Multi-source multi-path scheduling method for remote sensing data transmission

Abstract

The invention relates to the technical field of satellite network communication, and discloses a multi-source multi-path scheduling method for remote sensing data transmission, which comprises the following steps of: firstly, establishing an index table for remote sensing data by using a space-time name formed by a service type, an acquisition area, acquisition time, a node identifier and a data category; and the controller analyzes time and space constraints in the request, selects a plurality of source nodes meeting conditions from an index table, calculates priorities according to parameters such as residual bandwidth, round-trip delay and packet loss rate, and selects a plurality of non-overlapping high-quality paths. And then adding a service type, a path number, a path state report and a flow proportion field in the interest packet and the data packet, so that the source node sends segmented data in parallel according to a preset proportion. And the forwarding node reports the link change in real time, and the controller dynamically adjusts the path set and the shunt weight, thereby ensuring that the high-bandwidth and low-delay data transmission performance is continuously obtained in a topology and load rapid change scene.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite network communications, in particular to a multi-source and multi-path scheduling method for remote sensing data transmission. Background Art

[0002] Since the early days of TCP / IP-based satellite file relay, remote sensing data transmission has evolved from a serial architecture of "single-satellite direct link - ground queuing - centralized downlink" to disjointed forwarding using the DTN and CCSDS protocols, and more recently, to a content-driven architecture that integrates Named Data Networking (NDN) and Software-Defined Networking (SDN). NDN significantly reduces satellite-to-ground backhaul latency and increases link reuse through name-based routing and mid-transit caching; SDN, through control / forwarding separation, provides centralized intelligence for link orchestration and resource coordination in dynamic topologies. Simultaneously, constellation sizes have evolved from a dozen satellites to thousands of stars, with distributed caching nodes distributed across various platforms, including satellites, drones, and edge gateways. This has made multi-source, multi-path, parallel transmission a new trend in high-resolution remote sensing services. The academic community has proposed solutions such as ICN-Satellite, Map-NDN, and Hop-by-Hop Congestion Aware Multipath, attempting to utilize multiple replicas and link diversity. However, existing designs are generally limited to static directories or pure link indicator scheduling, lack explicit modeling of the "spatiotemporal coupling" characteristics of remote sensing data, and fail to natively support multipath traffic partitioning and link status feedback at the protocol layer.

[0003] Existing technologies suffer from the following major shortcomings: First, traditional NDN naming often uses business-layer concepts or hash digests as keys, ignoring the time and spatial scope of data collection. This makes it difficult to quickly identify versions and coverage areas for similar images, resulting in coarse index granularity and requiring multi-hop queries. Second, distributed caches only record "presence / absence" status and lack binding to node location or observation time, making it difficult for the controller to accurately select data sources that meet specific task windows. Third, multi-source parallel mechanisms often adopt a preemptive "first responder kills" strategy, failing to balance load based on bandwidth, latency, and packet loss. Protocol extensions are also lacking to identify paths and split ratios at the packet level, making it difficult to instantly redistribute traffic when link quality fluctuates. Fourth, link status monitoring typically relies on external SNMP / NetFlow or offline statistics from ground-based measurement and control stations, lacking real-time and coverage, and unable to support fine-grained scheduling in dynamic satellite-ground topologies. In summary, in highly dynamic, massively replicated, and bandwidth-heterogeneous air-space-ground-sea integrated environments, an integrated solution that simultaneously integrates "precise spatiotemporal indexing + multi-source and multi-path awareness + protocol-level traffic mapping + online adaptive scheduling" remains lacking. Summary of the Invention

[0004] (1) Technical Problems Solved: In response to the deficiencies of the prior art, the present invention provides a multi-source and multi-path scheduling method for remote sensing data transmission, which solves the above-mentioned problems.

[0005] (2) Technical solution: To achieve the above-mentioned purpose, the present invention provides the following technical solution: a multi-source multi-path scheduling method for remote sensing data transmission, the method comprising: S1: the controller uses a unified spatiotemporal naming structure to generate a unique name for each remote sensing data, and based on the node information, collection coverage area, and collection timestamp of the remote sensing data, writes the unique name and triple structure into the spatiotemporal replica index table.

[0006] S2: When receiving an interest packet, the controller extracts the time field and space field therein, and obtains a multi-source node set based on the time-space replica index table, and sorts the multi-source node set according to the link state matrix to obtain a priority node list; the link state matrix includes bandwidth, round-trip delay, and packet loss rate.

