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

Table 4

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

Method	LFM NLFM BPSK (LNB)			LFM NLFM QPSK (LNQ)			LFM BPSK QPSK (LBQ)			NLFM BPSK QPSK (NBQ)
	P	R	F1	P	R	F1	P	R	F1	P	R	F1
CCF	0.660	1.000	0.795	0.656	1.000	0.792	0.673	1.000	0.804	0.677	1.000	0.807
SC	0.591	1.000	0.743	0.664	1.000	0.798	0.583	1.000	0.736	0.584	1.000	0.738
WRFCCF	0.660	0.906	0.764	0.637	0.926	0.755	0.651	0.925	0.764	0.663	0.930	0.774
AnoGAN	0.540	0.950	0.688	0.557	0.918	0.693	0.522	0.985	0.682	0.536	0.963	0.689
EGBAD	0.545	0.985	0.701	0.516	0.986	0.677	0.562	0.986	0.716	0.513	0.995	0.677
GANomaly	0.762	0.622	0.685	0.948	0.740	0.831	0.545	0.846	0.663	0.852	0.600	0.704
DADA	0.759	0.951	0.845	0.872	0.965	0.916	0.736	0.991	0.845	0.740	0.978	0.843