Real-time data transmission and distribution system for aircraft with multi-network convergence

Abstract

This invention relates to the field of digital information transmission technology and discloses a real-time data transmission and distribution system for aircraft multi-network convergence, comprising: a physical layer status monitoring module, a transmission quality inference engine, an inference routing engine, and an image model calibration module. The system acquires physical layer statistics and infers transmission quality indicators based on the transport layer-physical layer image model to guide the routing engine in distributing data packets. Simultaneously, the calibration module uses actual transport layer feedback from data packets to dynamically correct the image model. This invention replaces active measurement with passive inference, avoiding information lag and resource congestion, and solves the drift failure problem of the inference model in complex environments by introducing a self-calibration closed loop.

Description

Technical Field

[0001] This invention relates to a real-time data transmission and distribution system for aircraft with multi-network integration, belonging to the field of digital information transmission technology. Background Technology

[0002] One of the core tasks is to plan the optimal transmission path for data packets in complex network environments. To this end, existing network decision-making systems generally rely on active measurement methods, in which the system actively sends out probe data packets to measure the quality of different transmission links, such as latency and packet loss, and then makes routing or distribution decisions based on these measurement results. However, when this decision-making method, which is effective in terrestrial networks, is applied to the specific scenario of multi-network convergence in aircraft, it exposes inherent operational constraints. The environment in which aircraft operate integrates high-latency satellite links, bandwidth-scarce backup links, and dynamically changing terrestrial cellular networks, making it logically difficult to balance active measurement behavior: on the one hand, for bandwidth-scarce links, probe packets themselves will crowd out the already valuable transmission resources, and the measurement behavior itself will interfere with service data; on the other hand, for high-latency links, due to the high-speed mobility of aircraft, when the system receives the measurement report from the previous moment, the aircraft's attitude or position may have already changed, and the actual state of the link is no longer the same as at that time, resulting in a lag in the decision-making basis.

[0003] In response to this situation, linear improvements attempted by those skilled in the art, such as adjusting the detection frequency, cannot fundamentally solve this constraint. If the detection frequency is reduced to avoid resource consumption, the link state information on which the decision is based will become increasingly outdated, and the system will be unable to respond to instantaneous changes in the link. If the detection frequency is increased to pursue real-time performance, it will exacerbate the crowding out of scarce link resources, and may even lead to routing oscillations due to frequent switching based on outdated information. This means that the system decision is always constrained by the trade-off between information lag and measurement overhead. In addition to the limitations of the aforementioned active measurement methods, there are other solutions in the art that attempt to alleviate network congestion at the application layer, such as the Chinese invention patent with publication number CN116389374A. The patent discloses an adaptive adjustment method for data distribution throughput of a multi-robot system in a SLAM scenario. This scheme attempts to estimate the throughput limit through an online learning algorithm and distinguish data types, discarding image frames with low utility to adapt to the network. However, the technical basis of this method relies on the observation of message latency to infer the network state. This mechanism, which depends on latency feedback from the transport layer or application layer, has inherent lag when applied to the high-latency, high-dynamic aircraft network environment of this invention. When the system makes decisions based on outdated latency data, the network link state of the aircraft may have already changed. This scheme still does not fundamentally solve the information lag problem under extreme conditions and cannot meet the requirements of real-time high-fidelity decision-making in multi-network fusion scenarios of aircraft.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a new data transmission decision mechanism that can obtain the real transmission quality of multiple heterogeneous links without relying on active detection or occupying service bandwidth, and then distribute data packets of different service levels in real time and stably. Summary of the Invention

[0005] This invention provides a real-time data transmission and distribution system for aircraft multi-network convergence. Its main purpose is to solve the problem of information lag or resource congestion caused by relying on active detection in the prior art, which makes it impossible to distribute aircraft multi-network data in real time and efficiently.

[0006] To achieve the above objectives, the present invention provides a real-time data transmission and distribution system for aircraft multi-network convergence, the system comprising:

[0007] The physical layer status monitoring module is used to obtain physical layer statistics from multiple heterogeneous network interfaces on the aircraft.

[0008] The transmission quality inference engine is used to infer physical layer statistics into transmission quality indicators that characterize multiple heterogeneous network interfaces based on the transport layer-physical layer mapping model.

