicators, and satisfying the following conditions: ω 1+ ω 2 + ω 3 =1; R min , R max , T min , T max , P min , P max These represent the minimum and maximum values ​​of the corresponding indicators among the candidate switching targets.

[0009] Furthermore, the calculation processes for the data rate, propagation delay, and packet loss rate indicators are respectively expressed as follows: Calculate the data rate based on Shannon's capacity theorem. R : ;in, B For channel bandwidth, P r and N These are the received signal power and the noise power, respectively. The calculation process for propagation delay is expressed as follows: ;in, V The speed of light; L ( h , h +1) is a satellite h and h Link distance between +1; T It is the average delay of all handover decisions in the satellite network, taking all satellites into account, and is expressed in milliseconds. H Indicates the overall switching decision; The calculation process for packet loss rate is expressed as follows: ;in, P Packet loss rate; P lost The number of lost data packets, that is, the total number of packets that failed to reach their destination during transmission; P sent The total number of data packets sent, i.e., the total number of data packets sent by the source node.

[0010] Furthermore, it also includes dynamically adjusting the weighting coefficients of the data rate, propagation latency, and packet loss rate in the overall utility cost of the link based on the user's service type, specifically including the following steps: For real-time control applications, increase the weighting of propagation latency and packet loss rate. ω 2 , ω 3 Reduce the weight of data rateω 1 ; For data-intensive businesses, increase the weight given to data rate. ω 1 Reduce the weight of propagation delay and packet loss rate ω 2 , ω 3 .

[0011] Furthermore, the weighting coefficient is dynamically penalized or enhanced based on whether the service quality threshold is met.

[0012] Furthermore, in the directed weighted dynamic graph model: The horizontal edges represent continuous service from the same satellite within adjacent time slices; The diagonal edges represent the switching connections between different satellites with overlapping coverage; Set up a virtual start node to connect all candidate satellite instances in the first time slice, and set up a virtual end node to connect all candidate satellite instances in the last time slice.

[0013] Furthermore, the shortest path algorithm is Dijkstra's algorithm.

[0014] Furthermore, the length of the prediction time window is configurable, and a time relaxation factor is introduced, allowing the actual switching time to be flexibly adjusted within the prediction time slice range.

[0015] Furthermore, the digital twin system continuously corrects the link quality prediction results by interacting with the physical LEO satellite network in real time to achieve a closed-loop optimization mechanism of "perception-prediction-decision-execution-feedback".

[0016] The present invention also includes a digital twin-driven handover optimization system for low-Earth orbit satellite networks, comprising: The digital model building module is used to collect satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network in real time, and to establish a digital twin system that evolves synchronously with the physical LEO satellite network. A dynamic graph construction module is used to predict the link quality between candidate satellites and user terminals within a future time window based on a digital twin system. The time window is divided into multiple consecutive time slices, and a directed weighted dynamic graph model is constructed with satellite-time slices as nodes and link comprehensive utility cost as edge weights. The link comprehensive utility cost is obtained by normalizing and weighting the data rate, propagation delay, and packet loss rate indicators. The search module is used to dynamically configure the weight coefficients of various QoS indicators in the comprehensive utility cost of the link according to the user's current service type; based on the configured weight coefficients, the shortest path algorithm is used to search for the optimal path from the virtual start node to the virtual end node in the directed weighted dynamic graph model, so as to determine the optimal handover time and target satellite sequence. The optimization module is used to send the optimal handover strategy to the user terminal for execution, and to correct the link quality prediction results in real time through the feedback mechanism of the digital twin system, forming a closed-loop optimization.

