Delphi-DNN:一种进近阶段飞行质量评估方法OA

Aiming at the problem of evaluating the flight quality in the landing approach phase,a method of construct-ing the index system and determining the weights based on the Delphi method and neural network regression analysis is proposed.Firstly,the evaluation objectives are clarified and six key indicators including altitude deviation,approach speed deviation,pitch angle deviation,slope,horizontal acceleration change and vertical acceleration are selected and weights are initially assigned.Then,experts in the field of aviation are invited to conduct multiple rounds of scoring through the Delphi method,and the weights of the indicators are calculated by combining the neural networks.Finally,these indicators are used as input features,and the final model is applied to the actual flight quality evaluation.The re-sults show that the method can accurately integrate expert knowledge and data-driven prediction to improve the accu-racy and reliability of the assessment.And the method provides a scientific and dynamic solution for flight quality evalu-ation.