[0009] The inference routing engine is used to select the target transmission interface for the data packets to be transmitted based on the transmission quality indicators output by the transmission quality inference engine, and to distribute the data packets through the target transmission interface.

[0010] The image model calibration module is used to obtain physical layer statistics, obtain the actual transport layer feedback of data packets distributed through the target transport interface, and dynamically correct the transport layer-physical layer image model when there is a preset deviation between the actual transport layer feedback and the transport quality indicators derived by the transport quality inference engine based on physical layer statistics.

[0011] Preferably, the inference routing engine also includes a slow passive check loop, which is used to passively monitor the transport layer feedback of data packets distributed via the target transport interface in order to generate actual transport layer feedback.

[0012] Preferably, the transport layer-physical layer mapping model is a static transport quality lookup table, and the mapping model calibration module is used to dynamically correct the entries in the transport quality lookup table when a preset deviation exists.

[0013] Preferably, the system also includes: a message service level identification module, used to classify data packets based on a preset policy rule base to determine the service level of the data packets; and a deductive routing engine, used to select a target transmission interface based on the service level of the data packets and transmission quality indicators.

[0014] Preferably, physical layer statistics include one or more of the following: signal-to-noise ratio (SNR), bit error rate (BER), and received signal strength indicator (RSSI).

[0015] Preferably, the transmission quality indicators include one or more of the following: the estimated packet loss rate or the estimated latency.

[0016] Preferably, the system further includes a downlink quality signaling encapsulation module; the downlink quality signaling encapsulation module is used to obtain the transmission quality indicators that characterize the downlink of the aircraft output by the transmission quality inference engine, generate the transmission quality indicators into status messages, and encapsulate the status messages in data packets distributed by the inference routing engine and sent to the ground station.

[0017] Preferably, the system also includes a communication health trend analysis module; the communication health trend analysis module is used to archive the transmission quality indicators output by the transmission quality inference engine within a historical time period, perform statistical trend analysis on the transmission quality indicators within the historical time period, and generate a communication system maintenance alarm when the result of the statistical trend analysis meets the preset maintenance threshold.

[0018] Preferably, the communication health trend analysis module is used to calculate the slope of the degradation trend when performing statistical trend analysis. : ,in, For transmission quality indicators at the first historical time point The rolling average, For transmission quality indicators at the second historical time point The rolling average value of the communication health trend analysis module in terms of the slope of the degradation trend. When the preset positive slope threshold is exceeded, a communication system maintenance alarm is generated.

[0019] Preferably, the system also includes a network mapping data collection module; the network mapping data collection module is used to obtain non-transmission quality event logs representing the operating status of heterogeneous network interfaces from the physical layer status monitoring module, obtain current geographical location information from the airborne navigation system, generate network mapping data packets based on the event logs and current geographical location information, and submit the network mapping data packets to the inference routing engine for low-priority distribution to the ground network planning server.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. The system of the present invention establishes a new data transmission decision-making method. It uses a cross-layer status monitoring module to obtain physical layer statistics, and generates transmission quality indicators that characterize the link status through a quality inference engine. These indicators serve as the basis for the distribution of the inference routing engine. This technical approach replaces the process of actively sending probe packets for link measurement with the passive acquisition of physical layer data. This allows the state perception of low-speed, high-latency channels to no longer come at the cost of occupying channel resources or acquiring outdated information, thus ensuring the real-time performance and effectiveness of data transmission decisions in complex network convergence environments.

[0022] 2. The transmission quality indicators of various heterogeneous networks were obtained through the quality inference engine. At the same time, the message service level identification module determined the service level of the data packets to be sent. When making distribution decisions, the inference routing engine can take into account the information of these two dimensions, guide high service level data packets to the inferred low latency links, and allocate low service level data packets to the inferred low cost links. This achieves differentiated and refined routing based on business intent and avoids blind competition between critical and non-critical data for high-quality links.

[0023] 3. Through the communication health trend analysis module, the transmission quality indicators generated by the quality projection engine in the historical period are archived and statistically trend analyzed. This mechanism reuses the instantaneous projection data that serves real-time routing decisions as a basis for non-real-time long-term hardware health status characterization, enabling the system to distinguish between instantaneous channel fluctuations and chronic hardware degradation, providing objective decision support for the predictive maintenance of the communication system, and expanding the management and maintenance capabilities of the data transmission system. Attached Figure Description

[0024] Figure 1 This is a system architecture diagram of the cross-layer deduction and closed-loop calibration of the present invention;

[0025] Figure 2 This is a diagram illustrating the link allocation effect of different service levels CoS in this invention.

