CN-122019223-A - Aviation fault behavior tree dynamic handling method for perception in flight phase

CN122019223ACN 122019223 ACN122019223 ACN 122019223ACN-122019223-A

Abstract

The application provides a dynamic handling method of an aviation fault behavior tree perceived in a flight phase, which comprises the steps of establishing a fault handling database independent of a main program, storing priority mapping relations of each flight phase, establishing a modularized behavior tree, key fault handling logic and fault tolerant branches, injecting an updating strategy into a special nonvolatile NVM (non-volatile memory) storage area through a dynamic loading interface, performing CRC (cyclic redundancy check) check on loaded data among three channels, enabling the data to be applied only if the data stored in the three channels are consistent, generating a handling sequence of two-stage sequencing based on real-time flight phases, sequentially executing the behavior subtrees, monitoring overtime states, starting the fault tolerant branches to execute emergency operation after overtime or handling failure, and automatically, intelligently and efficiently handling a solution of multi-level faults so as to cope with complex and changeable fault scenes in aircraft flight.

Inventors

LI HUI
WAN TIANCAI
WANG HAIFENG
CHEN YINCHAO
LUO CHUNFENG
SHI KE

Assignees

中国航空工业集团公司成都飞机设计研究所

Dates

Publication Date: 20260512
Application Date: 20251227

Claims (10)

1. A method for dynamically handling an aeronautical fault behavioral tree perceived during a flight phase, the method comprising: establishing a fault handling database independent of a main program, and storing priority mapping relations of all flight phases; establishing a modularized behavior tree, key fault handling logic and fault tolerant branches; Injecting an updating strategy into the special nonvolatile NVM storage area through the dynamic loading interface, and performing CRC check between three channels on the loaded data, wherein only the three channels of storage data are consistent and can be applied; Generating a two-stage ordered treatment sequence based on the real-time flight phase; And sequentially executing the behavior subtrees, monitoring the overtime state, and starting the fault-tolerant branch to execute emergency operation after overtime or treatment failure.
2. The method of claim 1, wherein the creating a fault handling database independent of the main program, storing each flight phase priority mapping relationship, comprises: A structured fault handling database is generated based on the historical operational data and the aircraft fault pattern library.
3. The method of claim 2, wherein the fault handling database contains fault IDs, fault level 1-4, level 1 faults being highest priority faults, priority mapping of flight phases, priorities of take-off, cruise, landing, ID of action subtree binding independently.
4. A method according to claim 3, wherein said building a modular behavioral tree, critical fault handling logic and fault tolerant branches comprises: The node type of the behavior tree is classified into a root node, a control node and an execution node, wherein the root node is used as a behavior tree entry to connect all fault handling branches, the control node comprises a condition execution node and a sequence node, the sequence node is used for judging logic control and fault occurrence, and each node of the execution node corresponds to one fault handling action and is used for carrying out specific fault handling actions; And a subtree multiplexing architecture is adopted, each subtree corresponds to one type of fault scene, an action tree ID field in the database points to a specific action subtree file, the action subtree library is independently stored, zero modification access of the main program is realized through memory mapping, and only the subtree file is replaced during updating without modifying the database or the main program.
5. The method of claim 4, wherein the injecting the update policy into the dedicated nonvolatile NVM storage via the dynamic loading interface and performing a CRC check between the three channels of the loaded data, only if the three channels store data are consistent, comprises: Maintainers update the fault handling database and the modularized behavior subtrees in the ground system; uploading an encrypted data packet containing a CRC (cyclic redundancy check) code through a standardized airborne interface, automatically verifying the data integrity by an airborne dynamic loading module, and writing a database into a nonvolatile memory (NVM); and comparing the three-channel CRC values in the NVM region through the redundancy management bus, and accessing the dynamic data region through the memory mapping interface by the main flight control program without recompilation or authentication of the main program in the whole process.
6. The method of claim 5, wherein the behavior sub-tree is a treatment logic flow defined in XML format.
7. The method of claim 6, wherein the generating a two-stage ordered treatment sequence based on real-time flight phases comprises: The system periodically performs two-stage sorting on all the valid priority objects, wherein the first-stage sorting is sorting according to the ascending order of the stage priority values, namely the higher the value is, the more the fault level is, the earlier the fault level is, the same priority time is sorting according to the ascending order of the occurrence time stamp, namely the first fault is handled first, the generated handling sequence precisely drives the follow-up action subtree to dispatch, and when the flight stage is switched, the priorities of all the fault handling times are immediately recalculated and the sorting is refreshed.
8. The method of claim 7, wherein the sequentially executing the behavioural subtrees and monitoring the timeout state, starting the fault tolerant branch to perform the emergency operation after a timeout or treatment failure, comprises: The system sequentially schedules the behavior subtrees associated with each fault time according to the output sequence of the dynamic priority judgment; During the execution of the subtree, the system activates the child nodes according to the logic flow topological structure sequence, wherein the Action node executes specific operation instructions, the condition node detects physical state parameters in real time, and simultaneously strictly compares the current treatment duration with a timeout threshold preset in a database, and if the treatment is completed within the threshold and the state verification is successful, logs are recorded and returned; if the handling is overtime or the handling target is not reached, a preset fault-tolerant branch in the subtree is immediately activated, the degradation operation is executed, and for faults occurring in the critical flight phase, if the fault-tolerant handling still fails, the system automatically sends an emergency instruction.
9. The method of claim 8, wherein the emergency command comprises an emergency return.
10. The method of claim 8, wherein the behavior sub-tree is dynamically bound by a predefined behavior sub-tree ID field in a fault handling database, ensuring that handling logic exactly matches fault policies.

