Research on inflight parameter identification and icing location detection of the aircraft

This paper introduces a research on inflight parameter identification and icing location detection of the aircraft. A quasi-state nonlinear iced aircraft model is constructed. A command input of the aircraft control surfaces is designed in both longitudinal and lateral/directional planes, based on which the Hinf parameter identification algorithm is implemented to provide inflight estimate of the aircraft dynamic parameters. Parameter estimates are adopted as inputs for the icing detection block, which in this paper is built up by using the Probabilistic Neural Network. A database corresponding to different icing locations and icing severities of the aircraft was generated for the training and test of the detection net. Based on the test results, the icing detection work presented in this paper is believed to be applicable for our future studies.