[0026] Figure 3 This is an overview diagram of the core mechanism for achieving highly reliable data distribution in this invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are only some embodiments of this invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] This invention provides a real-time data transmission and distribution system for aircraft with multi-network convergence. Built on an onboard computing platform, the system's core objective is to address the resource contention and information lag inherent in traditional active probing mechanisms when aircraft integrate multiple heterogeneous network interfaces, such as high-orbit satellites, terrestrial 5G, and onboard networks. Logically, the system architecture includes a physical layer status monitoring module, a transmission quality inference engine, an inference routing engine, and an image model calibration module. These modules work collaboratively to replace the traditional measurement-action paradigm with a zero-overhead cross-layer inference approach, achieving real-time, efficient, and highly reliable data packet distribution. The physical layer status monitoring module acts as a passive cross-layer information collector. It does not send any active probe data packets but instead accesses multiple heterogeneous network interfaces on the aircraft frequently and periodically (for example, every 100ms using a clock cycle) through standard management interfaces such as Simple Network Management Protocol (SNMP) or dedicated APIs. These interfaces correspond to underlying modems, and the physical layer status monitoring module obtains real-time data from these devices to maintain their links. The measured physical layer statistics provide raw, non-physical layer data input for routing decisions. These statistics may include one or more of the following: Signal-to-Noise Ratio (SNR), Bit Error Rate (BER), and Received Signal Strength Indicator (RSSI). A transmission quality inference engine converts the physical layer statistics acquired by the physical layer status monitoring module into transmission layer performance metrics necessary for routing decisions. The engine's input is physical layer statistics, such as SNR=7dB for the Satcom interface. Internally, it embeds a transmission layer-physical layer mapping model, enabling a simple and efficient implementation. In this model, the image model can be implemented as a static transmission quality inference lookup table. This lookup table is generated through offline laboratory calibration during the R&D phase. It establishes a deterministic mapping relationship between indicators and KPIs. The transmission quality inference engine performs an instantaneous lookup operation and can output a set of inferred transmission quality indicators. The transmission quality indicators can specifically include the inferred packet loss rate or the inferred latency. Taking the Satcom interface with an input of SNR=7dB as an example, its lookup output can correspond to Est.Loss=1.5% and Est.Latency=750ms.

[0029] The system may also include a message service level (SLS) identification module. This module classifies incoming data packets before the routing engine operates, based on a pre-defined policy rule base. This rule base determines the SLS of the data packets based on header fields such as the Differential Service Code Point (DSCP) tag, VLAN tag, source / destination IP address, or destination port number. For example, it might identify air traffic control (ATC) instruction packets as CoS-1, airline operations control (AOC) packets as CoS-2, and passenger Wi-Fi packets as CoS-3. The routing engine is the final decision-making and distribution execution unit for data packets. Upon receiving a data packet, it simultaneously identifies the SLS from the message service level. The module obtains the CoS level of the data packet and retrieves real-time transmission quality metrics for all available network interfaces from the transmission quality inference engine. Based on the data packet's service level and transmission quality metrics, the engine selects the target transmission interface. Specifically, it executes a service level-dependent cost function. For example, for a CoS-1 data packet, the cost function might be defined as minimizing only the inference latency (Cost = Est.Latency); while for a CoS-3 data packet, the cost function might be defined as minimizing the transmission cost (Cost = Est.Cost_Per_Bit). The inference routing engine then selects the network interface with the lowest computational cost as the target transmission interface for this data packet. The system distributes the data packet via the target transmission interface. To address the potential inaccuracies in the inference process caused by dynamic drift of the transport layer-physical layer mapping model in the physical environment (such as the emergence of new interference types), the system also includes a mapping model calibration module. This module is used to implement closed-loop adaptive correction of the inference model. To achieve this function, the inference routing engine may also include a slow passive verification loop. The function of this verification loop is to passively monitor the actual feedback of the transport layer, such as the TCP protocol, the actual retransmission rate, or the actual packet loss rate while the data packet is being distributed, thereby generating actual transport layer feedback. The mapping model calibration module runs in the background with low priority, acquiring physical layer statistics (such as SNR=10dB) and obtaining... The actual transport layer feedback (e.g., ActualLoss=5%) on the same link generated by the verification ring is obtained. At the same time, the transport quality index (e.g., Est.Loss=1%) derived based on the SNR=10dB is obtained from the transport quality inference engine. When there is a preset deviation between the actual transport layer feedback (5%) and the derived transport quality index (1%) (e.g., the deviation continues to exceed the threshold), the image model calibration module is activated. The transport layer-physical layer image model will be dynamically corrected. In the implementation where the model is a lookup table, the correction action is specifically to rewrite the corresponding entry in the lookup table and correct the Est.Loss value corresponding to SNR=10dB from 1% to 5%, which is closer to the actual situation.

