부산대학교 도서관

Parafoil is a kind of flexible aircraft, which has strong load capacity and long flight time but is easily disturbed by wind field. In the homing stage of parafoil from a high-altitude wind field to a low-altitude wind field, the low-altitude wind field is unmeasurable, which has a bad effect on the parafoil trajectory planning. To solve this problem, longitudinal prediction of the low-altitude wind field is proposed by intelligent processing of the high-altitude wind field data estimated by the parafoil. Since spatial wind field has the characteristics of hierarchical recursion and dynamic change, a deep deterministic policy gradient prediction model with Elman neural network as the core is proposed in this paper. Finally, the prediction effect of high accuracy and low-level precision attenuation, which provide reference information for the parafoil track planning, is realized. [ABSTRACT FROM AUTHOR]