A communication method and satellite communication system of a low earth orbit satellite

Abstract

The application relates to the technical field of satellite communication, and discloses a communication method of a low-orbit satellite and a satellite communication system; the method comprises six steps of S1 beam switching prediction, S2 preparatory link establishment and incremental context synchronization, S3 soft switching execution, S4 distributed inter-satellite route calculation, S5 dynamic beam resource allocation and S6 cross-module collaborative control; a switching prediction sequence is generated based on an SGP4 orbit prediction, a predictive soft switching is realized through incremental synchronization of differential state quantities encoded by TLV, a distributed on-satellite route is constructed by combining a two-hop adjacent local topology with ARIMA link quality prediction, a dynamic beam resource allocation is realized by adopting a QoS weighted demand factor model and a hysteresis mechanism, time coordination, target optimization and QoS end-to-end penetration are adopted to realize the synergy of the three mechanisms, the switching interruption is reduced to sub-millisecond level, the route delay is reduced by 25%, the system throughput is increased by 35%, and an additional 22% performance gain is brought by the synergy mechanism, and the defects of the prior art are solved.

Description

Technical Field

[0001] This invention belongs to the field of satellite communication technology, specifically a communication method and satellite communication system for low-Earth orbit satellites. Background Technology

[0002] With the rapid development of low Earth orbit satellite constellations, low Earth orbit satellite communication systems have become an important part of the next generation of communication networks due to their advantages such as global coverage, low latency, and high bandwidth. A typical low Earth orbit satellite constellation consists of hundreds to thousands of satellites that operate at high speeds in orbits 500 to 2,000 kilometers above the ground. The visibility time of a single satellite relative to a ground terminal is usually only a few minutes to a dozen minutes.

[0003] The existing low-Earth orbit satellite communication system mainly adopts the following technical architecture: the ground terminal establishes a connection with a single service satellite through the user link, the satellites forward data through the inter-satellite link, and the gateway station is responsible for the overall scheduling and management of the system; when the satellite moves out of the terminal's line of sight, the terminal performs a beam switching operation to transfer the communication link to the next visible satellite.

[0004] Existing technologies have conducted extensive research in three areas: handover, routing, and resource allocation. In terms of handover technology, existing solutions have proposed the idea of ​​soft handover with the principle of "connect first, disconnect later" to reduce handover interruptions. In terms of routing technology, existing research has explored distributed routing architectures in an attempt to overcome the latency problem of centralized scheduling. In terms of resource allocation, there are existing dynamic beam adjustment schemes that optimize resource allocation based on user load.

[0005] However, existing technologies have the following fundamental flaws: First, beam switching interruption is still not ideal; most existing soft handover schemes are passively triggered, that is, the handover process is only started after the signal quality deterioration is detected. There is a lack of active prediction mechanism based on orbit prediction, the handover preparation time is compressed, the context synchronization is insufficient, and the actual handover interruption is still tens of milliseconds. In addition, the context synchronization mostly adopts the full transmission method, which has a large amount of data and occupies a lot of inter-satellite link bandwidth.

[0006] Second, it is difficult to balance efficiency and accuracy in distributed routing. Existing distributed routing either maintains the entire network topology, resulting in huge computational and storage overhead, or uses a simple local greedy algorithm, leading to a decline in routing quality. More importantly, existing routes are all calculated based on the current link state, without considering the rapid changes in link quality caused by the relative motion of low-orbit satellites. By the time the route calculation is completed, the link state has already changed, causing route oscillations and frequent recalculations.

[0007] Third, the accuracy of dynamic resource allocation is insufficient; existing solutions mostly allocate resources simply based on the number of users without considering the QoS priority differences of different services, resulting in high-priority services not receiving sufficient resource guarantees during busy periods; at the same time, the lack of oscillation suppression mechanisms and frequent resource adjustments increase system overhead.

[0008] Fourth, and most critically, the three mechanisms of handover, routing, and resource allocation are designed and optimized independently, lacking cross-layer collaboration. In existing technologies, handover decisions do not consider the resource status of the target beam, and the asynchronous nature of routing updates and beam switching leads to packet loss. Resource allocation is unaware of changes in user distribution caused by handover. The optimization effects of the three mechanisms cancel each other out or even conflict with each other, resulting in overall system performance far lower than the sum of the theoretical gains of each part. This siloed design approach is a long-standing technical bias in the field, and those skilled in the art are generally accustomed to layered independent optimization, rarely considering the additional gains brought by cross-module collaboration.

[0009] Therefore, there is an urgent need for a low-Earth orbit satellite communication method that can achieve predictive seamless handover, predictive distributed routing, QoS-aware dynamic resource allocation, and overall performance leap through cross-module collaborative mechanisms, in order to solve the technical problems of excessive handover interruptions, frequent routing oscillations, inaccurate resource allocation, and especially the lack of collaborative optimization in existing technologies. Summary of the Invention