[0030] To address downlink routing failures caused by asymmetric uplink and downlink quality, the system may also include a downlink quality signaling encapsulation module. Since the downlink SNR, a statistical measure of the aircraft's receiver, is also obtained by the physical layer status monitoring module and extrapolated into downlink transmission quality indicators by the transmission quality extrapolation engine, the downlink quality signaling encapsulation module is used to obtain this set of downlink transmission quality indicators (e.g., [Satcom:Loss=90%], [5G:Loss=0.1%]) and generate a compact status message. Using signaling piggybacking, this status message is encapsulated in a reserved field or payload of any regular uplink data packet (e.g., an AOC report) distributed by the extrapolation routing engine. Upon receiving this uplink packet, the ground network equipment... The piggybacked information is parsed to instantly obtain the downlink status of the aircraft and optimize its downlink routing decisions accordingly, avoiding the transmission of critical data to failed links (such as Satcom). To extend the system's functionality from real-time tactical routing to long-term strategic maintenance, the system may also include a communication health trend analysis module. This module is activated during non-real-time periods (such as after aircraft landing) to archive transmission quality indicators (such as the projected packet loss rate) output by the transmission quality projection engine over a historical time period (such as the past 100 flight hours). Statistical trend analysis is performed on the transmission quality indicators over the historical time period. This analysis distinguishes between instantaneous channel fluctuations and / or chronic hardware degradation. In a specific implementation, the statistical trend analysis is used to calculate the degradation trend slope. Its calculation can follow ,in, This is the rolling average of the transmission quality metrics at the first historical point in time. Its rolling average at the second historical time point; when the results of statistical trend analysis (such as...) When the preset maintenance threshold is met, for example, if the slope of the degradation trend exceeds the preset positive slope threshold (e.g., +0.01% Loss / flight hours), it indicates that the link has a chronic performance degradation. The module will generate a communication system maintenance alarm for maintenance personnel to conduct predictive troubleshooting. In addition, to reuse the information byproducts of the physical layer to provide network planning value, the system may also include a network mapping data collection module. This module obtains non-transmission quality event logs generated by the modem during daily operation from the physical layer status monitoring module. For example, signal search / reacquisition events contain network coverage information. While obtaining such event logs, the collection module obtains the current geographical location information from the airborne navigation system and generates network mapping data packets based on the event logs and the current geographical location information. These data packets are marked as the lowest service level, such as CoS-4, and submitted to the inference routing engine. The inference routing engine uses its cost routing logic to automatically delay this non-urgent data packet until the lowest cost link, such as ground Wi-Fi, is available, before distributing it to the ground network planning server with low priority.

[0031] Example 1: During a high-speed flight mission, the aircraft simultaneously accesses a high-latency satellite communication interface (Satcom) and a dynamically available terrestrial cellular network interface (5G). When the aircraft is in the cruise phase, the 5G interface is temporarily unavailable, and the only communication entry and exit point for the system is the Satcom interface. At this time, the aircraft encounters a physical layer interference. This interference causes the physical layer signal-to-noise ratio (SNR), obtained by the physical layer state monitoring module, to be at a moderate level (e.g., 10 dB). However, the actual burst packet loss rate (Actual Loss) of the transmission layer is abnormally high, reaching 10%. This mapping relationship deviates from the system's expected value. In the initial static lookup table of the transport layer-physical layer mapping model, the projected packet loss rate Est.Loss corresponding to SNR=10dB is only 1%. Under this condition, when a high service level (CoS-1, taking air traffic control commands as an example) data packet arrives, the transport quality projection engine reports the projected packet loss rate of the Satcom interface to the projection routing engine based on its static lookup table. Based on this projection index, the projection routing engine distributes the CoS-1 data packet through the Satcom interface, but the data packet encounters an actual packet loss of 10% at the transport layer, resulting in transmission failure.

