Feature Extraction from Helicopter Using Time Domain Back Scattered Pulses

Authors

Abstract

Nowadays, due to the appearance of advanced aerial attack weapons, the usage of modern defence systems such as aerial target recognition systems could play a prominent role in a successful defence. The micromotion of rotating structures of radar targets is used, in many cases, as discriminant feature for their recognition. Possessing a general and accurate database of targets of interest can extremely affect the performance of feature extraction algorithms. Besides, the modeling of radar returns which contain informations about structural properties of the target, provides possibility of controlling the radar parameters and capability of arbitrarily selecting the target specifications. In this paper, the rotational micro-motions of the hovering helicopter main rotor blade tips are modeled as individual features of each helicopter in backscattered pulses with high range resolution ability. Although, the implementation of a pulse as narrow as the pulse used in this paper is not realizable with available technologies, it provides time domain analyzing of blade's micro-motions without taking their radar cross sections into account. In this paper, an algorithm is also designed and presented for estimating the rotational velocity of main rotor blade tips (i.e. v_rot) of four helicopter models. When the radar parameters such as pulse repetition frequency (PRF) and time on target (TOT) are improperly allocated, the backscattered pulses are not suitable to be used in the proposed v_rot estimation algorithm. Hence, above mentioned radar parameters are also evaluated.

Keywords