Satellite internet of things distributed node parallel transmission speed-up method

Abstract

The application discloses a satellite Internet of Things distributed node parallel transmission speed-up method, and relates to the technical field of satellite Internet of Things, and comprises the following steps: dynamic fragmentation, splitting into N fragments according to network load rate and node state, adding serial number, check code and redundant information, and the head identifier is {I_N_J}; intelligent routing, distributing fragments to distributed nodes through a scoring function based on satellite ephemeris, node load and link quality; dynamic path switching, monitoring link and node state, and switching to a better path when performance is not up to standard; fault-tolerant recombination, recombining according to serial numbers, and using error correction code to recover when there are few missing fragments, and retransmitting locally when there are many missing fragments. Through parallel transmission, dynamic switching and fault-tolerant recombination, the application improves transmission rate, throughput and robustness, balances node load, and optimizes the performance of satellite Internet of Things.

Description

Technical Field

[0001] This invention relates to the field of satellite Internet of Things (IoT) technology, and in particular to a method for accelerating parallel transmission of distributed nodes in satellite IoT. Background Technology

[0002] Satellite IoT provides critical connectivity for remote areas and oceans without terrestrial base station coverage, but current technologies face significant bottlenecks. While high-orbit satellites offer wide coverage, their orbital altitude of approximately 35,000 kilometers results in round-trip latency of 500-600 milliseconds, far exceeding the less than 20 milliseconds of fiber optic communication, severely limiting real-time performance and throughput. Even with low-orbit satellites reducing latency to 20-50 milliseconds, limited spectrum resources and transmission capacity still struggle to meet the demands of high-data-volume IoT applications such as video surveillance and remote sensing.

[0003] Traditional single-path serial transmission further exacerbates these problems, exhibiting low link utilization and susceptibility to single-point bottlenecks, failing to fully leverage the parallel capabilities of satellite and terrestrial networks. Existing solutions lack a collaborative mechanism for dynamic fragmentation and intelligent routing, unable to adjust transmission paths based on real-time link quality factors such as satellite ephemeris mobility and node load fluctuations. Furthermore, terminal nodes do not fully utilize edge computing capabilities for localized fragmentation encoding and redundancy control, resulting in limited transmission efficiency and fault tolerance. While comparative documents involve satellite network fragmentation or multi-path transmission, the former employs a static fragmentation strategy, neglecting dynamic link changes, while the latter fails to incorporate ephemeris prediction for route optimization, making it prone to packet loss due to frequent switching.

[0004] As the global scale of the Internet of Things (IoT) expands, the demand for satellite IoT transmission speed and reliability continues to increase, highlighting the shortcomings of existing technologies in areas such as bandwidth aggregation, load balancing, and dynamic adaptability. Therefore, developing a transmission algorithm that can utilize multiple nodes and links in parallel, combined with dynamic fragmentation, intelligent routing, and fault tolerance mechanisms, has become crucial for overcoming the transmission bottlenecks of satellite IoT. Summary of the Invention

[0005] This invention proposes a method for accelerating parallel transmission of distributed nodes in satellite Internet of Things (IoT) to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things includes the following steps: Dynamic fragmentation: Calculate the fragment size dn based on the current network load rate L and the real-time status of the nodes, and split the data packet into N data fragments and dynamically adjust them according to the rules: Each fragment is given a unique sequence number, a 16-bit CRC check code and erasure code redundancy information, and the header identifier is {I_N_J}, where N is the number of data fragments and J is the current fragment sequence number; Intelligent routing: Node k visibility time window based on satellite ephemeris prediction Real-time load rate Link quality signal-to-noise ratio mapping value Shards are allocated to distributed nodes using a scoring function; first, data for the next 120 seconds is retrieved from the ephemeris database. CPU load rate of data acquisition nodes and uplink signal-to-noise ratio (SNR), mapping SNR to =max(0,min(1,(SNR+10) / 30)), then calculate the score according to the weight coefficient configured by business type, and prioritize the allocation to the top M nodes (M=min(5,number of available nodes)); Dynamic path switching: The source node or relay node continuously monitors the link quality Q, node status load L, and remaining energy E. If it detects a surge in packet loss rate, a sharp increase in latency, excessive node load, or low energy on the current path, and there is a better backup path based on the updated score, then path switching is triggered before the fragment transmission is completed. Fault-tolerant reassembly: The receiving end reassembles fragments according to their sequence numbers, using RS(N,K) erasure coding, where K is the original number of fragments, R=NK is the number of redundant fragments, and R / K≤25%. When the number of missing fragments is ≤R, the data is directly restored; otherwise, a partial retransmission is initiated to the two nodes with the highest scores. The retransmitted fragment header is identified as {I_N_J_Rx}, where x is the retransmission count counter, initially x=1. Furthermore, the scoring function for intelligent routing is as follows: ,in This represents the real-time load rate of node k. This is the link quality signal-to-noise ratio mapping value; The visible time window for node k; This represents the total end-to-end transmission delay. The percentage of remaining energy at node k; For path packet loss rate; The storage and forwarding potential score for node k is given; α, β, γ, δ, ζ, and η are adjustable non-negative weight coefficients, which are adjusted to adapt to real-time and delay-tolerant services.