[0007] S3: The controller selects multiple non-overlapping paths based on the priority node list, generates an extended interest packet with service type, path number, path status and traffic ratio fields, and sends it to each source node.

[0008] S4: The forwarding node continuously reports the remaining bandwidth, round-trip delay, and packet loss rate through extended interest packets. The controller calculates the path fitness score based on link utilization and comprehensive delay, and allocates the traffic ratio of each path according to the path fitness score ratio.

[0009] S5: According to the multi-source node list and path traffic mapping issued by the controller, each source node sends the remote sensing data segment to the user end in parallel via the corresponding path based on the segment number; during the transmission process, the forwarding node continuously reports link changes, and the controller adjusts the path set and traffic ratio in real time to ensure overall transmission bandwidth and latency performance.

[0010] Furthermore, the S1 specifically includes: S11: the controller identifies the source node based on the remote sensing data , business type , Collection coverage area , collection timestamp , generated or cached node information and data types , generate a unique name for each remote sensing data.

[0011] S12: The unique name and triplet The structure is written to the time-space copy index table.

[0012] Furthermore, the S2 specifically includes: S21: when receiving the interest packet, the controller extracts the time field in the interest packet and spatial fields .

[0013] S22: Traverse each triple in the spatiotemporal replica index table ,when and Coverage area includes When the node information Join the candidate node set .

[0014] S23: For the candidate node set Each node in , the controller is based on the link state matrix Calculate the overall priority score ;in, are the round-trip delay, bandwidth and packet loss rate of the shortest path from the node to the user, is the preset weight constant.

[0015] S24: Press Sort from small to large, select the first nodes as the priority node list.

[0016] Furthermore, the S4 specifically includes: S41: the forwarding node reports the remaining bandwidth at a preset period , round-trip delay and packet loss rate .

[0017] S42: The controller calculates the link utilization using the following formula: : ;in, Indicates the currently occupied bandwidth. Indicates the upper limit of the physical bandwidth.

[0018] S43: The controller calculates the comprehensive delay using the following formula : ;in, Indicates the transmission delay, represents the propagation delay, represents the queuing delay, Indicates processing delay.

[0019] S44: The controller calculates the path fitness score using the following formula : ;in, is the adjustment weight, and .

[0020] S45: The controller allocates the flow ratio of each path according to the path fitness score ratio using the following formula: ;in, The bandwidth requirements requested for the user task.

[0021] Furthermore, the S5 specifically includes: S51: the controller numbers the remote sensing data Split and generate a path traffic mapping table.

[0022] S52: The source node adds a path number to the local data segment according to the mapping table, and sends the data packet in sequence via the corresponding path. The forwarding node forwards the data packet according to the path number.

[0023] S53: During the forwarding process, the forwarding node periodically encapsulates the link remaining bandwidth, round-trip delay, and packet loss rate into the extended interest packet.

[0024] S54: If the controller detects that the bandwidth of a certain path decreases or the delay exceeds a threshold, it recalculates the path fitness score and traffic ratio and updates the path traffic mapping table.

[0025] (III) Beneficial Effects: Compared with the prior art, the present invention provides a multi-source multi-path scheduling method for remote sensing data transmission, which has the following beneficial effects: 1. The multi-source multi-path scheduling method for remote sensing data transmission introduces the "acquisition area loc + acquisition timestamp time" secondary field in the NDN naming layer and writes the triplet synchronously. The index table enables precise positioning and version identification of image copies within the same mission window within seconds, avoiding the broadcast overhead of traditional "directory-hash" double-hop queries, significantly reducing index traversal latency and improving the retrieval efficiency of the air-ground fusion network.

[0026] 2. The multi-source multi-path scheduling method for remote sensing data transmission adopts bandwidth-delay-packet loss rate weighted scoring on the controller side. and press Traffic weight calculated proportionally The PathId and L_i fields are added to the interest packets and data packets, thereby realizing link quality-aware load balancing among multi-source nodes: high-score paths receive more diversion and low-score paths are automatically loaded down, which not only overcomes the single-source bottleneck caused by "first come, first served" but also ensures throughput and fairness under heterogeneous link conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Schematic diagram of the overall process of the method of the present invention.