[0010] The purpose of this invention is to overcome the above-mentioned defects of the prior art and provide a communication method and satellite communication system for low-orbit satellites. Through three core technological innovations—predictive soft handover, predictive distributed routing, and QoS-aware dynamic resource allocation—coupled with the synergistic effect of cross-module collaborative mechanisms, a significant leap in the overall performance of the system is achieved.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a communication method and satellite communication system for low-Earth orbit satellites, comprising the following steps: S1. Beam switching prediction: Obtain the terminal's position information and motion state parameters, combine them with the two-line element orbit prediction data of the on-orbit satellites, calculate the position change curves of each visible satellite based on the SGP4 orbit propagation model, and generate a beam switching prediction sequence containing the target satellite identifier and the expected switching time. S2. Preparatory Link Establishment and Incremental Context Synchronization: When the current serving beam maintains a communication connection and the signal-to-noise ratio is higher than a preset threshold, access is initiated to the target satellite in advance according to the beam switching prediction sequence and a preparatory beam link is established. The communication data context of the PDCP layer, ROHC layer and encryption layer is transmitted to the target satellite in the form of differential state variables through the inter-satellite link. S3. Distributed Inter-Satellite Routing Calculation: Based on the distributed inter-satellite routing protocol, each satellite node periodically broadcasts link status messages to neighboring satellites. Each node only maintains a local topology view within a two-hop adjacency range and independently calculates the optimal path for data forwarding by combining the link quality time series prediction value. S4. Dynamic beam resource allocation: Based on the user density distribution and QoS level classification of each service within the coverage area, the transmit power and carrier bandwidth resources of multiple beams are dynamically allocated proportionally using a weighted demand factor model. S5. Cross-module collaborative control: Establish a cross-module collaborative mechanism. The beam handover prediction module sends the predicted handover time value to the distributed routing module in advance to trigger route pre-calculation. The dynamic resource allocation module feeds back the remaining resource quota of each beam to the beam handover prediction module to optimize target selection. Service QoS level parameters are simultaneously input into three stages: soft handover priority scheduling, route cost weight calculation, and resource allocation weight configuration.

[0012] Preferably, step S1 specifically includes: calculating the instantaneous coordinates of each satellite in the geocentric-ground-fixed coordinate system within the prediction window based on the SGP4 model with a time step of 1 second; converting to the station-centric coordinate system through a coordinate transformation matrix, and calculating the azimuth, elevation, and slant range of the satellite relative to the terminal at each moment; determining that the satellite is visible when the elevation angle is greater than 10 degrees, and extracting the continuous visible time period as the service window; and generating a beam switching prediction sequence containing the target satellite ID, beam number, expected access time, and available bandwidth prediction value according to the start time of the service window, with the prediction window covering the next 30 minutes.

[0013] Preferably, step S2 specifically includes: when the signal-to-noise ratio of the current serving beam is higher than 10dB, sending a preliminary link establishment request to the predicted next target satellite; establishing a preliminary data channel after the target satellite completes access authentication; synchronizing incremental context to the target satellite through the inter-satellite link, wherein the incremental context adopts TLV encoding format, including PDCP sequence number differential mapping table, ROHC compressed context status number, encryption and integrity key count value, QoS flow tunnel endpoint parameters and unconfirmed PDU sequence number list, and the amount of data synchronized at one time is less than 10KB.

[0014] Preferably, step S3 specifically includes: each satellite node broadcasting a link status message to all neighboring satellites at a period of 500 milliseconds. The message carries available bandwidth, transmission delay, bit error rate, and the link quality trend value predicted by the ARIMA model for the next period; each node stores only the neighboring nodes and link information within a two-hop range to construct a local topology view; a minimum cost routing algorithm with delay constraints is used to calculate the forwarding table, and the cost function is Cost = α × (1 / available bandwidth) + β × delay + γ × bit error rate, where α, β, and γ are configurable weight coefficients; incremental recalculation is triggered when the link quality deteriorates by more than 30%.

[0015] Preferably, step S4 specifically includes: counting the number of users for each beam according to three priority levels: real-time session, streaming media, and background; and calculating the resource demand factor for each beam using the following formula: =w1× +w2× +w3× Where w1, w2, and w3 are the weight coefficients of the corresponding priorities, and the weight values ​​decrease sequentially as the priority decreases; the total power budget and total bandwidth pool are allocated according to the proportion of each beam demand factor to the total demand factor; a hysteresis threshold of 10% variation amplitude is introduced to suppress resource oscillation, and the resource adjustment cycle is 1 second.

[0016] Preferably, in step S5, the beam switching prediction module sends the predicted switching time value to the distributed routing module in advance to trigger route pre-calculation. Specifically, this includes: after generating the predicted switching time value, the beam switching prediction module sends a pre-calculation notification to the distributed routing module one route update cycle in advance, the notification carrying the target satellite identifier and the expected switching effective time; the distributed routing module calculates the new path through the target satellite in advance and sets the new routing table to take effect on time at the switching completion time.

[0017] Preferably, in step S5, the dynamic resource allocation module feeds back the remaining resource quota of each beam to the beam switching prediction module to optimize target selection. Specifically, this includes: the dynamic resource allocation module maintaining the remaining available bandwidth and remaining power quota of each beam in real time; after generating the initial candidate beam sequence, the beam switching prediction module queries the remaining resource quota of each candidate beam, prioritizes the beams that can meet the bandwidth requirements of the terminal services, and downgrades the beams with insufficient resources to the alternatives.

[0018] This invention provides a low-Earth orbit satellite communication system, comprising: The beam switching prediction module is used to obtain terminal position and motion state parameters, and generate beam switching prediction sequences based on the SGP4 model in combination with satellite orbit prediction data; The soft handover execution module is used to pre-establish a backup beam link during the current service beam communication and transmit the communication data context to the target satellite in the form of differential state variables via the inter-satellite link; The distributed routing module, deployed on each satellite node, is used to construct a local topology view within a two-hop adjacency range and independently calculate the optimal forwarding path based on the link quality prediction value. The dynamic resource allocation module is used to dynamically allocate multi-beam power and bandwidth resources based on the weighted demand factor model, according to user density distribution and service QoS level. The cross-module collaborative control module is used to send the handover time to the distributed routing module in advance to trigger route pre-calculation, to feed back beam resource quotas to the beam handover prediction module to optimize target selection, and to synchronously output QoS level parameters to the three modules of soft handover scheduling, route calculation and resource allocation.

