A Collective Ecosystem Platform On UAV Reliability

The development of UAV has a history of more than 100 years.With the continuous improvement of science and technology and the change of people's demand,UAV has gradually moved from the military field to the civilian field.At present,UAV has been used well in agricultural,aerial photography,meteorological monitoring,forest fire prevention,Electric power inspection,logistics transportation,etc.While bringing great convenience to people,people have put forward higher and higher requirements on the reliability level of UAV.However,UAV malfunctions still occur frequently.The UAV failure event accumulates a large amount of failure data,which contains abundant UAV failure information.These data,in the form of text data described in natural language,exist in the failure posts of the UAV community and in the news of the UAV accident.They are loosely distributed and not used efficiently.How to use these data efficiently to improve the reliability of UAV has become a problem worth studying.To solve this problem,this thesis proposes a fault diagnosis model and a Failure Mode and Effects Analysis model based on the text data of uav fault description,and then designs and implements a collective ecosystem platform for UAV reliability.The main work of this thesis is as follows:1.A total of 9135 UAV failure posts from the UAV community,276 UAV accident news from news websites and 541 UAV reliability literature from multiple literatures databases were collected.Based on these data,we create the UAV Failure Analysis Data Set – UAV-FAD,which contains the annotation information of the UAV failure category,failure risk factor and failure core entity,and provides the possibility of fault diagnosis and failure mode and effects analysis based on the UAV failure description text data.2.Based on UAV-FAD,the UAV fault diagnosis model–BERT-Diagnosis and UAV Failure Mode and Effects Analysis model–BERT-FMEA model are implemented,and the problem of fault diagnosis,failure mode and effect analysis based on UAV fault description text data is solved.3.Based on BERT-Diagnosis and BERT-FMEA,a collective ecosystem platform for UAV reliability is designed and implemented.It solves the problems of low troubleshooting efficiency,disorganized forum posts,single user group and low utilization rate of historical posts in the existing UAV community.