[0007] Furthermore, it also includes parallel upload optimization: each distributed node initiates upload on an independent frequency or time slot to avoid signal interference; during the upload process, the link status is monitored in real time, and erasure coding verification or retransmission is enabled for erroneous fragments. When RS(10,8) encoding is used, the 8 original fragments are expanded to 10, including 2 redundant fragments.

[0008] Furthermore, the weighting coefficients for real-time services are configured as follows: α=0.3, β=0.3, γ=0.2, δ=0.1, ζ=0.1, η=0, prioritizing low latency and link quality.

[0009] Furthermore, the weighting coefficients for delay-tolerant services are configured as follows: α=0.2, β=0.2, γ=0.1, δ=0.1, ζ=0.1, η=0.3, with S being the preferred choice. k Nodes with a speed greater than 0.8 perform store-and-forward operations to improve transmission robustness.

[0010] Furthermore, the fragment size dn in dynamic fragmentation is determined according to the following rules: when the average free memory of the node is >512MB, dn=1MB; when 256MB<free memory≤512MB, dn=0.5MB; when free memory≤256MB, dn=0.256MB, balancing the node storage pressure and transmission efficiency.

[0011] Furthermore, in dynamic path switching, the expected thresholds for path performance include: packet loss rate > 10%, latency > 500ms, node load rate > 80% or remaining energy < 20%. Meeting any one of these conditions will trigger a switching assessment.

[0012] Furthermore, in fault-tolerant reassembly, the receiving end sets a timeout period for fragment reception. =2× ,in For the predicted total end-to-end transmission delay, fragments that are not received within the timeout period are considered missing, and a recovery or retransmission process is initiated.

[0013] Furthermore, the storage-forwarding potential score of node k. The calculation formula is: =0.6×(available storage space / total storage space)+0.4×(number of connection opportunities in the next 24 hours / 10), comprehensively evaluate storage capacity and connection stability.

[0014] Furthermore, the system updates node state load L, remaining energy E, and link quality Q every 30 seconds, and updates satellite ephemeris predictions every 5 minutes. This ensures the real-time nature and accuracy of routing decisions.

[0015] Compared with existing technologies, the beneficial effects of this invention are: The satellite IoT distributed node parallel transmission speed-up method proposed in this invention significantly optimizes transmission performance through multi-dimensional technological innovation. In terms of transmission efficiency, by splitting data into multiple fragments and transmitting them in parallel through distributed nodes, bandwidth aggregation and concurrent transmission are achieved. Compared with the traditional single-path serial method, this significantly improves the overall data throughput and transmission rate, better meeting the needs of high-data-volume applications.

[0016] In terms of reliability and robustness, the algorithm introduces a dynamic path switching mechanism, which can monitor link quality and node status in real time. When a performance degradation of the current path is detected, it automatically switches to a better backup path, reducing transmission interruptions caused by single points of failure or link fluctuations. At the same time, it adopts erasure coding redundancy technology and a partial retransmission strategy. Even if the transmission of some fragments is blocked, data can be recovered through redundant information or only the missing parts can be retransmitted, which significantly improves the integrity and fault tolerance of data transmission.