[0017] This invention provides a digital twin-driven handover optimization method for low-Earth orbit satellite networks, which has the following beneficial effects: This invention constructs a dynamic graph model based on digital twins, jointly optimizing handover targets and handover timing to achieve real-time mapping and evolution prediction of the highly dynamic topology of low Earth orbit (LEO) satellite networks. By using a digital twin system to perceive the physical network status in real time, it predicts satellite orbits, link quality, and coverage relationships within future time windows, constructing a directed weighted graph model with a time dimension. Nodes represent "satellite-time slice" instances, and edge weights are dynamically weighted from multi-dimensional Quality of Service (QoS) indicators such as data rate, propagation delay, and packet loss rate. A shortest path algorithm is used to search for the optimal handover sequence, and a QoS-aware dynamic weight adjustment mechanism is introduced to adaptively optimize the weights of each indicator based on the service type, generating personalized handover targets and improving the accuracy and comprehensiveness of satellite handover optimization paths. This invention significantly reduces average latency and packet loss rate, improves data rate and service continuity without increasing additional resource overhead, effectively addressing differentiated QoS requirements in multi-service scenarios, and is suitable for highly dynamic LEO satellite communication systems. Attached Figure Description

[0018] Figure 1 This is an architecture diagram of a digital twin-driven handover optimization method for low-Earth orbit satellite networks oriented towards multiple service scenarios, as described in an embodiment of the present invention. Figure 2 This is a schematic flowchart illustrating a digital twin-driven handover optimization method for low-Earth orbit satellite networks oriented towards multiple service scenarios, as described in an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] like Figure 1As shown, digital twins, by sensing the real-time status of the physical satellite network, collect and synchronize multi-source data such as user location, satellite orbit, and link quality. After mapping through a virtual satellite network, a dynamic model reflecting the network's operational status is constructed. Based on this, digital twins provide network status prediction capabilities, dynamically injecting multi-temporal topology and multi-dimensional QoS information into the dynamic graph model, forming a directed weighted graph structure that evolves synchronously with the physical network. In this structure, horizontal edges represent continuous service from the same satellite in adjacent time slices (i.e., no handover required), while diagonal edges represent handover connections between different satellite instances with overlapping coverage. The weight of each directed edge is obtained by normalizing multi-dimensional QoS indicators (such as data rate, propagation delay, and packet loss rate) and then weighting them by a weighted sum. Through this method, combined with the predictive-feedback closed-loop mechanism of digital twins, joint optimal path planning for "handover target-handover timing" is achieved in the LEO network.

[0021] Satellite handover is affected by a variety of factors in practical applications, among which data rate, propagation delay and packet loss rate are the most significant factors. (1) Data rate refers to the amount of data transmitted through the communication channel per unit time, which is used to measure the data transmission capacity of the link. Its magnitude directly reflects the bandwidth utilization efficiency and information transmission speed of the communication system, and is the core indicator for evaluating communication quality and network performance. In LEO satellite networks, data rate not only determines the service experience that user terminals can obtain, but also affects the selection of handover strategies and the allocation of network resources, which is of great significance for ensuring service continuity and meeting diverse service needs. (2) Propagation delay refers to the time required for a signal to travel from the sending end to the receiving end in the transmission medium. In LEO satellite networks, propagation delay is mainly related to the distance between satellites and the speed of signal propagation. In multi-path routing transmission, the signal may be transmitted through different satellite paths, which will increase or decrease the overall propagation delay, depending on the length of the path and the network topology. (3) Packet loss rate refers to the ratio of the number of data packets lost during data transmission to the total number of data packets sent, which is an important indicator for measuring the reliability of network transmission. High packet loss rates lead to a decline in communication quality, manifesting as video stuttering, voice interruptions, or data retransmissions, severely impacting user experience. In LEO satellite networks, packet loss rates are more prone to fluctuation due to frequent handovers, link interruptions, and channel interference. Therefore, controlling and optimizing packet loss rates is crucial for ensuring service continuity, improving network stability, and guaranteeing the effective transmission of mission data.

[0022] Dynamic graphical models can characterize the time-varying topology of LEO satellite networks and jointly optimize handover targets and timing, thereby improving handover stability and response accuracy. Meanwhile, integrating digital twin technology to enhance state awareness and policy prediction capabilities has become a recent research focus.

