Cloud-edge collaborative flight management system and its route planning method

Abstract

The application provides a flight management system based on cloud-edge cooperation and a flight path planning method thereof, the flight management system comprising an airborne end, a cloud platform and an edge computing node; the flight path planning method of the flight management system comprises the following steps: step one, the cloud platform calculates an optimal flight path, predicts weather conditions on the flight path and corresponding real-time airspace information; step two, the cloud platform completes flight path optimization according to the data of step one and sends the optimized flight path to the edge computing node and the airborne end; step three, when a sudden situation occurs, the edge computing node calculates a corresponding escape path, sends the escape path to the airborne end and the air traffic control center, and the air traffic control center decides whether the aircraft executes the escape path; the application can solve the defects of traditional FMS in dynamic planning capability, computing power limitation and the like, and provides an extensible, highly reliable and low-cost solution for the next generation of flight management systems.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of civil aircraft avionics system, and in particular to a flight management system based on cloud-edge collaboration and a flight route planning method thereof. BACKGROUND

[0002] The traditional Flight Management System (FMS) is the core of modern avionics system. Since the 1970s, it has become the core avionics equipment of modern commercial aircraft. Its core functions include flight planning, trajectory prediction, flight guidance, navigation calculation, performance calculation, radio tuning, etc. Its main role is to reduce the operational burden of pilots, improve flight comfort and economy through route optimization, ensure flight safety, make flight more stable and comfortable, and save fuel.

[0003] The traditional FMS has two typical technical implementations:

[0004] (1) Adopting an integrated modular avionics (IMA) system architecture, the FMS software resides in the avionics core processing system (ACPS) platform.

[0005] (2) Relying on a special embedded computer (such as a PowerPC architecture processor) and a customized hardware module (ARINC429 / AFDX bus interface).

[0006] The data management part of the existing flight management system stores the navigation database locally and updates it regularly through physical media. With the increasing congestion of airspace, the increasing frequency of extreme weather, the International Civil Aviation Organization (ICAO) international aviation carbon offset and emission reduction plan (CORSIA) requiring net zero carbon emissions by 2050, and the need for airlines to reduce fuel costs, these factors have put higher requirements on the flight route planning function of the flight management system. Although the traditional FMS technology is mature, its inherent defects are increasingly prominent in the background of the digital transformation of the aviation industry.

[0007] The main problems of the flight management system and its flight route planning method in the prior art include the following aspects:

[0008] (1) Static route planning, unable to dynamically respond to real-time environmental changes

[0009] The route planning function of the traditional FMS plans the entire route before takeoff according to the aircraft performance and predicted weather data, and cannot integrate dynamic information such as weather and airspace traffic in real time, resulting in low flight efficiency. Especially when flying across the ocean or in complex airspace, sudden weather requires manual intervention.

[0010] (2) Insufficient real-time data processing capability

[0011] Civil aircraft has strict requirements on the weight and power of airborne equipment, which leads to the limitation of the traditional FMS local computing power (typical FMS computing power <1 TOPS), making it difficult to support real-time route optimization and a large amount of data processing. At the same time, the traditional FMS relies on ACARS to transmit weather update data, with a response delay of 5-10 minutes, which cannot meet the demand of high-density airspace.

[0012] (3) Limited function expansion

[0013] The flight management system in the prior art usually adopts strong coupling of software and hardware, and new functions need to be re-certified, which has a long cycle.

[0014] (4) Lack of ecological synergy

[0015] The traditional FMS cannot interact with the air traffic control system, airline operation center (AOC) and the like in real time, resulting in low utilization of airspace resources, and data cannot be shared across flights and airlines, and the potential of global optimization is not released.

[0016] Therefore, it is necessary to provide a flight management system based on cloud-edge collaboration and a route planning method thereof to solve the fundamental defects of the traditional FMS in dynamic planning capability and computing power limitation, and to provide an extensible, high-reliable and low-cost solution for the next generation of flight management system, which is especially suitable for high-density airspace, emerging electric aviation (eVTOL) and unmanned aerial vehicle logistics scenarios. SUMMARY

[0017] To solve the problems in the prior art, the present application provides a flight management system based on cloud-edge collaboration and a route planning method thereof, which can solve the fundamental defects of the traditional FMS in dynamic planning capability and computing power limitation, and provide an extensible, high-reliable and low-cost solution for the next generation of flight management system, which is especially suitable for high-density airspace, emerging electric aviation (eVTOL) and unmanned aerial vehicle logistics scenarios.

