Font Size: a A A

Research On Flight Training Evaluation Model And Its Method In Civil Aviation

Posted on:2014-01-21Degree:MasterType:Thesis
Country:ChinaCandidate:Y B ShaoFull Text:PDF
GTID:2272330473953834Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of civil aviation, the pilots become scarcely and the training demand of pilots is greatly increased. Due to long pilot training process and high cost, how to evaluate the pilot accurately in the training course becomes more and more important. Now, training evaluation is implemented only relying on experience of the coaches and simple statistics which lacks of corresponding scientific and objective evaluation method to help them make a more reasonable evaluation. Based on the above problems, the model is evaluated using both subjective and objective evaluation methods.The main flight parameters are presented to construct three flight training evaluation models in three important stages. The normal membership functions are given to normalize the each flight parameter to eliminate the ill effects of disagreements of the index type and unit based on the experts’ experience and advice.Fuzzy comprehensive evaluation method is provided briefly. Three stages of pilot training are evaluated respectively using the single stage fuzzy comprehensive evaluation method and the evaluation results are combined to evaluate the whole training process with multilevel fuzzy comprehensive evaluation method.Projection pursuit is simplely introduced and the commonly used projection index and projection model are also discussed. Adaptive differential evolution algorithm with bi-directional randomly mutation strategy is designed to optimize the projection index of projection pursuit model from a practical perspective. A new mutation operator is proposed which combines several typical mutation operators with the help of symbols function. The randomized crossover rate and adaptive mutation rate greatly simplify the complex parameter setting and improve the versatility of the algorithm. The improved algorithm has been tested on standard test functions and experimental results verifies the effectiveness of the proposed algorithm which provide the projection pursuit model a powerful optimization tool.Projection pursuit method is applied to the pilot training assessment. The projection pursuit classification model, projection pursuit evaluation model and projection pursuit principle component analysis are used to evaluate the three stages. The combination of three stages in the entire training process was also evaluated and the recommended model is given in training evaluation after comparison.A simple summary is made at the end of the article. The deficiency of the pilot training evaluation model proposed in this paper is analyzed and some opinion and outlook is given to evolve the work to do in the future.
Keywords/Search Tags:Flight training assessment, fuzzy comprehensive evaluation, projection pursuit, principal component analysis, differential evolution, bi-directional randomly search, multi-strategy variation
PDF Full Text Request
Related items