[0023] Based on the aforementioned digital twin and dynamic graph model, this invention proposes a digital twin-driven handover optimization method for low-Earth orbit satellite networks oriented towards multi-service scenarios, constructing a digital twin state matrix DT[ t ][ si Based on a time-directed graph, and introducing dynamic feedback and weight adaptation mechanisms, the global optimum of the end-to-end handover sequence is achieved. DT is introduced into the satellite network, and the optimal satellite handover is selected by verifying the performance of all potential handover satellites using corresponding indicators, rather than relying on a specific scheme. Multi-dimensional QoS indicator parameters are input into the constructed digital twin-driven handover optimization method for low-Earth orbit satellite networks oriented towards multiple service scenarios. The adaptive handover strategy is optimized based on the QoS awareness graph driven by digital twin feedback in the interval [0, ...]. T [Within step size] t Traversing visible satellite time instances in adjacent time slots, including: from the digital twin state matrix DT when coverage and distance constraints are satisfied. t ][ si The QoS vector is read and corrected based on historical feedback. After normalization, the edge cost is calculated according to the weight vector, and the result is added to the weighted directed graph. Then, virtual start and end nodes are added and the Dijkstra algorithm is used to find the globally optimal handover path. The weights can be adaptively adjusted according to the service type to achieve multi-indicator differentiated optimization.

[0024] The satellite handover scenario is built based on the STK 12.1 satellite constellation simulator. The constellation consists of 1584 LEO satellites distributed across 72 orbital planes (22 satellites per orbit), with an orbital altitude of 550 km and an orbital inclination of 53°. The parameters are based on the Starlink Phase 1 constellation configuration. The LEO satellite network performs satellite handover, including: In the design phase of the digital twin-driven handover optimization method for LEO satellite networks, this invention models the handover target selection problem as a multi-objective optimization problem. Specifically, to achieve coordinated optimization of handover strategies and execution timing under differentiated service requirements while considering multi-dimensional QoS indicators such as data rate, propagation delay, and packet loss rate, a multi-objective optimization model is established by normalizing these three indicators into a comprehensive link utility cost. During the handover target selection process, the weights of each indicator are dynamically adjusted based on the future network state predicted by the digital twin. This invention proposes a custom handover sequence selection process. In the adaptive handover strategy for LEO satellite networks driven by a digital twin based on dynamic graph modeling, the custom handover sequence selection process combines future topology evolution predictions with multi-dimensional QoS requirements to determine the optimal handover timing and target satellite sequence within the next service period. An optimal target satellite selection mechanism based on dynamic feedback from the digital twin is constructed, including: using the global state matrix constructed by the digital twin to select the target satellite with the best overall performance from the candidate service satellite set. This mechanism can comprehensively evaluate multi-dimensional QoS indicators such as data rate, propagation latency, and packet loss rate, and dynamically correct predicted values ​​through real-time feedback, thereby improving link stability and service continuity during handover decisions.

[0025] like Figure 2 As shown, the method specifically includes the following steps: S1. Construct a digital twin system to collect and synchronize satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network in real time, and establish a virtual model that evolves synchronously with the physical network.

[0026] Specifically, digital twins collect and synchronize multi-source data such as user location, satellite orbit, and link quality by sensing the status of physical satellite networks in real time, and then construct a dynamic model that reflects the network's operating status after mapping through a virtual satellite network.

[0027] S2. Based on the construction of a digital twin system, the link quality between each candidate satellite and the user terminal is predicted. The prediction time window is divided into multiple time slices, and a directed weighted dynamic graph model is constructed with "satellite-time slice" as nodes and the comprehensive utility cost of the link as the edge weight. A utility function is constructed in the directed weighted graph. The length of the prediction time window is configurable and includes a time relaxation factor, allowing the switching time to be flexibly adjusted within a certain range to balance link stability and switching flexibility. The digital twin system interacts with the physical network in real time to provide feedback correction to the link quality prediction results, forming a closed-loop optimization mechanism of "perception-prediction-decision-execution-feedback".