[0018] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a flight management system based on cloud-edge collaboration, comprising:

[0019] An airborne end deployed on an airplane for data security verification and flight management of the airplane;

[0020] A cloud platform deployed on the ground for route optimization based on route historical data learning, weather data learning, airspace data learning and airplane performance data learning, and sending the optimized route information to the airborne end and the edge computing node;

[0021] An edge computing node is deployed near an airport, an air traffic control center, a navigation station or a 5G base station, and is used to predict meteorological phenomena and airspace conditions that have a significant impact on the flight of an aircraft within the coverage area of the edge computing node, and to send the coordinates of the dangerous area and the recommended flight path to the airborne terminal according to the prediction results; at the same time, a real-time airspace situation map is constructed, and a relief path is calculated and sent to the airborne terminal and the air traffic control center, and the air traffic control center decides whether the aircraft executes the relief path.

[0022] In the preferred technical solution, the cloud platform comprises a flight route historical data analysis module, a meteorological data analysis module, an airspace data analysis module, an aircraft performance database and a flight route optimization model.

[0023] The flight route historical data analysis module analyzes the flight data of all flights on the corresponding flight route to obtain the optimal flight path of the aircraft, and sends the optimal flight path to the flight route optimization model.

[0024] The meteorological data analysis module is used to receive current meteorological data and historical meteorological data from a meteorological center, predict the meteorological conditions on the flight route, and send the meteorological conditions to the flight route optimization model.

[0025] The airspace data analysis module is used to calculate real-time airspace information when encountering thunderstorms, airspace control and similar emergencies, and send the real-time airspace information to the flight route optimization model.

[0026] The aircraft performance database is used to store complete aircraft performance parameters of the corresponding aircraft.

[0027] The flight route optimization model is used to optimize the current flight route based on the data provided by the flight route historical data analysis module, the meteorological data analysis module, the airspace data analysis module and the aircraft performance database, and to send the optimized flight route information to the edge computing node and the airborne terminal.

[0028] In the preferred technical solution, the edge computing node comprises a node meteorological data analysis module, a node airspace analysis module and a relief path calculation module.

[0029] The node meteorological data analysis module is used to receive real-time meteorological data from a meteorological center, airport meteorological radar and aircraft meteorological radar data, predict extreme meteorological phenomena such as wind shear and thunderstorms that have a significant impact on the flight of an aircraft within the next fifteen minutes, and send the coordinates of the dangerous area and the recommended flight path to the airborne terminal through ACARS or CPDLC.

[0030] The node airspace analysis module receives the position and speed information of surrounding aircraft through ADS-B, constructs a real-time airspace situation map, and sends the current airspace situation to the relief path calculation module.

[0031] The relief path calculation module receives data of the node weather data analysis module, the node airspace analysis module and the cloud platform route optimization model, calculates the corresponding relief path, and sends the corresponding relief path to the airborne terminal and the air traffic control center, so that the air traffic control center determines whether the aircraft executes the relief path.

[0032] In the preferred technical solution, the airborne terminal comprises a data security verification module and an airborne FMS; the data security verification module is used for security verification of all data sent by the cloud platform and the edge computing node to the airborne terminal, and the data is sent to the airborne FMS after security verification and confirmation of data security; and the airborne FMS is used for flight management of the aircraft.

[0033] In the preferred technical solution, the cloud-edge collaborative flight management system further comprises a federated learning algorithm module, which is used for encryption processing of corresponding data uploaded by each airline to the cloud platform.