[0019] Compared with the prior art, the beneficial effects of the present invention are: The handover performance is greatly improved. Based on orbit prediction, the present invention realizes active predictive handover. With the help of incremental context synchronization mechanism, the handover interruption delay is further reduced from tens of milliseconds in the existing soft handover scheme to less than 1 millisecond. The handover preparation lead time is increased by more than 5 times, while the amount of context synchronization data is reduced by more than 80%. This invention improves both routing efficiency and accuracy; the two-hop local topology reduces the routing computation complexity from O(n^2) to O(n^2). 2 The computation time is reduced to the O(1) level, and storage overhead is reduced by 90%; combined with the pre-calculation mechanism of link quality prediction, the frequency of route recalculation is reduced by more than 60%, and the average end-to-end latency of data packets is reduced by 25%. Enhanced resource utilization and QoS assurance; the QoS-weighted demand factor model increases the resource allocation for high-priority services by over 50%, and the hysteresis mechanism reduces resource adjustment frequency by 70%; overall system throughput is increased by 35% in scenarios with uneven user distribution. The cross-module collaboration produces outstanding synergistic effects; this is the core creative contribution of this invention: time collaboration reduces packet loss rate during handover by more than 90%, target optimization reduces handover blocking rate by 40%, and QoS end-to-end penetration improves end-to-end latency consistency of high-priority services by 35%; the overall performance gain brought by the collaboration of the three mechanisms is significantly higher than the sum of the effects of the three technologies applied individually. This synergistic effect of 1+1+1>3 is unpredictable by existing technologies and has outstanding substantive characteristics and significant progress. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0021] In the attached diagram: Figure 1 This is a schematic diagram of the overall architecture of a low-Earth orbit satellite communication system provided in an embodiment of the present invention; Figure 2 A flowchart illustrating the low-Earth orbit satellite communication method provided in an embodiment of the present invention; Figure 3 A timing diagram illustrating the predictive soft handover mechanism provided in an embodiment of the present invention; Figure 4This is a structural block diagram of the distributed inter-satellite routing module provided in an embodiment of the present invention; Figure 5 A schematic diagram illustrating the principle of dynamic beam resource allocation provided in an embodiment of the present invention; Figure 6 A schematic diagram of the architecture of the cross-module collaborative control mechanism provided in an embodiment of the present invention; In the diagram: 101, User terminal; 102, Phased array antenna module; 103, Positioning unit; 104, Motion sensor; 110, Low Earth orbit satellite; 111, User beam array; 112, Inter-satellite link terminal; 113, On-board processing unit; 114, Ephemeris storage unit; 120, Gateway station; 400, Distributed routing module; 401, Link status broadcasting unit; 402, Local topology maintenance unit; 403, Path calculation unit; 501, Resource demand statistics unit; 502, Demand factor calculation unit; 503, Resource allocation calculation unit; 504, Beamforming control unit; 600, Cross-module collaborative control module; 601, Beam switching prediction module; 602, Distributed routing module; 603, Dynamic resource allocation module. Detailed Implementation

[0022] 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. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0023] This embodiment provides a communication method and satellite communication system for low-Earth orbit satellites, applicable to the Walker constellation communication system composed of 72 low-Earth orbit satellites. The system achieves highly reliable, low-latency, high-capacity, and QoS-guaranteed satellite communication services through three core technologies: predictive soft handover, predictive distributed routing, and QoS-aware dynamic resource allocation, coupled with the synergistic effect of cross-module collaborative control mechanisms.

[0024] I. System Overall Architecture Please refer to Figure 1 The overall architecture of the low-orbit satellite communication system in this embodiment includes three layers: terminal layer, satellite constellation layer, and ground gateway layer.

[0025] The terminal layer includes multiple user terminals 101, which can be various satellite communication terminals such as handheld satellite terminals, vehicle-mounted terminals, shipborne terminals, airborne terminals, or Internet of Things terminals. Each user terminal 101 is equipped with a phased array antenna module 102, a positioning unit 103, and a motion sensor 104. The positioning unit 103 adopts a multi-mode GNSS receiver, supports joint positioning of GPS, Beidou, and Galileo systems, has a horizontal positioning accuracy better than 1 meter, and an output frequency of 1 Hz. The motion sensor 104 includes a three-axis accelerometer and a three-axis gyroscope with ranges of ±10g and ±500° / s, respectively, used to measure the terminal's three-dimensional velocity vector and attitude, and assist in Doppler frequency shift pre-compensation.

[0026] The satellite constellation consists of 72 low-Earth orbit (LEO) satellites 110. Each LEO satellite 110 operates in a near-circular orbit at an altitude of 550 kilometers with an orbital inclination of 53 degrees. They are distributed across 6 orbital planes, with 12 satellites in each plane. The orbital phases are evenly distributed, achieving continuous coverage of the global region within 53 degrees latitude. Each LEO satellite 110 carries a user beam array 111, an inter-satellite link terminal 112, an on-board processing unit 113, and an ephemeris storage unit 114.

[0027] The user beam array 111 employs digital multi-beamforming technology, which is achieved by a 64-element phased array antenna in conjunction with a digital beamformer. It can simultaneously form 16 independent spot beams, each with a 3dB beamwidth of 2.5 degrees. The pointing, transmit power, and bandwidth allocation can be independently adjusted. The total downlink capacity of a single satellite is 10Gbps, and the total transmit power is 200W.

[0028] Inter-satellite link terminal 112 adopts a laser communication system. Each satellite is equipped with 4 laser terminals, which are used to establish high-speed inter-satellite links with neighboring satellites in front and behind and adjacent satellites in the left and right orbital planes. The single link rate can reach 10Gbps and the link distance can reach up to 5000 kilometers.

[0029] The on-board processing unit 113 is an on-board computing platform that adopts a multi-core ARM architecture, with a computing power of 10,000 DMIPS and 4GB of memory. It performs functions such as baseband processing, routing calculation, resource scheduling, and collaborative control.

[0030] The ephemeris storage unit 114 stores orbital prediction data for all 72 satellites in the network, using a two-line element (TLE) format. The prediction accuracy is better than 10 meters, and the update cycle is 24 hours. The data is uploaded and updated by the ground gateway station through the telemetry and control link.