[0028] Specifically, in a digital twin-driven LEO network, to maintain a communication connection that meets predetermined QoS requirements, during the predicted time interval [0, ...T The process requires a series of satellite handovers. T Divided into equal parts n An equal time slice t i In each time slice t i Within this time slice, all candidate satellites covering the user's location are represented as a single time instance, i.e., "the coverage status of a satellite in this time slice" (e.g., ...). S 1 , t 1 , S 1 , t 2 , S 1 , t 3 They represent satellites S 1 (Instances at different time slices). Each physical satellite will generate several corresponding time instances at different time slices, forming a set. S ={ S 1 , t 1 , S 1 , t 2 ,…, S 1 , t n , S 2 , t 1 , S 2 , t 2 ,…, S k , t n},in k For the total number of satellites, n Let be the total number of time slices. These time instances are treated as graph nodes, and the switching decision criteria are mapped as directed weighted edges, which can be used in the prediction interval [0, ]. T The utility function in the directed weighted graph is constructed by using a directed weighted dynamic graph model with the link comprehensive utility cost as the edge weight.

[0029] In an exemplary embodiment, the utility function in the directed weighted graph is as shown in Equation (1): (1) In the formula, Rsj-i , T sj-i and P sj-i They represent time respectively i Time Satellite s j Data rate, propagation latency, and packet loss rate; ω 1 , ω 2 , ω 3 The weighting coefficients for the three types of indicators, and satisfying the following conditions: ω 1 + ω 2 + ω 3 =1; R min , R max , T min , T max , P min , P max These represent the minimum and maximum values ​​of the corresponding indicators among the candidate switching targets.

[0030] Based on the above utility function, the search for the optimal switching path can be modeled as the following optimization objective: (2) In the formula, p Represented as satellite sj exist i Time to satellite sk exist l The modeling approach is a set of paths for the handover strategy at any given time. This approach can obtain the optimal satellite handover strategy that meets QoS requirements through shortest path search, while considering multiple performance indicators.

[0031] This section will focus on several key QoS metrics, including data rate, propagation latency, and packet loss rate.

[0032] The following formula gives the data rate R calculated based on Shannon's capacity theorem: (3) in, B For channel bandwidth, P r and N These are the received signal power and the noise power, respectively. The received signal power is:

[0033] (4) in, x This is the cloud and rain attenuation coefficient, with units of dB / km; ℎ For satellite altitude, D For the distance of propagation; V Represents the speed of light. f c is the carrier frequency, and is the Rician small-scale fading coefficient; P t For transmission power, G t and G r These are the antenna gains for the transmitting and receiving ends, respectively.

[0034] The following formula gives the method for calculating propagation delay: (5) in, V The speed of light; L ( h , h +1) is a satellite h and h Link distance between +1; T It is the average delay of all handover decisions in the satellite network, taking all satellites into account, and is expressed in milliseconds. H This indicates the overall switching decision.

[0035] The following formula gives the method for calculating packet loss rate: (6) in, P Packet loss rate; P lost The number of lost data packets, that is, the total number of packets that failed to reach their destination during transmission; P sent The total number of data packets sent, i.e., the total number of data packets sent by the source node.

[0036] Optionally, this invention also includes constructing a directed weighted graph model. The digital twin, by sensing the real-time status of the physical satellite network, collects and synchronizes multi-source data such as user location, satellite orbit, and link quality. After mapping through a virtual satellite network, it constructs a dynamic model that reflects the network's operational status. Based on this, the digital twin provides network status prediction capabilities, dynamically injecting multi-temporal topology and multi-dimensional QoS information into the dynamic graph model, forming a directed weighted graph structure that evolves synchronously with the physical network. In this structure, horizontal edges represent continuous service from the same satellite in adjacent time slices (i.e., no handover required), and diagonal edges represent handover connections between different satellite instances with overlapping coverage. The weight of each directed edge is obtained by normalizing multi-dimensional QoS indicators (such as data rate, propagation delay, and packet loss rate) and then weighting them by a weighted sum. Through this method, combined with the prediction-feedback closed-loop mechanism of the digital twin, joint optimal path planning for "handover target-handover timing" is achieved in the LEO network.

