Target localization using machine learning

Abstract

The miniaturization of sensors has made possible the use of these tiny components in hostile environments for monitoring purposes in the form of sensor networks. Due to the fact that these networks often work in a data centric manner, it is desirable to use machine learning techniques in data aggregation and control. In this paper we give a brief introduction to sensor networks. One of the first attempts to solve the Geolocation problem using Support Vector Regression (SVR) is then discussed. We propose a method to localize a stationary, hostile radar using the Time Difference of Arrival (TDoA) of a characteristic pulse emitted by the radar. Our proposal uses three different Unmanned Aerial ..