[0032] The slow passive check loop built into the routing engine monitors retransmission or failure feedback at the transport layer (such as TCP) to obtain the actual transport layer feedback for the data packet, i.e., ActualLoss=10%. The image model calibration module obtains this actual transport layer feedback in the background and compares it with the transport quality index (Est.Loss=1%) derived by the transport quality inference engine based on the same physical layer statistic (SNR=10dB). When the system determines that there is a preset deviation between the actual value and the derived value, the image model calibration module is activated and performs dynamic correction, dynamically adjusting the derived packet loss rate Est.Loss entry corresponding to the SNR=10dB interval in the transport layer-physical layer image model (lookup table) from 1% to 10%. %; Later, as the aircraft approaches land, the 5G interface becomes available, the physical layer status monitoring module acquires its physical layer statistics, and the transmission quality inference engine infers its transmission quality indicators (e.g., Est.Loss=0.1%, Est.Latency=50ms); At this time, when another CoS-1 data packet arrives, the link status obtained by the inference routing engine from the transmission quality inference engine changes to: Satcom interface (Est.Loss=10%, after calibration) and 5G interface (Est.Loss=0.1%). Based on the low latency and low packet loss rate cost function of CoS-1 data packets, the inference routing engine instantaneously selects the 5G interface with better transmission quality indicators as the target transmission interface and distributes data packets through this interface.

[0033] Example 2: To objectively verify the cross-layer inference method adopted by the system of this invention, and to assess its response capability to instantaneous link changes in highly dynamic aviation scenarios compared to the active measurement method, a network simulation test platform was constructed. This platform simulates a high-speed aircraft, with interfaces including: a high-latency satellite interface (Satcom), with a baseline round-trip time (RTT) set to 600ms and a baseline packet loss rate of 0.1%; and a low-latency ground interface (5G), with a baseline RTT of 40ms and a baseline packet loss rate of 0.01%. The simulation platform can simulate changes in physical layer channel quality and accurately measure the actual latency and packet loss of data packets. A control group and the sample group of this invention were set up. The control group adopted the active measurement method, in which its routing decision unit sent probe packets (Ping) to the Satcom and 5G interfaces at a fixed period of 1.0 second, and updated the routing table based on the returned measurement latency and packet loss rate. The sample group of this invention... A cross-layer inference method was adopted, without sending any active probe packets. The physical layer status monitoring module obtained physical layer statistics (SNR) from the virtual interface of the simulation platform at a period of 100ms. The transmission quality inference engine infers the transmission quality indicators instantaneously based on the transport layer-physical layer mapping model (static lookup table), and the inference routing engine distributes the data accordingly. The test scenario was set as follows: the 5G interface encountered instantaneous physical layer deep fading. At the test time T=0.5s, its SNR dropped instantaneously from a stable 25.3dB to 4.1dB (corresponding to an actual packet loss rate >50%). After this fading lasted for 800ms, the SNR instantaneously recovered to 26.1dB at T=1.3s. The test service flow was a continuous, constant-rate CoS-1 level simulated data packet. During the test fading event, the decision-making behavior of the two systems and the transmission performance of the CoS-1 data packets were recorded in Table 1.

[0034] Table 1: Comparison of System Response under Instantaneous Fading Scenarios

[0035]