[0037] Optionally, based on the aforementioned directed weighted graph model, after completing the optimal target satellite evaluation, further cross-temporal global handover path optimization is required. To this end, this invention designs a digital twin QoS-aware dynamic graph model adaptive handover strategy. This invention constructs a digital twin state matrix DT[ t ][ si Based on a time-directed graph, and by introducing dynamic feedback and weight adaptation mechanisms, the global optimum of the end-to-end switching sequence is achieved.

[0038] S3. Based on the user's service type, set and dynamically adjust the weighting coefficients of data rate, propagation latency, and packet loss rate in the overall utility cost. For real-time control services, increase the weighting of propagation latency and packet loss rate. ω 2 , ω 3 Reduce the weight of data rate ω 1 For data-intensive businesses, increase the weight given to data rate. ω 1 Reduce the weight of propagation delay and packet loss rate ω 2 , ω 3 The weighting coefficients are dynamically penalized or enhanced based on whether the Quality of Service (QoS) threshold is met.

[0039] Specifically, during the handover target selection process, the weights of various indicators are dynamically adjusted based on the future network state predicted by the digital twin. In this invention, a dynamic weight adjustment method is designed that can penalize links that do not meet specific QoS thresholds. This ensures service continuity while achieving personalized optimization and traffic balancing of the handover target, effectively reducing the risk of service interruption caused by frequent handovers or link performance degradation.

[0040] Optionally, it also includes: data rate, which refers to the amount of data transmitted through a communication channel per unit time, used to measure the data transmission capacity of a link. Its magnitude directly reflects the bandwidth utilization efficiency and information transmission speed of the communication system, and is a core indicator for evaluating communication quality and network performance.

[0041] Optionally, it also includes: propagation delay refers to the time required for a signal to travel from the transmitting end to the receiving end in the transmission medium. In LEO satellite networks, propagation delay is mainly related to the distance between satellites and the signal propagation speed.

[0042] Optionally, it also includes: Packet loss rate, which is the ratio of the number of data packets lost during data transmission to the total number of data packets sent, is an important indicator for measuring the reliability of network transmission. A high packet loss rate will lead to a decline in communication quality, manifested as video stuttering, voice interruption, or data retransmission, seriously affecting the user experience.

[0043] S4. In the directed weighted dynamic graph model, the optimal path from the starting node to the ending node is searched using the shortest path algorithm to determine the optimal handover timing and target satellite sequence. The optimal handover strategy is then distributed to the user terminal for execution, and the prediction model is corrected in real time through the feedback mechanism of the digital twin system.

[0044] Horizontal edges represent continuous service from the same satellite within adjacent time slots, while diagonal edges represent handover connections between different satellites. The diagram uses a virtual start node to connect all candidate satellite instances in the first time slot and a virtual end node to receive all candidate satellite instances in the last time slot. The shortest path algorithm used is Dijkstra's algorithm.

[0045] Specifically, in this invention, the optimal target satellite selection mechanism is a key step in constructing a switching optimization mechanism that integrates dynamic sensing and adaptive scheduling. First, the algorithm initializes the digital twin state matrix DT[ t ][ si ] and historical cache H and in the time interval [0, T [Within step size Δ] tThe system iteratively processes the visible satellite set. Within each time slot, it calculates the propagation distance and delay, evaluates the link gain and received power, and obtains the data rate, propagation delay, and packet loss rate. If historical data exists, the indicators are corrected based on real-time feedback. After completing calculations for all time slots, the system directly selects the satellite with the highest utility value as the handover target based on the comprehensive performance indicators of each candidate satellite in the digital twin state matrix, thus achieving stable and efficient target satellite decision-making at the handover moment.

