New helicopter model identification method based on flight test data

doi:10.1017/s0001924000005789

New helicopter model identification method based on flight test data

2016-01-27

0

PRE

AI

DOI：10.1017/s0001924000005789

原文链接

原文求助

摘要

En 中文

AbstractA helicopter model has been identified and validated for flight conditions defined by altitudes, speeds, loadings and centre of gravity positions. To identify the helicopter models, 2-3-1-1 multistep control inputs were performed by the pilot to excite all helicopter modes. Then, each estimated signal has to remain in tolerance margins defined by the Federal Aviation Administration. Three methods were used to observe the system outputs from its states: a fuzzy logic method, a linear method optimised with a neural network algorithm and a classical method. Because of random effects when gathering data, classical method did not give good enough results. The fuzzy logic method was not robust enough so that output plots showed peaks that could be felt by the pilot. Then, because the model could be implemented in a simulator for the pilot training, the pilot feedback is very useful in order to compare the reality with the results of the mathematical model. When the outputs are obtained from the measured state variables, the linear method gave the best results.

New helicopter model identification method based on flight test data

摘要

期刊

机构

引用论文