[0031] The ground gateway station layer includes multiple gateway stations 120, with a total of 8 gateway stations deployed globally. They are mainly responsible for system network access control, billing management, ephemeris uploading, and interconnection with the terrestrial public network. Unlike the traditional centralized architecture, in this embodiment, the gateway station 120 does not participate in real-time routing calculation and beam scheduling. It only provides initial configuration and periodic ephemeris updates when the satellite enters the network. All real-time decisions are made on the satellite.

[0032] II. Overview of Methods and Procedures Please refer to Figure 2 The low-orbit satellite communication method in this embodiment includes the following five steps: S1 beam switching prediction, S2 pre-link establishment and incremental context synchronization, S3 soft handover execution, S4 distributed inter-satellite routing calculation, S5 dynamic beam resource allocation, and S6 cross-module collaborative control throughout the entire process.

[0033] S1, Beam switching prediction.

[0034] User terminal 101 obtains its own location information such as latitude, longitude, and altitude through positioning unit 103, and obtains three-dimensional velocity vector through motion sensor 104. It sends the above parameters to the currently serving low-Earth orbit satellite 110 through uplink control channel, with a reporting period of 5 seconds. The on-board processing unit 113 of the currently serving low-Earth orbit satellite 110 reads the TLE orbit parameters of all visible satellites from ephemeris storage unit 114, and calculates the visibility window, pitch angle change curve, and Doppler frequency shift change curve of each satellite relative to the terminal within the next 30 minutes, in combination with the terminal's position and motion status.

[0035] The specific calculation process is as follows: (1) Orbit propagation calculation: Based on the SGP4 orbit propagation model, the two lines of metadata in TLE are used as input, and the instantaneous position coordinates (X,Y,Z) of each satellite in the prediction window are calculated with a time step of 1 second. The coordinate system adopted is the geocentric geofixed coordinate system (ECEF). The input of the SGP4 model includes the satellite number, inclination, right ascension of the ascending node, eccentricity, argument of perigee, mean perigee, and mean motion orbital elements in the first line of TLE. The output is the satellite position vector and velocity vector at the corresponding time.

[0036] (2) Coordinate transformation: The ECEF coordinates of the satellite and the geodetic coordinates (longitude λ, latitude φ, altitude h) of the terminal are transformed. First, the geodetic coordinates of the terminal are converted to ECEF coordinates, and then the transformation is performed to the station center coordinate system (ENU) through the station center transformation matrix. The station center transformation matrix is: R=[-sinλ,cosλ,0;-sinφcosλ,-sinφsinλ,cosφ;cosφcosλ,cosφsinλ,sinφ], where φ is the latitude of the terminal and λ is the longitude of the terminal.

[0037] (3) Visibility determination: Calculate the satellite azimuth, elevation angle and slant range at each moment in the station center coordinate system; when the elevation angle is greater than 10 degrees, it is determined to be visible, and the continuous visible time period is the service window of the satellite; considering the influence of atmospheric refraction, the elevation angle is corrected by about 0.5 degrees at low elevation angles.

[0038] (4) Doppler calculation: Calculate the radial velocity of each visible satellite relative to the terminal. Thus, the Doppler frequency shift fd = × / c, where Where is the carrier frequency and c is the speed of light; this Doppler estimate is used for frequency offset pre-compensation of the terminal.

[0039] (5) Predicted sequence generation: Select the satellite sequence that will enter the terminal's field of view in the future, sort them according to the service start time, and generate a beam switching predicted sequence; each record includes: target satellite ID, corresponding beam number, expected access time, expected disconnection time and maximum available bandwidth predicted value; In this embodiment, a single prediction generates a sequence containing 5 candidate switching targets, with a prediction accuracy better than 95% and a switching time prediction error of less than 2 seconds.

[0040] S2, Preparatory Link Establishment and Incremental Context Synchronization.

[0041] Please refer to Figure 3 According to the timing diagram, when the beam signal quality of the currently serving satellite S1 is still much higher than the handover threshold (i.e., the received signal-to-noise ratio is greater than 10dB, and there are still about 30 seconds before normal handover), the system starts the handover preparation process in advance.

[0042] At time T1, the current serving satellite S1 sends a preparatory link establishment request to the predicted next target satellite S2 through the inter-satellite link terminal 112 according to the beam switching prediction sequence. The request carries the terminal identifier IMSI, service type identifier, required bandwidth and QoS level. After receiving the request, the target satellite S2 checks the remaining resources of the corresponding beam. If the requirements are met, it allocates the corresponding bandwidth and power resources and sends an access preamble signal to the terminal.

[0043] At time T2, the terminal detects the preparatory beam signal of the target satellite S2 through the second beam of the phased array antenna module 102, and sends a random access request after completing downlink time and frequency synchronization. The target satellite S2 completes the terminal authentication and establishes the bearer at each layer of PDCP, RLC and MAC, forming a preparatory data channel. At this time, the terminal has a link connection with both satellites at the same time, but the service data is still mainly transmitted through the current serving satellite S1. The preparatory link only transmits a small number of probe packets to maintain link synchronization. The probe packet transmission period is 200 milliseconds.

[0044] At time T3, the current serving satellite S1 synchronizes the incremental communication context with the target satellite S2 through the inter-satellite link terminal 112. Unlike the traditional full synchronization, this embodiment adopts an incremental differential synchronization mechanism, which only transmits the changes in state variables. The context data adopts the TLV (Type-Length-Value) three-level encoding format, with the type field occupying 1 byte, the length field occupying 2 bytes, and the value field having a variable length.

[0045] The specific fields to be synchronized include: --PDCP layer context (type code 0x01): superframe number HFN (4 bytes), next expected receive sequence number RX_NEXT (2 bytes), sender-side maximum acknowledgment sequence number TX_MAX_ACK (2 bytes); only transmits the difference from the baseline state, which is established during the first synchronization or full context refresh.