[0017] Furthermore, the intelligent routing strategy combines satellite ephemeris prediction, node load, and link quality for scoring, achieving reasonable fragmentation allocation, effectively balancing the load of each node, and avoiding single-point overload issues. The dynamic fragmentation mechanism can adjust the size and number of fragments based on network load and satellite visibility windows, enhancing adaptability to dynamic network environments. The overall solution fully utilizes the collaborative capabilities of distributed nodes, improving satellite IoT transmission performance while providing strong support for stable communication in complex environments. Attached Figure Description

[0018] Figure 1 This is a schematic block diagram of a satellite Internet of Things distributed node parallel transmission speed-up method proposed in this invention; Figure 2 Line chart comparing the number of shards under different network loads; Figure 3 A bar chart comparing transmission performance before and after path switching; Figure 4 Line graph showing packet recovery success rate under different redundancy ratios; Figure 5 A pie chart comparing the node scoring weights for real-time and delay-tolerant services. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0022] Reference Figures 1 to 5 A method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things (IoT) includes the following steps: Dynamic fragmentation: The fragment size dn is calculated based on the current network load rate L and the real-time status of the nodes. The data packet is split into N data fragments and dynamically adjusted according to the following rules: When L < 50%, N = [S / 1] (S is the data packet size in MB); when L ≥ 50% or the shortest satellite visibility window... When the time is less than 60 seconds, N = min([S / 0.256], 32). Each fragment is given a unique sequence number, a 16-bit CRC checksum, and erasure coding redundancy information. The header identifier is {I_N_J} (where N is the number of data fragments and J is the current fragment sequence number).

[0023] Intelligent routing: Node k visibility time window based on satellite ephemeris prediction Real-time load rate Link quality signal-to-noise ratio mapping value Shards are allocated to distributed nodes using a scoring function. First, the ephemeris database is retrieved for the next 120 seconds. CPU load rate of data acquisition nodes and uplink signal-to-noise ratio (SNR), mapping SNR to =max(0,min(1,(SNR+10) / 30)), then calculate the score according to the weight coefficient configured by business type, and prioritize the allocation to the top M nodes (M=min(5,number of available nodes)).

[0024] Dynamic path switching: The source node or relay node continuously monitors the link quality Q and node status (load L, remaining energy E). If a surge in packet loss rate, a sharp increase in latency, excessive node load, or low energy is detected on the current path, and there is a better backup path based on the updated score, then path switching is triggered before the fragment transmission is completed.

[0025] Fault-tolerant reassembly: The receiving end reassembles fragments according to their sequence numbers, using RS(N,K) erasure coding (K is the original number of fragments, R=NK is the number of redundant fragments, R / K≤25%). When the number of missing fragments is ≤R, the data is directly restored; otherwise, a partial retransmission is initiated to the two nodes with the highest scores. The retransmission fragment header is identified as {I_N_J_Rx} (x is the retransmission count counter, initially x=1).

[0026] This invention also includes: a scoring function for intelligent routing: ,in The real-time load rate (0~1) of node k. This is the link quality signal-to-noise ratio mapping value (0~1); The visible time window (in seconds) for node k; The total end-to-end transmission delay (including queuing, propagation, and transmission delays, in seconds); The percentage of remaining energy at node k (0~1); The path packet loss rate (0~1); The storage and forwarding potential score for node k is 0~1; α, β, γ, δ, ζ, and η are adjustable non-negative weight coefficients, which are adjusted to adapt to real-time and delay-tolerant services.

[0027] This invention also includes: parallel upload optimization: each distributed node initiates upload at an independent frequency or time slot to avoid signal interference; during the upload process, the link status is monitored in real time, and erasure coding verification or retransmission is enabled for erroneous fragments. When RS(10,8) encoding is used, the 8 original fragments are expanded to 10 (including 2 redundant fragments).

[0028] The present invention also includes: the weighting coefficients of real-time services are configured as follows: α=0.3, β=0.3, γ=0.2, δ=0.1, ζ=0.1, η=0, to prioritize low latency and link quality.