[0028] Figure 2 Schematic diagram of the timing interaction of the method of the present invention.

[0029] Figure 3 Schematic diagram of the three-layer architecture of the method of the present invention.

[0030] Figure 4 Schematic diagram of the basic components of the architecture of the method of the present invention.

[0031] Figure 5 Schematic diagram of the packet format of the method of the present invention.

[0032] Figure 6 Schematic diagram of the flow distribution ratio of the method of the present invention.

[0033] Figure 7 Schematic diagram of the first experimental scenario of the method of the present invention.

[0034] Figure 8 Schematic diagram of the second experimental scenario of the method of the present invention.

[0035] Figure 9 This is a diagram showing the effect of the interest packet sending frequency on latency according to the method of the present invention.

[0036] Figure 10 This is a throughput effect diagram of the multi-source and multi-path scenario of the method of the present invention.

[0037] Figure 11 This is a throughput comparison diagram of different scheduling algorithms of the method of the present invention.

[0038] Figure 12 This is a diagram showing the download volume effects of different ground stations using different scheduling strategies according to the method of the present invention. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] In order to enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0041] See also Figure 1-6 , Figure 1 Schematic diagram of the overall process of the method of the present invention; Figure 2 Schematic diagram of the timing interaction of the method of the present invention; Figure 3 Schematic diagram of the three-layer architecture of the method of the present invention; Figure 4 A schematic diagram of the basic components of the architecture of the method of the present invention; Figure 5Schematic diagram of the package format of the method of the present invention; Figure 6 Schematic diagram of the traffic distribution ratio of the method of the present invention; the present invention provides a multi-source multi-path scheduling method for remote sensing data transmission, the method comprising: S1: the controller generates a unique name for each remote sensing data using a unified spatiotemporal naming structure, and writes the unique name and triple structure into the spatiotemporal replica index table based on the node information, acquisition coverage area, and acquisition timestamp of the remote sensing data.

[0042] S2: When receiving an interest packet, the controller extracts the time field and space field therein, and obtains a multi-source node set based on the time-space replica index table, and sorts the multi-source node set according to the link state matrix to obtain a priority node list; the link state matrix includes bandwidth, round-trip delay, and packet loss rate.

[0043] S3: The controller selects multiple non-overlapping paths based on the priority node list, generates an extended interest packet with service type, path number, path status and traffic ratio fields, and sends it to each source node.

[0044] S4: The forwarding node continuously reports the remaining bandwidth, round-trip delay, and packet loss rate through extended interest packets. The controller calculates the path fitness score based on link utilization and comprehensive delay, and allocates the traffic ratio of each path according to the path fitness score ratio.

[0045] S5: According to the multi-source node list and path traffic mapping issued by the controller, each source node sends the remote sensing data segment to the user end in parallel via the corresponding path based on the segment number; during the transmission process, the forwarding node continuously reports link changes, and the controller adjusts the path set and traffic ratio in real time to ensure overall transmission bandwidth and latency performance.

[0046] Furthermore, in order to quickly and accurately locate remote sensing data copies in an integrated environment of air, space, land and sea, it is necessary to give each piece of data a unique name that can carry spatial and temporal semantics, that is, through the "collection coverage area" " and "Collection Timestamp " and other spatiotemporal attributes to construct corresponding naming entries; in the naming phase, the controller identifies the source node based on Business Type , Collection coverage area , collection timestamp , generated or cached node information and data types Generate names for remote sensing data and use triples Write the spatiotemporal replica index table. This naming-indexing design can achieve retrieval of multiple replicas of data while ensuring global uniqueness of the naming, that is, it can quickly return the matching source node set when facing large-scale constellations and edge caches; and because the index is only maintained and updated by the controller, and the other nodes only hold local cache records, external entities cannot forge or tamper with the spatiotemporal mapping, thereby ensuring the security of the retrieval results. The S1 specifically includes: S11: The controller identifies the source node based on the remote sensing data , business type , Collection coverage area , collection timestamp , generated or cached node information and data types , generate a unique name for each remote sensing data.

[0047] S12: The unique name and triplet The structure is written to the time-space copy index table.

