Deep Anomaly Detection with Attention (DADA): A Novel Approach for Identifying Multipath Interference in Radar Signals

Table 2

Performance metrics for the seven methods (test set with two multipath signals).

Method	LFM NLFM (LN)			LFM BPSK (LB)			LFM QPSK (LQ)
	P	R	F1	P	R	F1	P	R	F1
CCF	0.639	1.000	0.779	0.671	0.997	0.802	0.667	1.000	0.800
SC	0.659	1.000	0.794	0.591	1.000	0.743	0.660	1.000	0.795
WRFCCF	0.675	0.950	0.789	0.610	0.700	0.651	0.621	0.927	0.744
AnoGAN	0.755	0.902	0.822	0.520	0.942	0.670	0.528	0.992	0.689
EGBAD	0.673	0.965	0.793	0.503	0.995	0.668	0.540	0.995	0.700
GANomaly	0.731	1.000	0.844	0.541	0.553	0.547	0.576	0.657	0.614
DADA	0.910	0.962	0.936	0.711	0.990	0.828	0.911	0.848	0.878