Acoustics-based detection and tracking of aircraft wakes

Abstract

The accurate detection and tracking of aircraft wake is important for minimizing the hazard to aircraft and increasing the efficiency of airports around the world. Passive methods of locating these wakes involve detecting the noise generated by wake acoustic sources and using it to estimate their position. These sources, which are nothing but vortices, generate noise in very low-frequency bands where traditional methods such as Acoustic Beamforming suffer from spatial aliasing and poor resolution. To address them, this project proposes a more robust source localization method based on Deep Learning, in the form of a Convolutional Neural Network. By studying the phenomenon of wake acoustic sources through famous literature and tests, a miniature problem resembling the real problem was simulated to demonstrate the feasibility of the method. The results showed that the network was able to resolve complex source distributions at low frequencies which suggest that a deep learning framework can be successfully built and applied for the detection and tracking of aircraft wakes. The Circulation of the flow around the aircraft wing also gives valuable information about the noise sources. It has applications in vortex trajectory tracing and vortex sound generation and thus, modeling the flow around the airfoil using Panel Methods has also been discussed.