[0029] This invention also includes: the weighting coefficients for delay-tolerant services are configured as follows: α=0.2, β=0.2, γ=0.1, δ=0.1, ζ=0.1, η=0.3, with priority given to... Nodes with a speed greater than 0.8 perform store-and-forward operations to improve transmission robustness.

[0030] In this invention, the fragment size dn in dynamic fragmentation is determined according to the following rules: when the average free memory of the node is >512MB, dn=1MB; when 256MB<free memory≤512MB, dn=0.5MB; when free memory≤256MB, dn=0.256MB, thus balancing the node storage pressure and transmission efficiency.

[0031] This invention also includes: in dynamic path switching, the expected threshold for path performance includes: packet loss rate > 10%, latency > 500ms, node load rate > 80% or remaining energy < 20%, and a switching evaluation is triggered when any one of these conditions is met.

[0032] This invention also includes: in fault-tolerant reassembly, the receiving end sets a fragment reception timeout period. =2× ( (The predicted total end-to-end transmission delay) Fragments that are not received within the timeout period are considered missing, and a recovery or retransmission process is initiated.

[0033] This invention also includes: the store-and-forward potential score of node k. The calculation formula is: =0.6×(available storage space / total storage space)+0.4×(number of connection opportunities in the next 24 hours / 10), comprehensively evaluate storage capacity and connection stability.

[0034] This invention also includes: the system updating node status (L, E) and link quality Q every 30 seconds, and updating satellite ephemeris predictions every 5 minutes. This ensures the real-time nature and accuracy of routing decisions.

[0035] Example: System Architecture and Initialization This algorithm is applied to satellite IoT systems that include edge gateway terminals, satellite direct-connect terminals, and distributed nodes. Nodes are deployed in gateway or terminal devices and support ephemeris data acquisition, link quality monitoring, and fragmentation processing. During system initialization, node identifiers (unique IDs), available link parameters (frequency points, time slots), transmission priorities (real-time / delay tolerance) need to be configured. The system also needs to access a satellite orbit database (providing satellite visibility time predictions for the next 120 seconds), a local link quality history database (storing the average SNR over the past 10 minutes), a node energy monitoring module (sampling frequency 1Hz, accuracy ±1%), and a dynamic weighting strategy table (preset coefficient templates corresponding to the service types).

[0036] When the source terminal (such as an IoT device) sends a data packet, the algorithm first performs dynamic fragmentation: Fragment number calculation: based on the data packet size S (in MB), average network load rate L (average CPU utilization of all nodes in the network), and the shortest satellite visibility window. (corresponding to all nodes) Minimum value) determines the number of fragments N: If L < 50%, N = ceil(S / 1) (rounded up, e.g., N = 3 when S = 2.3MB); if L ≥ 50% or <60s, N=min(ceil(S / 0.256),32) (maximum limit of 32 fragments to avoid excessive fragmentation and reassembly overhead). Fragmentation processing: Split data packets according to the calculation results. The size of each fragment is dn=S / N (error ≤5%). Add a 16-bit CRC checksum (polynomial 0x1021) and RS erasure code redundancy (coding efficiency K / N≥80%). In the header identifier {I_N_J}, N is represented by 8 bits and J is represented by 16 bits (supports a maximum of 256 fragments). For example, the 3rd fragment of 10 fragments is identified as {0A_03} (hexadecimal).

[0037] Intelligent routing decision parameter acquisition: Obtain the visibility time of each node for the next 120 seconds from the ephemeris database. CPU load rate of data acquisition nodes (5s moving average), uplink SNR (sampled every 2s, averaged 10 times), mapped to Q k =max(0,min(1,(SNR+10) / 30)) (when SNR=-10dB) =0, 20dB =1); calculate end-to-end delay =Queueing delay (based on (Estimation) + Propagation delay (satellite distance / speed of light) + Transmission delay (fragment size / link rate); Obtain the remaining energy of the node. (Current battery level / Rated battery level), Historical packet loss rate (Packet loss percentage of nearly 100 data packets), store-and-forward potential (Available storage space percentage × 0.6 + Number of connections in the next 24 hours / 10 × 0.4). Scoring calculation:

[0038] The weights for real-time services are α=0.3, β=0.3, γ=0.2, δ=0.1, ζ=0.1, and η=0; the weights for delay-tolerant services are α=0.2, β=0.2, γ=0.1, δ=0.1, ζ=0.1, and η=0.3. Sharding allocation: Shards are preferentially allocated to... The top M nodes (M = min(5, number of available nodes)) have the following number of shards allocated to each node: Positive correlation (e.g., the node with the highest score is allocated 20% of the shards).

