Research On Flight Performance Evaluation In Plateau Airport Terminal Area Based On BP Neural Network

The flight performance evaluation is one of the important ways to improve the safety of civil aviation operation, which has great significance in ensuring the operation safety and efficiency of the plateau complex airport. However, there’s not a perfect flight performance evaluation system for the performance quantitative assessment owning to lack of experience in supervising and guaranteeing plateau operation safety.This paper focused on the research of pilots’ eye-movement characteristics and flight performance evaluation in plateau complex airport terminal area. Firstly, by further arranging and digging the eye movement and performance experimental data, repeated measurements and multiple analysis of variance method was applied to analysis the interactive influence on eye-movement data, such as fixation duration and fixation frequency, under the condition of different degree of hypoxia and task difficulty. Then pilots’ attention transfer laws under the different hypoxia conditions were expanding explored based on the theory of Markov chain. All these work made the existing research conclusions more complete. Secondly, Google Earth software was used for simulating the flight path of the non-hypoxia and hypoxia group of three plateau airports in order to analyze the flight path differences in different hypoxia conditions, the special segment, special point height and the average slope information were also discussed. The method and contents of the qualitative evaluation of flight training quality and pilots’performance have been enriched through the visual flight path analysis. Thirdly, according to the above and existing research results, the method of literature research and expert consultation were adopted to extracted the complex plateau airport terminal area flight performance evaluation indicators, and then the partial correlation analysis were applied to determine seven key indicators about the flight performance evaluation. Finally, the flight performance evaluation model based on BP neural network of plateau complex airport terminal area was built by using MATLAB as modeling platform so as to achieve the training, testing and performance prediction of the model. The result showed a high degree of convergence and small error as well as a good performance evaluation level.In this paper, the further study of relation between pilots’ eye-movements and flight performance was conducted which grounds on plateau complex airport experimental data.The plateau complex airport terminal area performance evaluation model based on BP neural network is driven by flight performance evaluation research upon the experimental data, which is significant to the future in-depth research on big data that promotes civil aviation flight performance quantitative evaluation.