Design And Implementation Of Ground Fault Diagnosis Software Of UAV Flight Control System

At the beginning of birth of UAV, UAV is only used as a target drone, later widelyapplied to areas such as electronic interference, communication, surveillance,anti-submarine, air reconnaissance, which shows great practical value. But these mustbe built on the UAV flight control system working correctly. Currently, some of thefault diagnosis system for UAV flight control system can neither learn from thediagnostic experience nor improve the performance of fault diagnosis by changing thestructure of the system itself. In response to these issues, focusing on case-basedreasoning expert system, based on in-depth analysis of fault cases organizationalstructure, design and implement a set of intelligent UAV flight control system faultdiagnosis expert system:1. Complete system database design using relational database theory. Byreasonable decomposition relational schema of system, make it meet the requirementsof the third normal form of relational schema, and solve the problem of datadependencies due to updating data abnormally. By encapsulating database accessinterface ADO, it simplifies the operation of the data in the database. Encapsulates thedatabase access classes again to form a table access classes combined with a databasetable, and hides all the details about the database, leaving only the class interface.2. Use the case-based reasoning algorithm to realize the reasoning module of case.The principle of nearest neighbor reasoning algorithm is that retrieve a historical case infault case base which is the the most similar with the description of input faultinformation.3. Use the maintenance algorithm based on the coverage and the rough set to realizethe weight maintenance module of case signs. Comparatively speaking, the maintenancealgorithm based on the coverage is faster but not good enough, while the other is slowbut good enough.4. Use the attribute reduction algorithm based on discernibility matrix to realize thereduction module of case signs. Before the procedure of reduction, traditional algorithmneeds to take up a lot of time and space to generate the identification matrix, but the improved algorithm extracts attribute value while simplifying, which eliminates theneed for intermediate links generated identification matrix and saves time and space andimproves the efficiency of the implementation of the reduction algorithm.Finally, use the equivalence partitioning technology to test and validate the variousfunctional blocks of the fault diagnosis software. Through the analysis of the test results,the system successfully meets the design requirements, which proves the correctness ofsystem design.