İnsansız Hava Araçları Akustik Yön Kestirimi

Abstract

With the advances in unmanned aerial vehicle (UAV) technology, the variety of UAVs has increased and UAVs has become widely used in many applications. Having capabilities such as remote control, image transfer and load carrying has caused UAVs to be used in illegal applications such as smuggling, spying and terrorism as well as being used in applications beneficial to humanity. The use of UAVs in illegal applications also has brought about detection and countermeasure studies. In the literature, it has been seen that the problem of UAV detection has been solved with four main approaches such as RF signal analysis, image processing techniques, radar signal processing techniques and acoustic signal processing techniques. In acoustic signal processing techniques, the target does not need to be in the line of sight. In addition, these techniques are effective against autonomous targets. Acoustic signal processing techniques are generally preferred due to these advantages and their low cost. Within the scope of this thesis, it is aimed to develop an algorithm that can be used in UAV direction estimation by using the acoustic signals emitted throughout the flight environment during the UAV flight. Firstly, various parameters utilized in solving the array signal processing problems and the effects of these parameters on performance have been investigated. As a result of these investigations, it was decided to use a uniform circular array structure consisting of 24 microphones. Later, in order to examine the characteristics of the UAV acoustic signals, the acoustic signals in the environment during the UAV flight were recorded using this microphone array. Upon examination of the obtained data, it was seen that the UAV acoustic signals contain harmonic frequency components spread over a wide band. It has also been observed that the dominant frequency component of these harmonic frequency components varies in different UAV and different flight mode situations. During the algorithm development studies, the delay and sum beamforming (DASB), Capon beamforming (CB) and multiple signal classification (MUSIC) methods, which are widely used in the literature for narrowband signal case, were used. Incoherent methods (IM) and coherent signal subspace method (CSSM) were also used for the broadband signal case. Algorithm performances were examined using field data recorded according to various flight scenarios. When the output of the algorithms are examined, it has been seen that the direction estimation results performed by using the incoherent delay and sum beamforming method (IDASB) and the coherent signal subspace method are more effective than the incoherent multiple signal classification method (IMUSIC) at long distances. But it has been observed that in terms of resolution and target separation incoherent multiple signal classification method is more effective than other methods. When the developed algorithms were examined in terms of processing load, it was seen that the processing load increased parallel to the resolution increase in the algorithm.