[0034] In the preferred technical solution, the federated learning algorithm module comprises a local training module and a federated learning core module; the local training module is used for training of a model by each airline using local data, generation of a local model without data leaving the local area, and transmission of the trained model parameters to the federated learning core module; and the federated learning core module is used for encryption and aggregation of the parameters of the local training model of each airline, generation of a global model, and encryption and pushing of the global model parameters to the local model of each airline, so as to perform global model optimization under the premise of ensuring data privacy of each airline.

[0035] Another object of the present application is to provide a route planning method of a cloud-edge collaborative flight management system, which comprises the following steps:

[0036] Step one, the route historical data analysis module analyzes all flight data of all flights uploaded by all airlines of the corresponding route by big data technology, obtains the optimal flight path of the corresponding aircraft, and sends the optimal flight path to the route optimization model; the weather data analysis module receives current weather data and historical same-period weather data from the weather center, predicts the weather condition on the flight route, and sends the weather condition to the route optimization model; when a thunderstorm, airspace control or similar emergency situation is encountered, the airspace data analysis module calculates corresponding real-time airspace information in combination with the current number of aircraft in the airspace and flight trajectory information, and sends the real-time airspace information to the route optimization model;

[0037] Step two, the route optimization model completes route optimization of the current flight according to data provided by the route historical data analysis module, the weather data analysis module, the airspace data analysis module and the aircraft performance database, and sends the optimized route information to the edge computing node and the airborne terminal.

[0038] Step three, when encountering wind shear scenes, thunderstorms or airspace control similar sudden situations, the node weather data analysis module receives real-time weather data from the weather center, airport weather radar and aircraft weather radar data, predicts the future fifteen minutes of wind shear, thunderstorm similar weather scenes that have a major impact on the flight of the aircraft, and sends the dangerous area coordinates and the recommended flight path to the onboard end through ACARS or CPDLC; the node airspace analysis module receives the position and speed information of the surrounding aircraft through ADS-B, constructs a real-time airspace situation map, and sends the current airspace situation to the escape path calculation module; the escape path calculation module receives data from the node weather data analysis module, the node airspace analysis module and the cloud platform flight path optimization model, calculates the corresponding escape path, and sends the corresponding escape path to the onboard end and the air traffic control center, and the air traffic control center decides whether the aircraft executes the escape path.

[0039] Preferably, in the step one, before the airline uploads the flight data of all flights, the corresponding data of each airline is encrypted by the federated learning algorithm module.

[0040] Preferably, in the step two, when the optimized flight path information is sent to the onboard end, the data sent to the onboard end is security verified.

[0041] Preferably, in the step three, when the escape instruction of the corresponding escape path is sent to the onboard end, the data sent to the onboard end is security verified.

[0042] Compared with the prior art, the flight management system based on cloud-end edge collaboration and the flight path planning method have the following beneficial effects:

[0043] 1. The flight management system based on cloud-end edge collaboration and the flight path planning method can solve the fundamental defects of the traditional FMS in dynamic planning capability, computing power limitation and the like, and provide an extensible, high-reliable and low-cost solution for the next generation flight management system, and is especially suitable for high-density airspace, emerging electric aviation (eVTOL) and unmanned aerial vehicle logistics scenarios.

[0044] 2. The flight management system based on cloud-end edge collaboration and the flight path planning method adopts the architecture of cloud, edge node and onboard FMS, solves the shortage of insufficient computing power of the onboard FMS under the premise of ensuring safety; on the basis of breaking through the computing power, a flight path model driven by federated learning is introduced, the local training model of each airline is used, and the parameters are encrypted and uploaded to the cloud for aggregation, so that the data privacy is protected and the global fuel efficiency is improved, the introduction of the AI algorithm reduces the rejection rate of the FMS generated path; the traditional avionics hardware is broken and the third party such as a weather supplier, an airline, an air traffic control and the like is allowed to participate in the operation calculation of the FMS, so that the problem of data island of the traditional FMS is solved.

[0045] 3、The application can improve the fuel efficiency of the flight path planning through cloud AI global optimization and edge real-time correction; the airspace utilization rate is improved and flight delays are reduced by introducing airspace data into the federated learning model; the computing power of the traditional FMS is improved through the cooperation of cloud-edge-onboard three-level computing power, and compared with the pure cloud solution, the demand for overall computing power is reduced and the efficiency is improved.

