Deep Learning based Real Time Acoustic UAV Detection using Smartphone as Edge Computing Device

UAV technology is currently being used in various fields such as agriculture, communication, logistics, 
and is expected to be used in more fields in the future. 
The global Unmanned Aerial Vehicle (UAV) market was valued at US $56.7 Billion in 2021 and is estimated to reach a valuation of 
US $106.03 Billion by 2030 at a Compound Annual Growth Rate(CAGR) of 7.5% from 2022 to 2030. 
However, with great power comes great responsibility. 
Unfortunately, as drone technology advances, incidents of careless misuse, military surveillance, and malicious activity of drones have increased. 
Specially, malicious drones threaten to stadiums, prisons, and oil & gas because of their ability to carry payloads bypassing ground security. 
Drones also allow criminals to plot a heist, or hack into your phone or laptop. 
Drone detection is an important issue not only to prevent unfortunate accidents caused by drones, but also to prevent crime by detecting malicious drones.

1. Easy to use
Many solutions use machine learning models with high performance to detect drones. 
However, in order to apply a high-performance machine learning model, a computer with appropriate performance is required. 
Therefore, in real life, if the user needs to check in real time if there is a drone nearby, 
there is a possibility that there will be restrictions on its use. 

2. Use the application
Even when users do not have radar, microphones, etc., other methods are needed to locate the drone. 
Hence, in this paper, we propose a method for finding drones using a machine learning-based smartphone application. 
Worldwide smartphone ownership rate in 2021 is estimated to be 67%, the same level since 2018 [1]. 
Since most people are using smartphones, this solution is very practical. 
When a drone detection application is developed, users can install the application with just a few touches and check if there is a drone near.

[1] F. Laricchia, “Global smartphone penetration 2016-2021,” Statista, 17-Jan-2023. [Online]. Available: https://www.statista.com/statistics/203734/global-smartphone-penetration-per-capita-since-2005/. [Accessed: 30-Jan-2023].

Application
 Android Studio Electric Eel | 2022.1.1 (Build #AI-221.6008.13.2211.9477386, built on January 11, 2023)
  - Gradle Plugin Version 7.4.0
  - Gradle Version 7.5
  - Target SDK Version API 32
 External Library
   - jLibrosa 1.1.8
   - tensorflow lite 2.5.0
   - Android Wave Recorder 1.7.0