[0036] Referring to Table 1, at T=0.5s, the 5G link SNR momentarily dropped. The physical layer state monitoring module of this invention's sample group acquired the fading SNR at T=0.6s (lagging by 100ms sampling period). The transmission quality inference engine instantly predicted that the 5G link packet loss rate rose to 55.8%. The inference routing engine immediately switched all CoS-1 data packets to the Satcom interface, which had better transmission quality indicators at the time (inferred packet loss rate ≈0.1%). The actual latency of the data packets increased accordingly to about 605ms, but the actual packet loss rate remained at a low level (0.11%). In contrast, the control group's active probe packets only returned failure reports at T=1.0s, with a decision lag of 500ms. This caused it to continue sending CoS-1 data packets to the control group during the time window from T=0.5s to T=1.0s. The malfunctioning 5G interface caused a 55.8% loss of data packets during that period. Furthermore, when the SNR recovered at T=1.3s, the sample group of this invention deduced the recovery state at T=1.4s (lagging by 100ms) and instantly switched back to the high-quality 5G interface. In contrast, the control group did not switch back to the 5G interface until the next detection cycle at T=2.0s, causing data to unnecessarily remain on the high-latency link for 600ms. Experimental data showed that when the network physical layer state changes rapidly at the millisecond level, the active measurement method suffers from decision lag due to its measurement cycle and latency overhead, leading to the loss of critical data or improper routing selection. The cross-layer deduction method adopted in this invention passively acquires physical layer statistics with zero overhead, which can synchronize the instantaneous changes of the physical layer and convert them into transmission quality indicators, avoiding decision lag.

[0037] Example 3: This example combines Figures 1 to 3 A description of the real-time data transmission and distribution system for aircraft multi-network convergence, such as... Figure 1 As shown, the diagram illustrates a physical layer status monitoring module that acquires physical layer statistics such as SNR, BER, and RSSI from heterogeneous network interfaces and 5G, as well as a message service level identification module that determines the service level (CoS) of data packets to be transmitted. The physical layer status monitoring module outputs physical layer statistics to the transmission quality inference engine, which infers the inferred transmission quality indicators such as packet loss rate and latency. These indicators, along with the service level, are fed into the inference routing engine, which executes a decision to select a target transmission interface for data packet distribution. Simultaneously, the actual transmission layer feedback generated during this distribution process comes from the slow passive check loop and is input into the image model calibration module along with the physical layer statistics. After the calibration module dynamically corrects the image model, it feeds the results back to the transmission quality inference engine, forming a self-calibration closed loop.

[0038] like Figure 2As shown in the chart, with service level as the horizontal axis, the distribution of data packets of four different service levels—CoS-1 (ATC commands), CoS-2 (AOC data), CoS-3 (passenger Wi-Fi), and CoS-4 (network mapping)—on two dimensions: satellite link utilization and 5G link utilization. The data shows that the high-priority CoS-1 primarily utilizes 5G links (85%), while the low-priority CoS-4 overwhelmingly utilizes satellite links (90%). Figure 3 As shown in the figure, the cross-layer inference mechanism is implemented by passively acquiring physical layer statistics and inferring transmission quality indicators; differentiated service routing is implemented by identifying the service level of data packets and implementing the cost function that depends on the service level; closed-loop adaptive correction is implemented by acquiring actual transport layer feedback and dynamically correcting the image model; and long-term health trend analysis is implemented by archiving historical transmission quality indicators and generating communication system maintenance alarms.

[0039] Example 4: This example illustrates the transport layer-physical layer mapping model within the transport quality inference engine, i.e., the static transport quality inference lookup table. This is a standardized engineering procedure that performs offline calibration during the R&D phase to eliminate the problem of unclear model origin. In the calibration scenario, the initial state definition of the enabling environment includes: a network impairment simulator that can accurately simulate various channel conditions, a specific type of airborne modem to be calibrated (e.g., a Satcom Modem), a traffic generator, and a high-precision network analyzer. The problem to be solved by calibration is to establish a reproducible mapping relationship between physical layer statistics, such as SNR, and transport layer performance indicators, namely packet loss rate (Loss) and latency (Latency).

[0040] The calibration process employs a parameter scanning method: First, using a network impairment simulator, a specific physical layer signal-to-noise ratio (SNR) value is set, with the scanning range taken as -5dB to +25dB, and the scanning step size set to 1dB. The initial SNR is set to -5dB. Second, the traffic generator sends a large sample of data packets, such as N=10000 1024-byte UDP packets, to the network analyzer via the modem interface. Third, the network analyzer measures and records the actual transport layer packet loss rate at that SNR point. =(N-number of receivers) / N) and average one-way delay ( Fourth, repeat steps one through three, incrementing the SNR value until all scan points at +25dB are completed, obtaining a set of raw (SNR, The calibration dataset is divided into three triplets. After calibration, the original calibration dataset is subjected to interval statistical processing to generate lookup table entries. Taking the <5dB interval of the Satcom interface as an example, all data points with SNR <5dB in the calibration dataset are extracted, and the statistical mean or 95th percentile of these points is calculated. If the statistical value is 14.8%, then 15.0% (with a certain engineering margin) is used as the extrapolated packet loss rate (Est.Loss) for this interval. Similarly, the statistical mean of this interval is calculated, such as 1180ms, and its margin value of 1200ms is used as the extrapolated latency (Est.Latency). This generates a specific entry in the lookup table: Satcom, <5dB, >10e-5, 15.0%, 1200ms. Through this procedure, static transmission quality extrapolation lookup tables with engineering reproducibility can be established for different types of network interfaces (such as 5G interfaces) for use by the transmission quality extrapolation engine during flight.