[0046] --ROHC Compression Context (Type Code 0x02): Compression context status number for each IP stream (1 byte per stream), only for streams whose status has changed.

[0047] -- Encryption and Integrity Protection Context (Type Code 0x03): Encryption key COUNT value (4 bytes), Integrity key COUNT value (4 bytes).

[0048] --QoS flow tunnel parameters (type code 0x04): GTP tunnel endpoint identifier TEID (4 bytes) and QoS configuration parameters for each QoS flow.

[0049] --Unacknowledged PDU list (type code 0x05): A list of unacknowledged PDU sequence numbers sent in AM mode, using bitmap encoding.

[0050] The total data volume of incremental context synchronization is less than 10KB, and it can be transmitted within a single inter-satellite link time slot (1 millisecond), with minimal bandwidth consumption on the inter-satellite link. After synchronization is completed, the target satellite sends an acknowledgment message. If no acknowledgment is received within the timeout period, it will retransmit once.

[0051] S3, soft switch execution.

[0052] At time T4, the signal quality of the currently serving satellite S1 drops to the handover trigger threshold (signal-to-noise ratio equal to 3dB), and the system officially initiates the handover. The terminal simultaneously sends uplink data to both satellites, and the two downlink satellites simultaneously send the same data to the terminal. The terminal uses maximum ratio combining (MRC) to receive the two signals, improving the received signal-to-noise ratio. This dual-link parallel transmission phase lasts for about 2 seconds, ensuring that no data is lost during the handover process.

[0053] At time T5, if the link quality of the target satellite S2 is consistently higher than that of the current serving satellite S1 for more than 1 second, the system determines that the handover is complete, disconnects the link connection with the original serving satellite S1, and switches all service traffic to the target satellite S2. Since the target satellite has obtained the complete communication context in advance, there is no need to re-establish the bearer, negotiate parameters, and synchronize the state after the handover. The service interruption time perceived by the user is less than 1 millisecond, achieving a truly seamless handover.

[0054] S4, Distributed Inter-Satellite Routing Calculation.

[0055] Please refer to Figure 4 Each satellite's distributed routing module 400 includes three sub-modules: a link status broadcasting unit 401, a local topology maintenance unit 402, and a path calculation unit 403.

[0056] Link status broadcast unit 401 sends link status broadcast messages to all neighboring satellites via four inter-satellite links at a period of 500 milliseconds. The message format is: source node ID (2 bytes), sequence number (2 bytes), number of link entries (1 byte), and several link entries. Each link entry includes: peer node ID (2 bytes), available bandwidth (2 bytes, in Mbps), average latency (2 bytes, in milliseconds), and bit error rate (2 bytes, in milliseconds). ), Link quality prediction trend value (1 byte, -100 to +100, representing the expected percentage change in the next period).

[0057] Link quality prediction uses an ARIMA(2,1,1) time series model. The model input is the link available bandwidth measurement value within the last 5 minutes, with a sampling interval of 100 milliseconds and a total of 3000 sample points. The model order (p=2, d=1, q=1) is determined by the AIC information criterion, and the parameters are updated online using the recursive least squares method with an update period of 5 seconds. The output is the predicted link available bandwidth value for the next period (500 milliseconds later). The purpose of introducing the predicted value is to detect changes in link quality caused by the relative motion of satellites in advance, so that the route adjustment is forward-looking. During the cold start phase, when the historical data is less than 1 minute, the current measurement value is used directly as the predicted value.

[0058] Local topology maintenance unit 402 collects link-state messages from adjacent satellites and constructs a local network topology view within a two-hop range; the topology maintenance rules are as follows: (1) Direct Neighbor (One Hop): Satellites directly connected via inter-satellite links obtain link status from local measurement data.

[0059] (2) Two-hop neighbor: the neighbor of the neighbor, extracted from the link state message sent by the neighbor.

[0060] (3) Information on nodes that are three hops or further away is not stored and is discarded directly.

[0061] Each topology record has an aging timestamp. Links that have not received an update for more than 3 broadcast cycles (i.e., 1.5 seconds) are marked as invalid and removed from the topology view. Unlike the traditional OSPF protocol, which maintains the topology of the entire network of 72 nodes, this embodiment only maintains a two-hop local topology. The topology view of each node contains only about 10 nodes and 20 links, which greatly reduces storage overhead and computational complexity.

[0062] The path calculation unit 403 uses a minimum cost routing algorithm with time delay constraints to calculate the forwarding table; the algorithm pseudocode is as follows: Input: Local topology graph G, source node s, destination node d, maximum time delay constraint Output: Optimal next hop node 1. Initialize the distance array dist[] to infinity and the predecessor array prev[] to null. 2.dist[s]=0 3. Use Dijkstra's algorithm to calculate the minimum delay path from s to all nodes. 4. Filter out those that satisfy dist[d] <= The set of all candidate paths P 5. If P is empty, relax the constraint to 1.5× Re-filter 6. Calculate the cost for each path in P: Cost = α × (1 / available bandwidth of the path) + β × total delay of the path + γ × bit error rate of the path. 7. Select the path with the lowest cost and return its first hop node as the path. α, β, and γ are weighting coefficients that can be dynamically adjusted according to the service type. The default values ​​are α=0.3, β=0.5, and γ=0.2. For latency-sensitive voice call services, the values ​​are adjusted to α=0.2, β=0.7, and γ=0.1. For bandwidth-sensitive video transmission services, the values ​​are adjusted to α=0.6, β=0.3, and γ=0.1.

[0063] When the quality degradation of an inter-satellite link exceeds a preset threshold (i.e., the available bandwidth decreases by more than 30% or the latency increases by more than 50%), a local route recalculation is triggered. The recalculation adopts an incremental approach: first, the set of destination addresses affected by the link is determined, and only the next hop is recalculated for these destination addresses, rather than updating the forwarding table of all destination addresses, which further reduces the computational overhead.