[0039] During the dynamic path switching mechanism, the source node monitors link quality and node status every 500ms: Link quality metrics: packet loss rate (the percentage of 10 consecutive fragments lost), latency (the time difference between fragment transmission and acknowledgment); Node status metrics: load rate. Remaining energy Triggering conditions: Packet loss rate > 10%, latency > 500ms, L k >80% or <20%, and there are spare nodes The current node is more than 10% higher than the current node; the switching process is as follows: mark the current fragment transmission interrupted, reallocate the unfinished part to the standby node, and the new node uses the original fragment identifier but adds a switching mark (such as {I_N_J_S}, where S is the number of switching).

[0040] Parallel Upload and Fault-Tolerant Reassembly: Parallel Upload: Each node uploads fragments on an independent frequency (interval ≥ 50kHz) or time slot (allocated according to TDMA mechanism). Before uploading, the integrity of the locally cached fragments is verified (CRC check). If an error occurs, a local redundant copy is used for retransmission. When using RS(10,8) encoding, 8 original fragments correspond to 2 redundant fragments, ensuring recovery when ≤ 2 fragments are lost in a single link. Fault-Tolerant Reassembly: The receiver maintains a fragment reception table, sorted by J sequence number, and sets a timeout period. =2× (like =200ms =400ms); if the number of missing fragments ≤ R (number of redundant fragments), recover the complete data using the RS decoder; if the number of missing fragments > R, send to The two highest nodes send retransmission requests, with the retransmission fragment identifier being {I_N_J_Rx} (x=1,2...), and the number of retransmissions is limited to ≤3. After reassembly, the overall CRC check is calculated, and if it passes, it is delivered to the application layer; otherwise, a full retransmission is triggered (only for the scenario of initial reassembly failure).

[0041] The system dynamically updates the node status. , ) and link quality ( , Updated every 30 seconds; satellite ephemeris data ( Updated every 5 minutes to ensure accurate prediction of the visible time window; the weight coefficients are finely adjusted every hour based on the overall network performance (throughput, packet loss rate) (fluctuation range ≤10%).

[0042] Effect data characterization and interpretation Average throughput lower Significant improvement Transmission delay longer Significantly shortened Packet loss recovery success rate lower Significantly improved Node load balancing Poor Significant improvement The above data is based on comparative tests under the same satellite link conditions (3 high-orbit satellites and 5 distributed nodes). The improvement in average throughput stems from bandwidth aggregation through parallel transmission across multiple nodes, fully utilizing the idle resources of each link; the reduction in transmission latency is attributed to the intelligent routing's priority selection of low-latency paths and the reduced waiting time due to fragmented parallel processing; the improved packet loss recovery success rate reflects the synergistic effect of RS erasure coding and partial retransmission, enabling recovery through redundant information even when multiple links experience simultaneous packet loss; the improved node load balancing is attributed to the scoring function's distribution of traffic to high-load nodes, avoiding performance bottlenecks caused by single-point overload. Overall, this algorithm comprehensively improves the transmission performance of satellite IoT through a distributed parallel mechanism and dynamic optimization strategy.

