Research And Implementation Of Unmanned Aerial Vehicle Flight Training Assessment System

The current flight training assessment system is mainly based on manual evaluation by flight training experts,which is extremely susceptible to personal experience,personal characteristics and other factors.This will lead to low reliability of the results of the evaluation of flight student performance and increase the cost of time of flight training.At the same time,the general flight training system has poor scalability.After the system is developed and put into use,the original system will be heavily modified and may even need to be re-developed if new requirements are to be added.Which shortens the entire life cycle of the flight training system and increases the cost of development and maintenance of the flight training system.In order to solve some problems in the above flight training assessment system,this paper is proposed to use the method of entropy correction G2 in flight training performance evaluation to solve the shortcomings of manual evaluation of experts,which based on the X-Plane platform-related visual display theory and technology.The designed system has adopted independent methods such as training models,flight training subjects?visual display systems and flight evaluation methods,which is greatly improved the versatility and maintainability of the system.The main research contents shown in this paper are as follows,Based on the characteristics of flight data in flight training,a classification method based on random forest is used to identify the flight phases in flight training.Thus,accurate recognition and fast division of the unmanned aerial vehicle flight training stage are realized.Aiming at the shortcomings of manual assessment by flight training experts,a method for determining the weights of assessment indicators based on the entropy modified G2 method is proposed to determine the weights of indicators at each flight stage.Furthermore,through such a combination of subjective and objective assessment methods,a more reliable assessment of flight training performance is achieved.In view of the lack of prediction of flight training results in the flight training assessment system,this paper is used a gray prediction model to achieve a sustainable prediction of the training results of flight students,so that flight training experts can better understand training situation of flight students and timely adjust flight training arrangements.This will greatly improves the efficiency of flight training.Based on the characteristics of the X-Plane platform,the overall design of the flight training assessment system is determined and the flight training assessment system is completed.In addition,the function and performance of the system are tested and verified.