Deep learning based doppler radar for micro UAS detection and classification
Authors:
- 1 November 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 924-929
Abstract
In this paper, a radar sensor is proposed for the automated detection and classification of micro unmanned aerial systems (UASs), using Doppler signatures and their spectral correlation functions (SCFs). Our proposed system effectively detects and identifies UASs (within the radar beam width) by employing a Deep Belief Network (DBN) to classify the SCF signature patterns. The proposed system is experimentally verified using 3 UASs sensed with a 2.4 GHz continuous-wave doppler radar, which is set up in a laboratory environment. The experiment results show that a Doppler radar sensor is able to detect and classify UASs with an accuracy above 90% based on the automated classification of the radar signature SCF using DBN-based classifier.Keywords
This publication has 7 references indexed in Scilit:
- Machine Learning2014
- Gaussian-Bernoulli deep Boltzmann machine2013
- Cyclic spectral analysis in practiceMechanical Systems and Signal Processing, 2007
- A Fast Learning Algorithm for Deep Belief NetsNeural Computation, 2006
- Cyclostationarity: Half a century of researchSignal Processing, 2006
- A new approach to signal classification using spectral correlation and neural networks2005
- Higher-order cyclostationarity properties of sampled time-seriesSignal Processing, 1996