[0043] The above are merely preferred embodiments 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 method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things, characterized in that, Includes the following steps: Dynamic fragmentation: Calculate the fragment size dn based on the current network load rate L and the real-time status of the nodes, and split the data packet into N data fragments: when L < 50%, N = [S / 1]; when L ≥ 50% or the shortest satellite visibility window... When the time is less than 60 seconds, N = min([S / 0.256], 32), where S is the data packet size in MB; each fragment is given a unique sequence number, a 16-bit CRC checksum, and erasure coding redundancy information, and the header identifier is {I_N_J}, where N is the number of data fragments and J is the current fragment sequence number; Intelligent routing: Node k visibility time window based on satellite ephemeris prediction CPU real-time load rate of node k Link quality signal-to-noise ratio mapping value Shards are distributed to distributed nodes using a scoring function: first, the ephemeris database is retrieved for the next 120 seconds. Collect the real-time CPU load rate of node k. and uplink signal-to-noise ratio (SNR), mapping SNR to =max(0,min(1,(SNR+10) / 30)), then calculate the score according to the weight coefficient configured according to the business type, and prioritize the allocation to the top M nodes, where M=min(5,number of available nodes); Parallel upload optimization: Each distributed node initiates upload on an independent frequency point or time slot; during the upload process, the link status is monitored in real time, and erasure coding verification or retransmission is enabled for erroneous fragments. When RS(10,8) encoding is used, the 8 original fragments are expanded to 10, including 2 redundant fragments. Dynamic path switching: The source node or relay node continuously monitors the link quality Q, node status load L, and remaining energy E. If the current path packet loss rate, latency, node load exceeds the threshold or energy is below the threshold, and there is a backup path based on the updated score, then path switching is triggered before the fragment transmission is completed. Fault-tolerant reassembly: The receiving end reassembles fragments according to their sequence numbers, using RS(N,K) erasure coding, where K is the original number of fragments and R=NK is the number of redundant fragments. When the number of missing fragments is ≤R, the data is directly restored; otherwise, a partial retransmission is initiated to the two nodes with the highest scores. The retransmission fragment header is identified as {I_N_J_Rx}, where x is the retransmission count counter, initially x=1.

2. The method for accelerating parallel transmission of distributed nodes in satellite IoT according to claim 1, characterized in that, The scoring function for intelligent routing is: in, The real-time CPU load rate of node k; This is the link quality signal-to-noise ratio mapping value; The visible time window for node k; This represents the total end-to-end transmission delay. The percentage of remaining energy at node k; For path packet loss rate; The storage-forwarding potential score for node k; α, β, γ, δ, ζ, η are adjustable non-negative weight coefficients, which are adjusted to adapt to real-time and delay-tolerant services; the storage-forwarding potential score for node k. The calculation formula is: =0.6×(Available storage space / Total storage space)+0.4×(Number of connection opportunities in the next 24 hours / 10).

3. The method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things according to claim 2, characterized in that, The weighting coefficients for real-time services are configured as follows: α=0.3, β=0.3, γ=0.2, δ=0.1, ζ=0.1, η=0.

4. The method for accelerating parallel transmission of distributed nodes in satellite IoT according to claim 2, characterized in that, The weighting coefficients for delay-tolerant services are configured as follows: α=0.2, β=0.2, γ=0.1, δ=0.1, ζ=0.1, η=0.3, with priority given to... Nodes with a speed greater than 0.8 perform store-and-forward operations.

5. The method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things according to claim 1, characterized in that, The shard size dn in dynamic sharding is determined according to the following rules: when the average free memory per node is > 512MB, dn = 1MB; when 256MB < average free memory per node ≤ 512MB, dn = 0.5MB. When the free memory is ≤256MB, dn=0.256MB.

6. The method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things according to claim 1, characterized in that, In dynamic path switching, the expected threshold for path performance includes: path switching is triggered when any one of the following conditions is met: packet loss rate > 10%, latency > 500ms, node load rate > 80%, or remaining energy < 20%. The latency is the time difference between fragment sending and confirmation.

7. The method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things according to claim 1, characterized in that, In fault-tolerant reassembly, the receiving end sets a timeout period for fragmented reception. =2× ,in For the predicted total end-to-end transmission delay, fragments that are not received within the timeout period are considered missing, and a recovery or retransmission process is initiated.

8. The method for accelerating parallel transmission of distributed nodes in a satellite Internet of Things according to claim 1, characterized in that, The node status load (L), remaining energy (E), and link quality (Q) are updated every 30 seconds, and the satellite ephemeris predictions are updated every 5 minutes. .

Images