[0064] In this embodiment, the time taken for a single full route calculation is less than 10 milliseconds, and the time taken for incremental recalculation is less than 2 milliseconds.

[0065] To avoid routing loops, a split horizon mechanism is used: the routing information is not sent back from the interface that received the routing information; at the same time, the maximum number of hops is set to 10, and packets exceeding the number of hops are directly discarded.

[0066] S5, Dynamic Beam Resource Allocation.

[0067] Please refer to Figure 5 The satellite's dynamic resource allocation module monitors the load status of each beam in real time and periodically adjusts the resource allocation. This module includes a resource demand statistics unit 501, a demand factor calculation unit 502, a resource allocation calculation unit 503, and a beamforming control unit 504.

[0068] First, the resource demand statistics unit 501 counts the number of online users in each beam coverage area and classifies them by service type. The service types are divided into three categories: real-time conversation (voice calls, video conferencing), streaming media (live video streaming, video on demand), and background (web browsing, file transfer, IoT data), arranged in descending order of QoS priority. The statistics period is 1 second.

[0069] Then, the demand factor calculation unit 502 calculates the resource demand factor for each beam using the following formula: =w1× +w2× +w3× ,in , , , w1, w2, w3 are the number of users of the three types of services in the i-th beam, respectively, and w1, w2, w3 are the corresponding priority weights.

[0070] In this embodiment, typical values ​​are w1=3, w2=2, and w3=1. The weights decrease sequentially as the priority decreases, ensuring that high-priority services receive more resource allocation. The weight parameters can be configured remotely through the ground network management system.

[0071] Next, the resource allocation calculation unit 503 allocates the total power budget (200W) and the total bandwidth pool (500MHz) proportionally based on the sum of the demand factors of all 16 beams; that is, the power allocated to the i-th beam. =200× / Σ (Unit: Watts) Allocated bandwidth =500× / Σ (Unit: MHz)

[0072] To avoid frequent resource oscillations, a hysteresis mechanism is adopted for allocation adjustment: the relative change between the new allocation value and the current allocation value is calculated, and no adjustment is made if it is less than 10%. Only when it exceeds the threshold is a reallocation triggered. The allocation adjustment cycle is 1 second, which can respond to changes in hot spots in a timely manner and ensure system stability. Simulation shows that this hysteresis mechanism can reduce the number of resource adjustments by 70% and significantly reduce system overhead.

[0073] Finally, the beamforming control unit 504 controls the digital multi-beamformer to adjust the complex weighting coefficients of each beam according to the calculated power and bandwidth parameters of each beam, so as to realize the dynamic reconfiguration of resources. The power adjustment is achieved by adjusting the amplitude of the weighting coefficients, with an adjustment step size of 0.1dB. The bandwidth adjustment is achieved by adjusting the subcarrier allocation, with a granularity of 1 resource block (180kHz).

[0074] Through the aforementioned dynamic allocation mechanism, system resources are automatically tilted towards densely populated urban areas and high-priority services.

[0075] III. S6 Cross-Module Collaborative Control Mechanism Please refer to Figure 6 The core innovation of this invention lies in the cross-module collaborative control module 600, which connects the beam switching prediction module 601, the distributed routing module 602, and the dynamic resource allocation module 603 to achieve data sharing and collaborative decision-making. The collaborative mechanism includes three dimensions: time collaboration, target optimization, and end-to-end QoS.

[0076] First, time coordination.

[0077] After the beam switching prediction module 601 outputs the predicted switching time, the cross-module collaborative control module 600 sends a route pre-calculation notification message to the distributed routing module 602 one route update cycle (i.e., 500 milliseconds) in advance. The notification message format is: message type (1 byte, 0x10 indicates route pre-calculation notification), target satellite ID (2 bytes), and expected switching effective timestamp (4 bytes, millisecond precision). After receiving the notification, the distributed routing module 602 calculates the new path through the target satellite in advance and marks the new routing table entry as pending, setting it to take effect atomically and on time at the switching completion time.

[0078] Time synchronization uses satellite local clocks, and each satellite clock is periodically calibrated for time deviation via inter-satellite links, with a calibration accuracy better than 1 millisecond. The handover effective time uses an absolute timestamp method to ensure that route handover and beam handover are completed at the same time. In this way, the route also switches to the new path simultaneously at the moment the beam handover is completed, avoiding packet loss and detours caused by route inconsistency during the handover. Simulation shows that the time coordination mechanism can reduce the packet loss rate during handover from about 5% without coordination to below 0.5%, a reduction of more than 90%.

[0079] Second, target beam selection.

[0080] When the beam switching prediction module 601 initially generates the candidate beam sequence, it only considers signal quality and service duration. The cross-module collaborative control module 600 queries the dynamic resource allocation module 603 for the remaining resource quota information of each candidate beam. The query message carries the candidate beam ID list, and the return message carries the remaining available bandwidth and remaining power value of the corresponding beam.

[0081] After receiving resource information, the candidate sequence is reordered: first, beams with remaining bandwidth that can meet the peak demand of terminal services are selected as the first tier and sorted by signal quality; beams with insufficient remaining bandwidth but that can meet basic requirements are selected as the second tier; beams that cannot meet the requirements at all are removed from the candidate list. This avoids handover blocking or service degradation caused by switching to beams with insufficient resources. Simulation shows that the target optimization mechanism can reduce the handover blocking rate by more than 40%.

[0082] Third, QoS parameters are applied end-to-end.

[0083] Once the QoS level parameters of the service are determined during session establishment, they are simultaneously mapped and distributed to the three modules by the cross-module collaborative control module 600. --In the soft handover execution module, it is mapped to the handover priority parameter; high QoS level services get higher handover priority and start the preparatory link establishment more time in advance (40 seconds in advance for high priority, 30 seconds in advance for medium priority, and 20 seconds in advance for low priority), so as to ensure higher handover reliability for high priority services.