Description

Aviation fault behavior tree dynamic handling method for perception in flight phase Technical Field The application belongs to the technical field of flight control, and particularly relates to a dynamic handling method for an aviation fault behavior tree perceived in a flight stage. Background The complexity and high risk characteristics of aircraft flight systems present significant technical challenges. Aircraft are often made up of a number of highly complex systems, and failure of any one of the subsystems during flight can cause the aircraft to enter a special condition. It is worth noting that the consequences generated after faults occur in different systems are significantly different, and faults of the same type in the same system can also generate different influences in different flight phases. Furthermore, system faults in aircraft often exhibit concurrency and strong coupling characteristics, which make fault handling more complex and difficult. The method can rapidly and effectively cope with various faults in the airplane flight, and has important significance for improving the task completion rate and reducing the accident probability. Currently there is a lack of effective emergency handling means to meet this need, existing aircraft fault handling methods rely primarily on pilot manual operations or fixed rule based automated systems, but these methods suffer from the following significant drawbacks: the manual dependency is strong, and the pilot needs to quickly identify faults and perform complex operations in a short time, which is easily affected by human factors and is limited by the reaction speed. The traditional automation system adopts preprogrammed linear logic, and the stiff rule is difficult to cope with the dynamic scene of multi-fault concurrency, and the processing strategy cannot be flexibly adjusted according to the actual situation. When multiple failures occur simultaneously, existing systems lack an effective priority management mechanism, and cannot dynamically allocate a treatment order according to the severity of the failure, which may result in critical failures not being properly handled in time. The current aircraft fault handling adopts static priority, and the difference of the same fault in different flight phases cannot be distinguished. In addition, current aircraft fault strategies are coupled with the main program depth, and modification logic needs to be authenticated and loaded again, which may lead to aircraft shutdown. Therefore, there is an urgent need to develop a solution capable of automatically, intelligently and efficiently handling multi-level faults to cope with complex and varied fault scenarios in aircraft flight. Disclosure of Invention The invention aims to provide a dynamic handling method of an aviation fault behavior tree perceived in a flight phase, which is used for solving the problems that in the existing aircraft flight system, the updating needs overall authentication, the decision lacks scene adaptability and the handling flow is stiff. The application provides a dynamic handling method of an aviation fault behavior tree perceived in a flight phase, which comprises the following steps: establishing a fault handling database independent of a main program, and storing priority mapping relations of all flight phases; establishing a modularized behavior tree, key fault handling logic and fault tolerant branches; Injecting an updating strategy into the special nonvolatile NVM storage area through the dynamic loading interface, and performing CRC check between three channels on the loaded data, wherein only the three channels of storage data are consistent and can be applied; Generating a two-stage ordered treatment sequence based on the real-time flight phase; And sequentially executing the behavior subtrees, monitoring the overtime state, and starting the fault-tolerant branch to execute emergency operation after overtime or treatment failure. Preferably, the building a fault handling database independent of the main program, storing the priority mapping relation of each flight phase, includes: A structured fault handling database is generated based on the historical operational data and the aircraft fault pattern library. Preferably, the fault handling database contains fault IDs, fault level 1-4, level 1 faults are faults with highest priority, priority mapping of flight phases, priorities of take-off, cruising and landing are independently defined, and ID of action subtree binding is independently defined. Preferably, the building a modularized behavior tree, a key fault handling logic and a fault tolerant branch comprises: The node type of the behavior tree is classified into a root node, a control node and an execution node, wherein the root node is used as a behavior tree entry to connect all fault handling branches, the control node comprises a condition execution node and a sequence node, the sequence node is used for judging