[0041] Example 5: This example illustrates the preset policy rule base of the message service level identification module. A reproducible procedure defined before system deployment is implemented. The purpose of this procedure is to ensure that the inductive routing engine can perform differentiated routing decisions for different service data packets. The input to the procedure is the Network Interface Control Document (ICD) of the airborne avionics system, which defines the network traffic characteristics of all airborne applications. The procedure steps are: traversing the application list in the ICD, classifying their network traffic according to service importance, and extracting the five-tuple features of their packet header (source IP, destination IP, protocol number, source port, destination port) as matching keys. Taking the Air Traffic Control (ATC) data link application as an example, its service importance is the highest. The ICD defines its destination port number as X1, then the policy... The first rule generated in the rule base is: IF(target port=X1)THENServiceLevel=CoS-1; Airline Operations Control (AOC) applications (such as engine status feedback) are of secondary importance, and the ICD defines its target IP address range as Y1 to Y2, so the second rule is generated: IF(target IPIN[Y1,Y2])THENServiceLevel=CoS-2; Cabin Entertainment (IFE) applications are of the lowest importance, and the ICD defines its source IP network segment as Z1, so the default rule is generated: IF(source IPINZ1)THENServiceLevel=CoS-3; The policy rule base output by this procedure is embedded in the message service level identification module, providing a classification basis for its operation.

[0042] This embodiment further elaborates on the standardized offline calibration procedure for the preset maintenance threshold of the communication health trend analysis module in the specific implementation. The purpose of this procedure is to establish a statistically significant alarm baseline to distinguish between normal aging and abnormal chronic degradation of hardware. The input to the procedure is a historical transmission quality index dataset of the healthy fleet. This dataset contains the packet loss rate time series of each Satcom interface archived by the transmission quality inference engine over the past 2000 flight hours, taking 20 aircraft as an example. The calibration steps include: First, performing linear regression analysis on the packet loss rate time series of each interface in the dataset to calculate its degradation trend slope. Second, collect the values ​​of all health interfaces to form a statistical distribution of the normal aging slope, and calculate the mean (e.g., +0.003% Loss / flight hour) and standard deviation (e.g., 0.002% Loss / flight hour) of this distribution; Third, set alarm thresholds based on this statistical distribution, with a 3-sigma control limit (…). For example, the preset maintenance threshold is set to +0.009% Loss / flight hours. This threshold is loaded into the communication health trend analysis module. When the degradation trend slope of any subsequent interface exceeds 0.009%, the system determines that it deviates from the normal aging baseline and generates a maintenance alarm.

[0043] Example 6: This example illustrates a reproducible engineering implementation procedure for obtaining actual transport layer feedback using a slow passive verification ring. This procedure is executed at the operating system kernel level of the airborne computing platform, and its technical path is passively parsing network protocol stack statistics. Taking a Linux-based system as an example, the slow passive verification ring is configured as a low-priority background daemon process, which uses a slow sampling period (e.g., ...). A statistical snapshot operation is performed every 10 seconds; at the sampling point... At that time, the process reads the MIB statistics counter exposed by the kernel network subsystem, specifically the cumulative number of retransmitted packets in the TCP protocol. and the total number of messages sent The check loop then calculates the incremental retransmission rate over the past 10-second period. The calculation formula is as follows: This calculation result That is, it serves as the actual transmission layer feedback output to the image model calibration module.