[0046] 4、The real-time performance and reliability of the application are enhanced. The edge computing node processes the aircraft approach or departure stage task, which is more accurate than the traditional FMS and has lower delay than the pure cloud mode.

[0047] 5、The application has good data privacy protection, and adopts federated learning to ensure that the airline data does not leave the local, meeting the requirements of GDPR / CCAR-121-R7. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The software architecture diagram of the flight management system based on cloud-edge cooperation in embodiment 1.

[0049] Figure 2 The flight path planning process schematic diagram of the application.

[0050] Figure 3 The flight path planning process schematic diagram when encountering a wind shear scenario.

[0051] Figure 4 The architecture schematic diagram of the flight management system based on cloud-edge cooperation in embodiment 2.

[0052] Figure 5 The architecture schematic diagram of the federated learning algorithm module. DETAILED DESCRIPTION

[0053] The related terms in the application are explained as follows:

[0054] FMS: flight management system;

[0055] ACPS: avionics core processing system;

[0056] IMA: integrated modular avionics;

[0057] ICAO: International Civil Aviation Organization;

[0058] CORSIA: International Aviation Carbon Offsetting and Reduction Scheme;

[0059] TOPS: Tera Operations Per Second, a unit of processor computing power; 1TOPS represents that the processor can perform one trillion operations per second;

[0060] ACARS: Aircraft Communications Addressing and Reporting System

[0061] AOC: Airline Operation Center

[0062] CPDLC: Controler Pilot Data Link Communication

[0063] ADS-B: Automatic Dependent Surveillance-Broadcast

[0064] Referring to Figures 1 to 5 The cloud-edge collaborative flight management system architecture and the scanning method thereof are further described.

[0065] Embodiment 1

[0066] As Figure 1 shown: the cloud-edge collaborative flight management system includes a cloud platform, an edge computing node, and an onboard terminal; the cloud platform includes a route history data analysis module, a meteorological data analysis module, an airspace data analysis module, an aircraft performance database, and a route optimization model; the edge computing node includes a node meteorological data analysis module, a node airspace analysis module, and a relief path calculation module; and the onboard terminal includes an onboard FMS and a data security verification module.

[0067] The node of the edge computing module is deployed near the airport, air traffic control center, navigation station or 5G base station, ensuring low-delay coverage of the terminal area, approach corridor and key areas. The node weather data analysis module receives real-time weather data from the weather center, airport weather radar and aircraft weather radar data, mainly predicts the weather field that has a major impact on takeoff and landing such as wind shear and thunderstorm in the next fifteen minutes, and sends the coordinates of the dangerous area and the recommended flight path to the data security verification module through ACARS or CPDLC (Controller Pilot Data Link Communication; controller-pilot data link communication). After confirming the safety of the data, it is sent to the onboard FMS. The node airspace analysis module receives the position and speed information of the surrounding aircraft through ADS-B (Automatic Dependent Surveillance-Broadcast), constructs a real-time airspace situation map, and sends the airspace situation to the escape path calculation module. The escape path calculation module receives the output of the node weather data analysis module, node airspace analysis module and cloud platform flight path optimization model, and calculates the escape path according to the provisions of ICAO DOC 4444 "Air Traffic Management", such as the minimum fuel consumption path, the fastest escape path, the optimal flight path, etc. The escape instruction is issued to the data security verification module, and after confirming the safety of the data, it is sent to the onboard FMS and air traffic control center. The air traffic control center decides whether the aircraft will execute the escape path.

[0068] The cloud platform is deployed on the ground and supports the "historical data learning + weather data analysis" collaborative mechanism through powerful computing power to optimize the flight path globally. The flight path historical data analysis module analyzes all flight data of the flight path to obtain the optimal flight path through big data technology. The weather data analysis module receives weather data and historical weather data from the weather center to predict the weather conditions on the flight path. The airspace data analysis module provides airspace information for the flight path optimization model in the event of sudden situations such as thunderstorms and airspace restrictions, combined with information such as the number of aircraft in the airspace and flight trajectories. The aircraft performance database is a complete aircraft performance database. Due to the relationship between computing power and storage space, the onboard FMS generally performs a 10% reduction on the complete aircraft performance database. The flight path optimization model completes the flight path optimization based on the data provided by the flight path historical data analysis module, weather data analysis module, airspace data analysis module and aircraft performance database, and sends the flight path information to the data security verification module. After confirming the safety of the data, it is sent to the onboard FMS.