[0084] --In the distributed routing module 602, it is mapped to the routing cost weight parameter; high QoS level services use a higher latency weight coefficient, and the route selection is more inclined to low latency paths; medium QoS level uses the default weight; low QoS level uses a higher bandwidth weight, and prioritizes high bandwidth paths.

[0085] --In the dynamic resource allocation module 603, it is mapped to resource allocation weight parameters; that is, the aforementioned three-level weights w1, w2, and w3, with higher priority services receiving more resource quotas.

[0086] In this way, QoS assurance achieves end-to-end consistency from access switching, transmission routing to resource allocation, avoiding any link from becoming a QoS bottleneck. Simulation shows that end-to-end QoS implementation improves end-to-end latency consistency of high-priority services by 35% and reduces latency jitter by 40%.

[0087] IV. Anomaly Handling and Boundary Scenarios (1) Prediction failure handling: When the actual switching time deviates from the prediction time by more than 5 seconds, it is determined that the prediction has failed. The system will revert to the normal passive switching mode and record the deviation data for subsequent model optimization.

[0088] (2) Context synchronization failure: If incremental synchronization fails twice in a row, full context synchronization will be automatically triggered. The full data is about 50KB and will be transmitted in 5 time slots through the inter-satellite link.

[0089] (3) Route calculation error: When the local topology information is insufficient and the path cannot be calculated, the shortest path (by hop count) is used by default for forwarding, and the neighboring satellite is requested to supplement the topology information.

[0090] (4) Collaborative module failure: If the cross-module collaborative control module 600 malfunctions, the three major functional modules will automatically degrade to independent operation mode and work according to their local strategies to ensure that basic communication capabilities are not interrupted.

[0091] V. Simulation Verification and Performance Comparison To verify the technical effectiveness of this invention, a simulation platform was built using the STK satellite toolkit and the NS3 network simulator. The simulation parameters are as follows: constellation size of 72 satellites, orbital altitude of 550km, inclination of 53 degrees; 10,000 user terminals, distributed non-uniformly according to urban population density; service model of 30% voice, 40% video, and 30% background services; and the baseline scheme for comparison is the existing typical scheme of traditional hard handover + centralized routing + fixed resource allocation.

[0092] The main performance indicators are compared as follows: (1) Switching interrupt latency: The baseline solution is about 150 milliseconds, while the present invention is about 0.8 milliseconds, a reduction of 99.5%; (2) Switching packet loss rate: The baseline solution is about 8%, while the present invention is about 0.3%, a reduction of 96%; (3) Average end-to-end latency: The baseline scheme is about 45 milliseconds, while the present invention is about 34 milliseconds, a reduction of 24%; (4) Route recalculation frequency: The baseline scheme is about 2 times per second, while this invention is about 0.7 times per second, a reduction of 65%; (5) System throughput: The baseline solution is about 8Gbps per satellite, while this invention is about 10.8Gbps per satellite, an improvement of 35%; (6) Average latency of high-priority services: The baseline solution is about 52 milliseconds, while the present invention is about 32 milliseconds, a reduction of 38%.

[0093] In particular, to verify the synergistic effect of the collaboration mechanism, four scenarios were tested: predictive switching only, distributed routing only, dynamic resource allocation only, and all three plus collaboration. The results showed that the combined performance improvement of the three technologies when applied individually was about 30%, while the overall improvement reached 52% after adding the collaboration mechanism, with the additional 22% improvement coming from the synergistic effect. This fully demonstrates that the technical effect of 1+1+1>3 brought about by cross-module collaboration is something that those skilled in the art cannot easily expect based on existing technologies.

[0094] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A communication method for low-Earth orbit satellites, characterized in that, Includes the following steps: S1. Beam switching prediction: Obtain the location information and motion state parameters of the user terminal (101), combine the two-line element orbit prediction data of the low-orbit satellite (110) in orbit, calculate the position change curve of each visible low-orbit satellite (110) based on the SGP4 orbit propagation model, and generate a beam switching prediction sequence containing the target satellite identifier and the expected switching time. S2. Preparatory link establishment and incremental context synchronization: When the current service beam maintains communication connection and the signal-to-noise ratio is higher than the preset threshold, access is initiated to the target low-orbit satellite (110) in advance according to the beam switching prediction sequence and a preparatory beam link is established. The communication data context of the PDCP layer, ROHC layer and encryption layer is transmitted to the target low-orbit satellite (110) in the form of differential state variables through the inter-satellite link terminal (112). S3. Distributed Inter-Satellite Routing Calculation: Based on the distributed inter-satellite routing protocol, each low-orbit satellite (110) node periodically broadcasts link status messages to neighboring satellites. Each node only maintains a local topology view within a two-hop adjacency range and independently calculates the optimal path for data forwarding by combining the link quality time series prediction value. S4. Dynamic beam resource allocation: Based on the user density distribution and QoS level classification of each service within the coverage area, the transmit power and carrier bandwidth resources of the user beam array (111) are dynamically allocated proportionally using a weighted demand factor model. S5. Cross-module collaborative control: A cross-module collaborative mechanism is established. The beam switching prediction module (601) sends the predicted value of the switching time to the distributed routing module (602) in advance to trigger route pre-calculation. The dynamic resource allocation module (603) feeds back the remaining resource quota of each beam to the beam switching prediction module (601) to optimize target selection. The service QoS level parameters are simultaneously input into the three stages of soft handover priority scheduling, route cost weight calculation and resource allocation weight configuration.