[0048] Furthermore, in order to ensure that user requests can be quickly routed to the most appropriate data replica in a highly dynamic remote sensing network, it is necessary to complete the two-level screening of "time-space matching + link quality assessment" immediately when the interest packet arrives; therefore, when the controller receives the interest packet, it first extracts the time field and spatial fields , and retrieve the matching and Coverage area includes A collection of multiple source nodes ; Then, the controller maintains a link state matrix in real time, whose elements are composed of bandwidth, round-trip delay and packet loss rate, which is a set Calculate the comprehensive priority of each node in , and generates a priority node list from low to high priority. Therefore, the selected node fits the user task window in time and space, and also eliminates the risk of congestion or delay surge caused by the difference in link performance between nodes through unified measurement. The S2 specifically includes: S21: When receiving the interest packet, the controller extracts the time field in the interest packet and spatial fields .

[0049] S22: Traverse each triple in the spatiotemporal replica index table ,when and Coverage area includes When the node information Join the candidate node set .

[0050] S23: For the candidate node set Each node in , the controller is based on the link state matrix Calculate the overall priority score ;in, are the round-trip delay, bandwidth and packet loss rate of the shortest path from the node to the user, is the preset weight constant.

[0051] S24: Press Sort from small to large, select the first nodes as the priority node list.

[0052] Furthermore, to seamlessly communicate the controller's calculated "best source, best path, best traffic split ratio" decision to the forwarding plane and directly indicate "who takes which path and how much" at the protocol layer, the controller, after obtaining a list of priority nodes, first selects multiple non-overlapping end-to-end paths with complementary link metrics based on the topology and calculates a target traffic weight for each path. It then constructs an extended Interest packet, adding a service type field to distinguish between service scenarios, using a path number to identify the selected path sequence, using the path status to carry a real-time path score so that nodes along the path can report back the link quality, and using the target traffic weight to specify the traffic ratio borne by the path. Finally, the Interest packet is sent to each source node. This allows the source node to directly tag data blocks by path number and fragment them according to the traffic weight ratio at the sending end, avoiding mid-transmission switching. Furthermore, the real-time path score field provides real-time closed-loop feedback on link quality. The selection of multiple non-overlapping paths eliminates contention for hot links, enabling true multi-source parallel transmission and bandwidth aggregation, significantly improving throughput and reducing congestion risks.

[0053] Furthermore, in order to maintain the optimal bandwidth utilization and the lowest end-to-end delay in the long-delay, high-fluctuation satellite-to-ground link, the present invention requires each forwarding node to allocate the remaining bandwidth during the forwarding process. , round-trip delay and packet loss rate The extended interest packet is embedded with the path status field and reported periodically; based on this, the controller calculates the link utilization and the comprehensive delay including the four components of transmission, propagation, queuing, and processing in real time, and then gives a path fitness score; the higher the score, the more idle the path, the more abundant the bandwidth, and the more friendly the delay. The controller then distributes the to-be-sent traffic proportionally to each path, so that the high-scoring paths bear more traffic and the low-scoring paths automatically reduce the load; if the score of a path drops significantly with the reported data, the new score and traffic ratio can be brought back to the source node by the next round of extended interest packets, so that the throughput can be kept stable and the delay can be controlled when the link quality fluctuates or the node is congested. The S4 specifically includes: S41: The forwarding node reports the remaining bandwidth at a preset period , round-trip delay and packet loss rate .

[0054] S42: The controller calculates the link utilization using the following formula: : ;in, Indicates the currently occupied bandwidth. Indicates the upper limit of the physical bandwidth.

[0055] S43: The controller calculates the comprehensive delay using the following formula : ;in, Indicates the transmission delay, represents the propagation delay, represents the queuing delay, Indicates processing delay.

[0056] S44: The controller calculates the path fitness score using the following formula : ;in, is the adjustment weight, and .

[0057] S45: The controller allocates the flow ratio of each path according to the path fitness score ratio using the following formula: ;in, The bandwidth requirements requested for the user task.