[0069] For example, Figure 2The shown flight path planning flowchart. First, through the flight history data analysis module, the weather data analysis module, the airspace data analysis module and the flight performance database, the integration of multi-source real-time data such as airline strategy, weather, airspace and aircraft performance data is completed, and the data is processed to finally generate a candidate flight path set; second, the processed data is sent to the flight path optimization model, which generates constraint conditions according to the requirements of air traffic control, airlines and crews, and optimizes the candidate flight path set again. If a suitable optimal flight path cannot be automatically generated, manual intervention can be used to reduce the constraint condition requirements to finally generate the optimal flight path; third, the optimal flight path is sent to the edge computing node and the onboard FMS, respectively. The node weather data analysis module and the node airspace analysis module of the edge computing node analyze the weather data and airspace flow of the node area, and send the analysis data to the escape path calculation module to complete the real-time fine tuning of the optimal flight path to generate the final flight path and send it to the onboard FMS; finally, the data security verification module of the onboard FMS receives the cloud flight path data and the edge stage flight path data, respectively, and performs security verification on the data. After meeting the security verification conditions, the flight path data sent by the edge node is used preferentially. Only when the data sent by the edge node is invalid (security verification fails, not sent to the onboard end, etc.), the flight path data sent by the cloud is used. If both the sent data are invalid, the flight path data of the onboard FMS is used.

[0070] As Figure 3 The flight path planning flowchart under the wind shear scenario is shown. First, the node weather data analysis module of the edge computing completes the wind shear model prediction by combining the comprehensive weather radar, aircraft state, airspace data, etc.; second, according to the wind shear prediction result, it is confirmed whether to continue monitoring or to correct the path; third, the node airspace analysis module is used to complete the analysis of the airspace situation, and the escape path calculation module completes the path calculation of the shortest escape distance and the minimum fuel consumption by integrating all data and according to the regulations of ICAO DOC 4444《Air Traffic Management》; finally, after the air traffic control agrees, it is sent to the onboard FMS for execution under the condition of ensuring data security.

[0071] Example 2

[0072] As Figure 4As shown, Example 2 adds a federated learning algorithm module to Example 1. This module ensures data security for all airlines while allowing all participants to share the model through a process of local training → parameter encryption → secure aggregation → noise addition → model distribution. This module addresses the problems of limited data coverage by individual airlines and the inability of airlines to share core data (such as historical flight trajectories, fuel consumption, and engine performance) due to competition and compliance requirements, leading to insufficient data samples and poor generalization in traditional route optimization models. Through the federated learning algorithm module, each airline trains the model locally, uploading only encrypted model parameters; the original data remains local. The cloud aggregates the parameters to generate a global model, which can utilize industry-wide experience without direct data sharing. Finally, the data generated by the model is distributed to all participants.

[0073] like Figure 5 The diagram shows the structure of the federated learning algorithm module. This module is divided into a local training module and a federated learning core module. Through the local training module, each airline trains its model using local data (flight data / meteorological data, etc.), ensuring the data remains locally. The model parameters are then transmitted to the federated learning core module. The core module encrypts and aggregates the parameters of each airline's locally trained model, generating a global model. It then encrypts the global model parameters and pushes them back to each airline's local model, ensuring global model optimization while protecting airline data privacy.

[0074] The route planning method of the cloud-edge collaborative flight management system in this embodiment 2 is as follows:

[0075] Step 1: The route history data analysis module uses big data technology to analyze all flight data uploaded by all airlines along the corresponding route, derives the optimal flight path for the corresponding aircraft, and sends it to the route optimization model. The meteorological data analysis module receives current meteorological data and historical meteorological data for the same period from the meteorological center, predicts the meteorological conditions along the current flight route, and sends it to the route optimization model. When encountering emergencies such as thunderstorms or airspace control issues, the airspace data analysis module combines the current number of aircraft in the airspace and flight trajectory information to calculate the corresponding real-time airspace information and sends it to the route optimization model. Before airlines upload all flight data, the corresponding data from each airline is encrypted using a federated learning algorithm module.

