UAV Detection Using the Cepstral Feature with Logistic Regression

Abstract

The unmanned aerial vehicle system has been employed in various aspects, but the need for anti-unmanned aerial vehicle system technology is emerging due to privacy violation and bypass of a security system. In this paper, we propose a detection algorithm for an unmanned aerial vehicle system using acoustic sensors. The learned detection model is employed for the acoustic signal of the unmanned aerial vehicle system to obtain higher recognition performance. The cepstrum of the acoustic signal sampled during operation of the unmanned aerial vehicle system is applied to the feature vector and the logistic regression model is developed for the detection model. The learned model is verified through ten arbitrary cross-validations. The detection error for verification data is about 17.48%.