Doors in the Sky: Detection, Localization and Classification of Aerial Vehicles using Laser Mesh

TL;DR

This paper introduces a laser mesh-based system for detecting, localizing, and classifying aerial vehicles, offering an alternative to radar that can identify stealth and small UAVs effectively.

Contribution

It presents a novel laser mesh technique combined with machine learning for aerial vehicle detection, localization, and classification, outperforming traditional radar in certain scenarios.

Findings

Achieves simultaneous detection, classification, localization, and tracking of aerial targets.

Demonstrates effectiveness through simulation with various ML algorithms.

Creates an artificial database for training and testing ML models.

Abstract

The stealth technology and unmanned aerial vehicles (UAVs) are expected to dominate current and future aerial warfare. The radar systems at their maximum operating ranges, however, are not always able to detect stealth and small UAVs mainly due to their small radar cross-sections and/or low altitudes. In this paper, a novel technique as an alternative to radar technology is proposed. The proposed approach is based on creating a mesh structure of laser beams initiated from aerial platforms towards the ground. The laser mesh acts as a virtual net in the sky. Any aerial vehicle disrupting the path of the laser beams are detected and subsequently localized and tracked. As an additional feature, steering of the beams can be used for increased coverage and improved localization and classification performance. A database of different types of aerial vehicles is created artificially based on Gaussian distributions. The database is used to develop several machine learning (ML) models using different algorithms to classify a target. Overall, we demonstrated through simulations that our proposed model achieves simultaneous detection, classification, localization, and tracking of a target.