[0076] Step two, the route optimization model completes the route optimization of the current flight according to the data provided by the route historical data analysis module, the weather data analysis module, the airspace data analysis module and the aircraft performance database, and sends the optimized route information to the edge computing node and the onboard terminal; wherein when the optimized route information is sent to the onboard terminal, the data security verification module will perform security verification on the data sent to the onboard terminal.

[0077] Step three, when encountering wind shear scenarios, thunderstorms or airspace control similar sudden situations, the node weather data analysis module receives real-time weather data from the weather center, airport weather radar and aircraft weather radar data, predicts the future fifteen minutes of wind shear, thunderstorm similar weather scenes that have a major impact on the flight of the aircraft, and sends the dangerous area coordinates and the recommended flight path to the onboard terminal through ACARS or CPDLC; the node airspace analysis module receives the position and speed information of the surrounding aircraft through ADS-B, constructs a real-time airspace situation map, and sends the current airspace situation to the escape path calculation module; the escape path calculation module receives the data of the node weather data analysis module, the node airspace analysis module and the cloud platform route optimization model, calculates the corresponding escape path, and issues the escape instruction corresponding to the escape path to the onboard terminal and the air traffic control center, and the air traffic control center decides whether the aircraft executes the escape path; wherein when the escape instruction corresponding to the escape path is issued to the onboard terminal, the data security verification module will perform security verification on the data sent to the onboard terminal.

[0078] The above is only a preferred embodiment of the present application, and is not intended to limit the present application. It should be noted that any modifications, equivalent replacements and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A cloud-edge collaborative based flight management system, characterized in that: It comprises: The airborne terminal is deployed on the aircraft for data security verification and flight management of the aircraft; The cloud platform is deployed on the ground for route optimization based on route history data learning, weather data learning, airspace data learning and aircraft performance data learning, and sends the optimized route information to the airborne terminal and the edge computing node; The edge computing node is deployed near the airport, air traffic control center, navigation station or 5G base station, and is used to predict weather phenomena and airspace conditions that have a major impact on the flight of the aircraft within the coverage area of the edge computing node, and send the coordinates of the dangerous area and the recommended flight path to the airborne terminal according to the prediction results; At the same time, a real-time airspace situation map is constructed, and the relief path is calculated and sent to the airborne terminal and the air traffic control center, and the air traffic control center decides whether the aircraft executes the relief path; The edge computing node comprises a node weather data analysis module, a node airspace analysis module and a relief path calculation module; The node weather data analysis module is used to receive real-time weather data from the weather center, airport weather radar and aircraft weather radar data, predict weather phenomena that have a major impact on the flight of the aircraft within the next fifteen minutes, and send the coordinates of the dangerous area and the recommended flight path to the airborne terminal through ACARS or CPDLC; The node airspace analysis module receives the position and speed information of the surrounding aircraft through ADS-B, constructs a real-time airspace situation map, and sends the current airspace situation to the relief path calculation module; The relief path calculation module receives data from the node weather data analysis module, the node airspace analysis module and the cloud platform route optimization model, calculates the corresponding relief path, and sends the relief instructions of the corresponding relief path to the airborne terminal and the air traffic control center, and the air traffic control center decides whether the aircraft executes the relief path. 2.The cloud-edge collaboration based flight management system of claim 1, wherein: The cloud platform comprises a route history data analysis module, a weather data analysis module, an airspace data analysis module, an aircraft performance database and a route optimization model; The route history data analysis module analyzes the flight data of all flights on the corresponding route to obtain the optimal flight path of the aircraft through big data technology, and sends it to the route optimization model; The weather data analysis module is used to receive current weather data and historical weather data from the weather center, predict the weather conditions on the flight route, and send them to the route optimization model; The airspace data analysis module is used to calculate the corresponding real-time airspace information when encountering thunderstorms, airspace control and other sudden situations, and send it to the route optimization model; The aircraft performance database is used to store the complete aircraft performance parameters of the corresponding aircraft; The route optimization model is used to complete the route optimization of the current flight according to the data provided by the route history data analysis module, the weather data analysis module, the airspace data analysis module and the aircraft performance database, and sends the optimized route information to the edge computing node and the airborne terminal. 3.The cloud-edge collaboration based flight management system of claim 2, wherein: The onboard end comprises a data security verification module and an onboard FMS; the data security verification module is used for security verification of all data sent by the cloud platform and the edge computing node to the onboard end, and the data is sent to the onboard FMS after the security verification and confirmation of data security; and the onboard FMS is used for flight management of the airplane. 4.The cloud-edge collaboration based flight management system of claim 3, wherein: It also comprises a federal learning algorithm module, which is used for encryption processing of corresponding data uploaded by each airline to the cloud platform. 5.The cloud-edge collaboration based flight management system of claim 4, wherein: The federal learning algorithm module comprises a local training module and a federal learning core module; the local training module is used for training of a model by each airline using local data, generation of a local model without data leaving the local area, and transmission of the trained model parameters to the federal learning core module; and the federal learning core module is used for encryption and aggregation of parameters of the local training model of each airline to generate a global model, and encryption and pushing of the global model parameters to the local model of each airline, so as to optimize the global model under the premise of ensuring the data privacy of each airline.