[0058] Furthermore, in order to fully exploit the parallel advantages of multiple copies and multiple paths and maintain service quality when the link fluctuates, the controller first divides the entire remote sensing data into consecutively numbered , and generates a three-dimensional mapping table of "source node-path-segment"; after receiving the mapping, each source node directly writes the corresponding path number tag for each data segment locally and sends the packet synchronously through the designated path in the order of the number, thus forming a parallel transmission flow with multi-source segmentation, link multiplexing, and bandwidth superposition on the user side. At the same time, the forwarding node encapsulates the real-time measured remaining bandwidth, round-trip delay, and packet loss rate into the path status field and continuously reports it. The controller compares the target traffic ratio with the current path throughput difference. If it finds that the bandwidth of a certain path drops sharply or the delay soars, it immediately recalculates the fitness score of the new path and the target traffic ratio, dynamically rewrites the mapping table, and issues instructions to add or delete paths or adjust the segment ratio to achieve online load migration and bottleneck avoidance, thereby ensuring that the overall transmission process still maintains the target bandwidth and low latency indicators in a highly dynamic satellite-to-ground link environment.

[0059] The S5 specifically includes: S51: the controller counts the remote sensing data Split and generate a path traffic mapping table.

[0060] S52: The source node adds a path number to the local data segment according to the mapping table, and sends the data packet in sequence via the corresponding path. The forwarding node forwards the data packet according to the path number.

[0061] S53: During the forwarding process, the forwarding node periodically encapsulates the link remaining bandwidth, round-trip delay, and packet loss rate into the extended interest packet.

[0062] S54: If the controller detects that the bandwidth of a certain path decreases or the delay exceeds a threshold, it recalculates the path fitness score and traffic ratio and updates the path traffic mapping table.

[0063] To further verify the effect of the method of the present invention, please refer to Figure 8-12 , Figure 8 Schematic diagram of the second experimental scenario of the method of the present invention; Figure 9 This is a diagram showing the effect of the interest packet sending frequency on the delay of the method of the present invention; Figure 10 This is a throughput effect diagram of the multi-source multi-path scenario of the method of the present invention; Figure 11 This is a throughput comparison diagram of different scheduling algorithms of the method of the present invention; Figure 12 This is a diagram showing the download volume effects of different ground stations using different scheduling strategies according to the method of the present invention.

[0064] Figure 7 A "linear" scenario is shown where a single consumer requests data from two satellites in parallel on a satellite-to-ground link; Figure 8 This is expanded to a "mesh" scenario covering a global six-orbit constellation and multiple ground stations. Nodes numbered Node101-604 form a three-dimensional interconnected backbone, with ground stations in the Arctic, Kashgar, and Sanya participating simultaneously. The two topologies illustrate that this method supports both simple chaining and large-scale, intercontinental multi-hop forwarding and caching coordination, providing sufficient path and replica diversity for subsequent multi-metric scheduling.

[0065] like Figure 9 As shown in the figure, in the first scenario, when the interest packet sending frequency is increased from 1400pkt / s to 2100pkt / s, the average request latency of this method (ST-MMSS) increases linearly from about 32ms to 63ms, but is always 3-6ms lower than the ECMP and RAF synchronization test curves; especially at 1800pkt / s, ST-MMSS is about 6ms and 4ms lower than RAF and ECMP, respectively, indicating that the "spatiotemporal filtering + link scoring + diversion weight" mechanism effectively suppresses the delay expansion caused by queuing and retransmission under high concurrency.

[0066] like Figure 10As shown in the second scenario, fixed link delays of 10ms, 30ms, and 50ms were injected into the three candidate paths, respectively. It can be seen that the bandwidth of the three curves climbed within 10 seconds and stabilized on the 15-16Mbps platform line, with a fluctuation range of less than ±0.5Mbps. This shows that even if the heterogeneous RTT difference reaches 5 times, bandwidth weighting and online reallocation can still maintain high link utilization.

[0067] like Figure 11 As shown in the figure, under the same network conditions, the steady-state throughput of ST-MMSS is about 17Mbps, ECMP is about 16Mbps, and RAF is only 15Mbps; the overall improvement is about 6% for ECMP and about 13% for RAF. Figure 10 , which can be attributed to the fact that ST-MMSS dynamically migrates traffic according to real-time scores, avoiding the long-term bandwidth occupation of low-scoring paths caused by ECMP's equal distribution mismatch and RAF's path preemption.

[0068] like Figure 12 As shown, the effective download capacity of the three stations in the Arctic, Kashgar, and Sanya increased to 69MB, 66MB, and 71MB, respectively, during the same observation period. This represents a 13%-20% increase compared to RAF and a 5%-10% increase compared to ECMP. This result demonstrates that when multiple ground stations concurrently request imagery from the same spatiotemporal window, ST-MMSS's multi-source and multi-path weighting can automatically differentiate traffic based on each station's geographic location and link quality, thereby improving overall satellite-to-ground link utilization and shortening queuing times.