[0044] This embodiment further elaborates on the reproducible engineering implementation procedure for the network mapping data collection module to acquire non-transmission quality event logs. This procedure is based on the standard Simple Network Management Protocol (SNMP). During system deployment, the network mapping data collection module is configured as an SNMP manager and loads the proprietary Management Information Base (MIB) provided by the onboard modem (such as a 5G modem) manufacturer. The MIB defines specific Object Identifiers (OIDs) to store the latest status codes of network search events (e.g., 0x01 represents the start of the search, and 0x02 represents successful recapture). The network mapping data collection module sends SNMPGet requests to the modem in a low-priority polling manner, such as every 30 seconds, to read the value of the OID (i.e., the event log) and packages it with geographical location information to generate a network mapping data packet.

[0045] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An aircraft multi-network converged data real-time transmission and distribution system, characterized in that, The system comprises: a physical layer status monitoring module configured to obtain physical layer statistics from a plurality of heterogeneous network interfaces onboard an aircraft; a transport quality inference engine configured to infer the physical layer statistics into transport quality indicators characterizing the plurality of heterogeneous network interfaces based on a transport layer-physical layer mapping model; an inference routing engine configured to select a target transport interface for a data packet to be transmitted based on the transport quality indicators output by the transport quality inference engine, and to distribute the data packet via the target transport interface; a mapping model calibration module configured to obtain the physical layer statistics, to obtain actual transport layer feedback of the data packet distributed via the target transport interface, and to dynamically correct the transport layer-physical layer mapping model when a preset deviation exists between the actual transport layer feedback and the transport quality indicators inferred by the transport quality inference engine based on the physical layer statistics.

2. The real-time data transmission and distribution system for aircraft multi-network integration according to claim 1, wherein, The inference routing engine further comprises a slow passive check loop configured to passively monitor the transport layer feedback of the data packet distributed via the target transport interface to generate the actual transport layer feedback.

3. The real-time data transmission and distribution system for aircraft multi-network integration according to claim 1, wherein, The transport layer-physical layer mapping model is a static transport quality inference lookup table, and the mapping model calibration module is configured to dynamically correct entries in the transport quality inference lookup table when the preset deviation exists.

4. The real-time data transmission and distribution system for aircraft multi-network integration according to claim 1, wherein, The system further comprises a message service level identification module configured to classify the data packet based on a preset policy rule library to determine a service level of the data packet, and the inference routing engine is further configured to jointly select the target transport interface based on the service level of the data packet and the transport quality indicators.

5. The real-time data transmission and distribution system for aircraft multi-network integration according to claim 1, wherein, The physical layer statistics comprise one or more of a signal-to-noise ratio (SNR), a bit error rate (BER), and a received signal strength indication (RSSI).

6. The real-time data transmission and distribution system of claim 1, wherein, The transport quality indicators comprise one or more of an inferred packet loss rate or an inferred latency.

7. The real-time data transmission and distribution system of claim 1, wherein, The system further comprises a downlink quality signaling encapsulation module configured to obtain the transport quality indicators characterizing a downlink of the aircraft output by the transport quality inference engine, to generate the transport quality indicators into a status message, and to encapsulate the status message into the data packet distributed by the inference routing engine to a ground station.

8. The real-time data transmission and distribution system of claim 1, wherein, The system further comprises a communication health trend analysis module. The communication health trend analysis module is configured to archive the transport quality indicators output by the transport quality inference engine in a historical time period, to perform statistical trend analysis on the transport quality indicators in the historical time period, and to generate a communication system maintenance alarm when a result of the statistical trend analysis satisfies a preset maintenance threshold.

9. The real-time data transmission and distribution system of claim 8, wherein, The communication health trend analysis module is configured to calculate a degradation trend slope when performing statistical trend analysis : wherein is a rolling average of the transmission quality indicator at a first historical time point, is a rolling average of the transmission quality indicator at a second historical time point, is a rolling average of the transmission quality indicator at a second historical time point, The communication health trend analysis module is configured to generate a communication system maintenance alert when the degradation trend slope exceeds a pre-defined positive slope threshold.

10. The real-time data transmission and distribution system for aircraft multi-network fusion according to claim 1, wherein, The system further comprises a network mapping data collection module configured to obtain non-transport quality event logs characterizing running states of the heterogeneous network interfaces from the physical layer status monitoring module, to obtain current geographic location information from an onboard navigation system, to generate a network mapping data packet based on the event logs and the current geographic location information, and to submit the network mapping data packet to the inference routing engine for low-priority distribution to a ground network planning server.

Images