6. A flight route planning method based on a cloud-edge collaborative flight management system, characterized in that: It comprises the following steps: Step one: the route historical data analysis module analyzes all flight data of all flights uploaded by all airlines on the corresponding route by big data technology, obtains the optimal flight path of the corresponding airplane, and sends the optimal flight path to the route optimization model; the weather data analysis module receives current weather data and historical same-period weather data from the weather center, predicts the weather on the current flight route, and sends the weather to the route optimization model; when a thunderstorm, airspace control or similar sudden situation is encountered, the airspace data analysis module calculates the corresponding real-time airspace information in combination with the current number of airplanes in the airspace and flight trajectory information, and sends the real-time airspace information to the route optimization model; Step two: the route optimization model completes the route optimization of the current flight according to the data provided by the route historical data analysis module, the weather data analysis module, the airspace data analysis module and the airplane performance database, and sends the optimized route information to the edge computing node and the onboard end; Step three: when a wind shear scene, thunderstorm or airspace control or similar sudden situation is encountered, the node weather data analysis module receives real-time weather data from the weather center, airport weather radar and airplane weather radar data, predicts weather phenomena such as wind shear and thunderstorm that may have a major impact on the flight of the airplane in the next fifteen minutes, and sends the coordinates of the dangerous area and the recommended flight path to the onboard end through ACARS or CPDLC; the node airspace analysis module receives the position and speed information of surrounding airplanes through ADS-B, constructs a real-time airspace situation map, and sends the current airspace situation to the escape path calculation module; The escape path calculation module receives data from the node weather data analysis module, the node airspace analysis module and the cloud platform route optimization model, calculates the corresponding escape path, and sends the corresponding escape path to the onboard end and the air traffic control center, so that the air traffic control center decides whether the airplane executes the escape path.

7. The method of claim 6, wherein the method further comprises: receiving, by the cloud server, the flight plan from the UAV; and transmitting, by the cloud server, the flight plan to the UAV. In the step one, the corresponding data of each airline is encrypted by the federal learning algorithm module before the airline uploads all flight data of all flights.

8. The method of claim 7, wherein the method further comprises: receiving, by the cloud server, the flight plan from the UAV; and transmitting, by the cloud server, the flight plan to the UAV. The step two sends the optimized flight path information to the airborne terminal, and the data sent to the airborne terminal is verified.

9. The method of claim 8, wherein the method further comprises: receiving, by the cloud server, the flight plan from the UAV; and transmitting, by the cloud server, the flight plan to the UAV. The step three sends the corresponding release path to the airborne terminal, and the data sent to the airborne terminal is verified.

Images