[0046] This invention proposes a digital twin-driven handover optimization method for low Earth Orbit (LEO) satellite networks, oriented towards multiple service scenarios, based on multi-dimensional indicators such as data rate, propagation delay, and packet loss rate. The method constructs a dynamic graph model based on digital twins, jointly optimizing handover targets and handover timing to achieve real-time mapping and evolution prediction of the highly dynamic topology of LEO satellite networks. It introduces digital twin-driven orbit and link state forward prediction for real-time updating of candidate handover paths. A QoS-aware dynamic weighting mechanism is designed to normalize and adaptively adjust indicators such as data rate, propagation delay, and packet loss rate to generate handover targets for different service scenarios.

[0047] This invention relates to a digital twin-driven handover optimization method for low-Earth orbit satellite networks, applicable to multiple service scenarios, particularly suitable for dynamic communication application environments with high requirements for link continuity, network responsiveness, and quality of service assurance. In diverse service scenarios such as telemedicine, disaster early warning, intelligent transportation, emergency communication, and remote area networking, the digital twin-driven handover mechanism proposed in this invention enables real-time modeling and prediction of satellite node status, service requirements, and network topology evolution, thereby intelligently deciding on handover paths and timing, significantly improving network service continuity and resource utilization efficiency. Compared with existing technologies, this invention possesses stronger sensing, adaptive, and policy coordination capabilities, supporting stable connections in high-frequency handover environments, meeting the core requirements of next-generation satellite networks for agile service scheduling and dynamic network resource adaptation, and has broad engineering application prospects and promotional value.

[0048] Based on the same inventive concept, this invention also proposes a digital twin-driven handover optimization system for low-Earth orbit satellite networks, comprising: The digital model building module is used to collect satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network in real time, and to establish a digital twin system that evolves synchronously with the physical LEO satellite network.

[0049] The dynamic graph construction module is used to predict the link quality between candidate satellites and user terminals within a future time window based on a digital twin system. It divides the time window into multiple consecutive time slices and constructs a directed weighted dynamic graph model with "satellite-time slice" as nodes and link comprehensive utility cost as edge weights. The link comprehensive utility cost is obtained by normalizing and weighting the data rate, propagation delay, and packet loss rate indicators.

[0050] The search module is used to dynamically configure the weight coefficients of various QoS indicators in the comprehensive utility cost of the link according to the user's current service type; based on the configured weight coefficients, the shortest path algorithm is used to search for the optimal path from the virtual start node to the virtual end node in the directed weighted dynamic graph model, so as to determine the optimal handover time and target satellite sequence.

[0051] The optimization module is used to send the optimal handover strategy to the user terminal for execution, and to correct the link quality prediction results in real time through the feedback mechanism of the digital twin system, forming a closed-loop optimization.

[0052] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A digital twin-driven handover optimization method for low-Earth orbit satellite networks, characterized in that, Includes the following steps: By collecting real-time satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network, a digital twin system that evolves synchronously with the physical LEO satellite network is established. Based on a digital twin system, the link quality between each candidate satellite and the user terminal within a future time window is predicted. The time window is divided into multiple consecutive time slices, and a directed weighted dynamic graph model is constructed with satellite-time slice as nodes and link comprehensive utility cost as edge weights. The link comprehensive utility cost is obtained by normalizing and weighting the data rate, propagation delay, and packet loss rate indicators. Based on the user's current service type, dynamically configure the weight coefficients of various QoS indicators in the comprehensive utility cost of the link; based on the configured weight coefficients, use the shortest path algorithm to search for the optimal path from the virtual start node to the virtual end node in the directed weighted dynamic graph model, so as to determine the optimal handover timing and target satellite sequence; The optimal switching strategy is sent to the user terminal for execution, and the link quality prediction results are corrected in real time through the feedback mechanism of the digital twin system, forming a closed-loop optimization.

2. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, The utility function calculation process in the directed weighted dynamic graph model is expressed as follows: ； in, R sj-i , T sj-i and P sj-i They represent time respectively i Time Satellite s j Data rate, propagation latency, and packet loss rate; ω 1 , ω 2 , ω 3 The weighting coefficients for the three types of indicators, and satisfying the following conditions: ω 1 + ω 2 + ω 3 =1; R min , R max , T min , T max , P min , P max These represent the minimum and maximum values ​​of the corresponding indicators among the candidate switching targets.

3. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, The calculation processes for the data rate, propagation delay, and packet loss rate are respectively expressed as follows: Calculate the data rate based on Shannon's capacity theorem. R : ;in, B For channel bandwidth, P r and N These are the received signal power and the noise power, respectively. The calculation process for propagation delay is expressed as follows: ;in, V The speed of light; L ( h , h +1) is a satellite h and h Link distance between +1; T It is the average delay of all handover decisions in the satellite network, taking all satellites into account, and is expressed in milliseconds. H Indicates the overall switching decision; The calculation process for packet loss rate is expressed as follows: ;in, P Packet loss rate; P lost The number of lost data packets, that is, the total number of packets that failed to reach their destination during transmission; P sent The total number of data packets sent, i.e., the total number of data packets sent by the source node.

4. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, It also includes dynamically adjusting the weighting coefficients of the data rate, propagation latency, and packet loss rate in the overall utility cost of the link based on the user's service type, specifically including the following steps: For real-time control applications, increase the weighting of propagation latency and packet loss rate. ω 2 , ω 3 Reduce the weight of data rate ω 1 ; For data-intensive businesses, increase the weight given to data rate. ω 1 Reduce the weight of propagation delay and packet loss rate ω 2 , ω 3 .

5. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 4, characterized in that, The weighting coefficients are dynamically penalized or enhanced based on whether the service quality threshold is met.

6. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, In the directed weighted dynamic graph model: horizontal edges represent continuous service of the same satellite in adjacent time slots; diagonal edges represent handover connections between different satellites with overlapping coverage; a virtual start node is set to connect all candidate satellite instances in the first time slot, and a virtual end node is set to connect all candidate satellite instances in the last time slot.

7. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, The shortest path algorithm is Dijkstra's algorithm.

8. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, The length of the prediction time window is configurable, and a time relaxation factor is introduced to allow the actual switching time to be flexibly adjusted within the prediction time slice range.

9. The method for optimizing digital twin-driven handover in low-Earth orbit satellite networks according to claim 1, characterized in that, The digital twin system interacts with the physical LEO satellite network in real time to achieve a closed-loop optimization mechanism of "perception-prediction-decision-execution-feedback" and continuously correct the link quality prediction results.

10. A digital twin-driven switching optimization system for low-Earth orbit satellite networks, characterized in that, include: The digital model building module is used to collect satellite orbit parameters, user terminal locations, and link status information of the physical LEO satellite network in real time, and to establish a digital twin system that evolves synchronously with the physical LEO satellite network. A dynamic graph construction module is used to predict the link quality between candidate satellites and user terminals within a future time window based on a digital twin system. The time window is divided into multiple consecutive time slices, and a directed weighted dynamic graph model is constructed with satellite-time slices as nodes and link comprehensive utility cost as edge weights. The link comprehensive utility cost is obtained by normalizing and weighting the data rate, propagation delay, and packet loss rate indicators. The search module is used to dynamically configure the weight coefficients of various QoS indicators in the comprehensive utility cost of the link according to the user's current service type; based on the configured weight coefficients, the shortest path algorithm is used to search for the optimal path from the virtual start node to the virtual end node in the directed weighted dynamic graph model, so as to determine the optimal handover time and target satellite sequence. The optimization module is used to send the optimal handover strategy to the user terminal for execution, and to correct the link quality prediction results in real time through the feedback mechanism of the digital twin system, forming a closed-loop optimization.

Images