2. The communication method for low-Earth orbit satellites according to claim 1, characterized in that, Step S1 specifically includes: calculating the instantaneous coordinates of each low-orbit satellite (110) in the geocentric-ground-fixed coordinate system within the prediction window based on the SGP4 model with a time step of 1 second; converting to the station-centered coordinate system through a coordinate transformation matrix, and calculating the azimuth, elevation, and slant range of the low-orbit satellite (110) relative to the user terminal (101) at each moment; determining that the satellite is visible when the elevation angle is greater than 10 degrees, and extracting the continuous visible time period as the service window; and generating a beam switching prediction sequence containing the target satellite ID, beam number, expected access time, and available bandwidth prediction value according to the start time of the service window, with the prediction window covering the next 30 minutes.

3. The communication method for low-Earth orbit satellites according to claim 2, characterized in that, Step S2 specifically includes: when the signal-to-noise ratio of the current serving beam is higher than 10dB, sending a preliminary link establishment request to the predicted next target low-Earth orbit satellite (110); establishing a preliminary data channel after the target low-Earth orbit satellite (110) completes access authentication; synchronizing an incremental context to the target low-Earth orbit satellite (110) through the inter-satellite link terminal (112), wherein the incremental context adopts the TLV encoding format and includes a PDCP sequence number differential mapping table, ROHC compressed context status number, encryption and integrity key count value, QoS flow tunnel endpoint parameters and unconfirmed PDU sequence number list, and the amount of data synchronized at one time is less than 10KB.

4. The communication method for low-Earth orbit satellites according to claim 3, characterized in that, Step S3 specifically includes: each low-orbit satellite (110) node broadcasts a link status message to all neighboring satellites at a period of 500 milliseconds. The message carries available bandwidth, transmission delay, bit error rate, and the link quality trend value predicted by the ARIMA model for the next period. Each node stores only the neighboring nodes and link information within a two-hop range to construct a local topology view. The forwarding table is calculated using a minimum cost routing algorithm with delay constraints. The cost function is Cost = α × (1 / available bandwidth) + β × delay + γ × bit error rate, where α, β, and γ are configurable weight coefficients. Incremental recalculation is triggered when the link quality deteriorates by more than 30%.

5. The communication method for low-Earth orbit satellites according to claim 4, characterized in that, Step S4 specifically includes: counting the number of users for each beam according to three priority levels: real-time session, streaming media, and background; and calculating the resource demand factor for each beam using the following formula: =w1× +w2× +w3× Where w1, w2, and w3 are the weight coefficients of the corresponding priorities, and the weight values ​​decrease sequentially as the priority decreases; the total power budget and total bandwidth pool are allocated according to the proportion of each beam demand factor to the total demand factor; a hysteresis threshold of 10% variation amplitude is introduced to suppress resource oscillation, and the resource adjustment cycle is 1 second.

6. The communication method for low-Earth orbit satellites according to claim 5, characterized in that, In step S5, the beam switching prediction module (601) sends the predicted switching time value to the distributed routing module (602) in advance to trigger route pre-calculation. Specifically, after generating the predicted switching time value, the beam switching prediction module (601) sends a pre-calculation notification to the distributed routing module (602) one route update cycle in advance. The notification carries the target satellite identifier and the expected switching effective time. The distributed routing module (602) calculates the new path through the target low-orbit satellite (110) in advance and sets the new routing table to take effect on time at the switching completion time.

7. The communication method for low-Earth orbit satellites according to claim 6, characterized in that, In step S5, the dynamic resource allocation module (603) feeds back the remaining resource quota of each beam to the beam switching prediction module (601) to optimize the target selection. Specifically, the dynamic resource allocation module (603) maintains the remaining available bandwidth and remaining power quota of each beam in real time. After generating the initial candidate beam sequence, the beam switching prediction module (601) queries the remaining resource quota of each candidate beam, prioritizes the beams that can meet the service bandwidth requirements of the user terminal (101), and downgrades the beams with insufficient resources to the alternatives.

8. A low-Earth orbit satellite communication system, characterized in that, include: The beam switching prediction module (601) is used to obtain the position and motion state parameters of the user terminal (101) and generate a beam switching prediction sequence based on the SGP4 model in combination with the orbit prediction data of the low-orbit satellite (110). The soft switching execution module is used to pre-establish a backup beam link during the current service beam communication and transmit the communication data context to the target low-orbit satellite (110) in the form of differential state variables through the inter-satellite link terminal (112). The distributed routing module (602) is deployed on each low-Earth orbit satellite (110) node to construct a local topology view within a two-hop adjacency range and independently calculate the optimal forwarding path based on the link quality prediction value. The dynamic resource allocation module (603) is used to dynamically allocate the power and bandwidth resources of the user beam array (111) based on the weighted demand factor model, according to the user density distribution and service QoS level. The cross-module collaborative control module (600) is used to send the handover time to the distributed routing module (602) in advance to trigger route pre-calculation, to feed back the beam resource quota to the beam handover prediction module (601) to optimize target selection, and to synchronously output the QoS level parameters to the three modules of soft handover scheduling, route calculation and resource allocation.

9. The low-Earth orbit satellite communication system according to claim 8, characterized in that, The cross-module collaborative control module (600) includes: The time coordination unit is used to trigger route pre-calculation one routing cycle in advance based on the handover prediction time, so that the routing table update and beam handover take effect synchronously. The target selection unit is used to reorder candidate handover targets based on the remaining resource quota of each beam; The QoS unified management and control unit is used to map service QoS levels to handover priority parameters, routing cost weight parameters, and resource allocation weight parameters, and distribute them to the corresponding modules.

10. The low-Earth orbit satellite communication system according to claim 8, characterized in that, The soft handover execution module includes: The preparatory link establishment unit is used to initiate the access procedure to the target low-orbit satellite (110) in advance when the current link signal-to-noise ratio is higher than the 10dB threshold; Incremental context synchronization unit, used to transmit session context in TLV-encoded differential state form via inter-satellite link terminal (112); The dual-link control unit is used to maintain parallel transmission of the two links during the handover transition period, and the downlink uses the maximum ratio combining method for reception.

Images