[0069] In summary, multiple simulations show that within the range of link latency of 10ms-50ms and interest packet rate of 1400-2100pkt / s, the average throughput of the present invention is improved by approximately 15%-28%, end-to-end latency is reduced by 12%-20%, and it can maintain a leading download volume in a globally distributed scenario, demonstrating its efficiency and stability in high-concurrency, highly dynamic remote sensing data distribution.

[0070] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0071] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A multi-source multi-path scheduling method for remote sensing data transmission, characterized in that: The method includes: S1: the controller uses a unified spatiotemporal naming structure to generate a unique name for each remote sensing data, and writes the unique name and the triple structure into the spatiotemporal replica index table based on the node information, acquisition coverage area, and acquisition timestamp of the remote sensing data; S2: when receiving the interest packet, the controller extracts the time field and space field therein, and obtains a multi-source node set according to the spatiotemporal replica index table, and sorts the multi-source node set according to the link state matrix to obtain a priority node list; the link state matrix includes bandwidth, round-trip delay, and packet loss rate; S3: the controller selects multiple non-overlapping paths according to the priority node list, and generates a service class The extended interest packet with the type, path number, path status and traffic ratio fields is sent to each source node; S4: The forwarding node continuously reports the remaining bandwidth, round-trip delay and packet loss rate through the extended interest packet. The controller calculates the path fitness score through link utilization and comprehensive delay, and allocates the traffic ratio of each path according to the path fitness score ratio; S5: According to the multi-source node list and path traffic mapping issued by the controller, each source node sends the remote sensing data segment to the user end in parallel via the corresponding path according to the segment number; during the transmission process, the forwarding node continuously reports link changes, and the controller adjusts the path set and traffic ratio in real time to ensure the overall transmission bandwidth and delay performance.

2. The multi-source multi-path scheduling method for remote sensing data transmission according to claim 1, characterized in that: The S1 specifically includes: S11: the controller identifies the source node based on the remote sensing data , business type , Collection coverage area , collection timestamp , generated or cached node information and data types , generate a unique name for each remote sensing data; S12: the unique name and triple The structure is written to the time-space copy index table.

3. The multi-source multi-path scheduling method for remote sensing data transmission according to claim 2, characterized in that: The S2 specifically includes: S21: When receiving the interest packet, the controller extracts the time field in the interest packet and spatial fields ; S22: traverse each triple in the spatiotemporal replica index table ,when and Coverage area includes When the node information Join the candidate node set ; S23: For the candidate node set Each node in , the controller is based on the link state matrix Calculate the overall priority score ;in, are the round-trip delay, bandwidth and packet loss rate of the shortest path from the node to the user, is the preset weight constant; S24: press Sort from small to large, select the first nodes as the priority node list.

4. The multi-source multi-path scheduling method for remote sensing data transmission according to claim 3, characterized in that: The S4 specifically includes: S41: the forwarding node reports the remaining bandwidth at a preset period , round-trip delay and packet loss rate ; S42: The controller calculates the link utilization by the following formula : ;in, Indicates the currently occupied bandwidth. Indicates the upper limit of the physical bandwidth; S43: The controller calculates the comprehensive delay using the following formula : ;in, Indicates the transmission delay, represents the propagation delay, represents the queuing delay, Indicates processing delay; S44: The controller calculates the path fitness score using the following formula : ;in, is the adjustment weight, and ; S45: The controller allocates the flow ratio of each path according to the path fitness score ratio using the following formula: ;in, The bandwidth requirements requested for the user task.

5. The multi-source multi-path scheduling method for remote sensing data transmission according to claim 4, characterized in that: The S5 specifically includes: S51: the controller counts the remote sensing data Split and generate a path traffic mapping table; S52: After the source node adds a path number to the local data segment according to the mapping table, it sends the data packet in sequence through the corresponding path, and the forwarding node forwards the data packet according to the path number; S53: During the forwarding process, the forwarding node periodically encapsulates the link's remaining bandwidth, round-trip delay, and packet loss rate in the extended interest packet; S54: If the controller detects that the bandwidth of a certain path has decreased or the delay has exceeded the threshold, it recalculates the path fitness score and traffic ratio, and updates the path traffic mapping table.

Images