Table A1.

Summary of reviewed articles.

Art	Year	Est	AP	rPoint	fMap	fmRoom	mAlg	sAlg	mError	oError	sType
[41]	2021	pMap	168	57	IPIN2016	N	DRL		$0.92 \mathrm{m}$		RSSI
		pMap	589	1452	UTSIndoorLoc	Y	DRL		$1.72 \mathrm{m}$		RSSI
		pMap	520	993	UJIIndoorLoc	Y	DRL		$3.06 \mathrm{m}$	Only Building B1	RSSI
[22]	2021	pMap	96	80	JUIndoorLoc	Y	BayesNet	Dempster–Shafer		Accuracy = 80% between 3 and 3.6 m	RSSI
		pMap	520	993	UJIIndoorLoc	Y	BayesNet			Accuracy = 98% in 2 m	RSSI
[87]	2021	exp	7	116	1052 ${\mathrm{m}}^2$	Y	SISAE (NN)		$1.93 \mathrm{m}$	std = 1.34 m	RSSI
[44]	2021	exp	1	32	$49.9 {\mathrm{m}}^2$	N	CNN		$1.76 \mathrm{m}$		CSI
		exp	1	45	40 ${\mathrm{m}}^2$	N	CNN		$1.16 \mathrm{m}$		CSI
		exp	1	66	$48.8{\mathrm{m}}^2$	Y	CNN		$2.54 \mathrm{m}$		CSI
		exp	1	15	32 ${\mathrm{m}}^2$	N	CNN		$0.91 \mathrm{m}$		CSI
[45]	2021	sim	15	158	1160 ${\mathrm{m}}^2$	Y	ASDELM (ELM)			Accuracy = 85,90% in 1 m	CSI
		exp	22	47	384 ${\mathrm{m}}^2$	Y	ASDELM (ELM)			Accuracy = 77% in 1 m	CSI
[88]	2021	pMap	520	993	UJIIndoorLoc	Y	DNNIP			Accuracy = 89% building and floor	RSSI
[80]	2021	pMap	520	993	UJIIndoorLoc	Y	CHISEL (CNN)	autoencoder	$6.95$ $\mathrm{m}$	Accuracy = 99.6% building, 83.97% floor	RSSI
[46]	2021	exp	1	40	131.3 ${\mathrm{m}}^2$	Y	BPNN	adaptive genetic algorithm		Accuracy = 90.47% in 4 m	CSI
[30]	2021	pMap	520	993	UJIIndoorLoc	Y	NNELILS (NN)			67% to 78% localization accuracies	RSSI
		pMap	96	80	JUIndoorLoc	Y	NNELILS (NN)		$2.2 \mathrm{m}$ to $2.6 \mathrm{m}$		RSSI
[89]	2021	pMap	309	3951	Tampere	Y	CMDRNN (cnn)		$8.26 \mathrm{m}$	std = 1.31 m	RSSI
[21]	2021	pMap	520	993	UJIIndoorLoc	Y	CDAE i CNN		12.4 $\mathrm{m}$		RSSI
		pMap	152	670	Alcala Tutorial 2017	N	CDAE-CNN		1.05 $\mathrm{m}$		RSSI
[89]	2021	pMap	520	993	UJIIndoorLoc	Y	CMDRNN (cnn)		$8.26 \mathrm{m}$	std = 1.31 m	RSSI
[90]	2021	exp	113	30	3600 ${\mathrm{m}}^2$	Y	WiFiNet (cnn)			Accuracy = 91.89% in 2 m	RSSI
[81]	2021	pMap	520	993	UJIIndoorLoc	Y	DeepLocBox (NN)		9.07 $\mathrm{m}$		RSSI
[33]	2021	exp	15	150	200 ${\mathrm{m}}^2$	Y	SVM	M-LS	2.7 $\mathrm{m}$		RSSI
[47]	2021	exp	1	$N/A$	14 ${\mathrm{m}}^2$	N	NN		0.18 $\mathrm{m}$		CSI
		exp	2	$N/A$	18 ${\mathrm{m}}^2$	N	NN		0.03 $\mathrm{m}$		CSI
		exp	2	$N/A$	6.7 ${\mathrm{m}}^2$	Y	NN		0.08 $\mathrm{m}$		CSI
[48]	2021	exp	1	317	148.5 ${\mathrm{m}}^2$	Y	BLS(NN)		2.54 $\mathrm{m}$		CSI
		exp	1	176	126 ${\mathrm{m}}^2$	N	BLS(NN)		1.48 $\mathrm{m}$		CSI
[82]	2021	exp	6	132	460 ${\mathrm{m}}^2$	Y	Edgeloc(CapsNet)			99% under 2 m	RSSI
		pMap	520	993	UJIIndoorLoc	Y	Edgeloc(CapsNet)		7.93 $\mathrm{m}$		RSSI
[91]	2021	exp	1	210	600 ${\mathrm{m}}^2$	Y	MLR		4.03 $\mathrm{m}$		RSSI
[77]	2021	exp	436	654	WIFINE	Y	RNN		3.05 $\mathrm{m}$		RSSI
[92]	2021	exp	191	349	360 ${\mathrm{m}}^2$	N	DNN		1.08 $\mathrm{m}$		RSSI
[49]	2021	exp	1	$17,486$	CTW 2019 challenge	N	CNN		0.12 m		CSI
[93]	2021	exp	$N/A$	292	600 ${\mathrm{m}}^2$	Y	CNN		1.86 $\mathrm{m}$	Accuracy = 95% in 5.41 m	RSSI
		exp	$N/A$	262	1360 ${\mathrm{m}}^2$	Y	CNN		1.86 $\mathrm{m}$	Accuracy = 95% in 5.41 m	RSSI
[94]	2021	exp	12	680	6000 ${\mathrm{m}}^2$	N	DNN		3.6 $\mathrm{m}$		RSSI
		exp	12	170	6000 ${\mathrm{m}}^2$	N	DNN		3.7 $\mathrm{m}$		RSSI
		exp	12	40	6000 ${\mathrm{m}}^2$	N	DNN		3.8 $\mathrm{m}$		RSSI
[95]	2021	exp	4	54	69.35 ${\mathrm{m}}^2$	Y	ANN			Accuracy = 13.84% < 0.5 m & 23.07% 0.5 < 1 m	RSSI
[50]	2020	exp	3	21	45 ${\mathrm{m}}^2$	Y	CNN		1.27 $\mathrm{m}$	std = 0.68 m	CSI
[36]	2020	exp	4	264	112 ${\mathrm{m}}^2$	N	RF		1.68 $\mathrm{m}$		RSSI
[51]	2020	exp	4	63	75.6 ${\mathrm{m}}^2$	N	CNN		1.61 $\mathrm{m}$		CSI
		exp	4	$N/A$	44.8 ${\mathrm{m}}^2$	N	CNN		1.11 $\mathrm{m}$		CSI
		exp	4	$N/A$	16 ${\mathrm{m}}^2$	N	CNN		0.98 $\mathrm{m}$		CSI
[96]	2020	exp	4	10	169 ${\mathrm{m}}^2$	Y	CNN		0.98 $\mathrm{m}$		RSSI
[54]	2020	exp	5	34	55 ${\mathrm{m}}^2$	N	MLP	Regression	0.37 $\mathrm{m}$	RMSE = 0.84 m	SNR

art: Article; mAlg: Main algorithm used; est: Experimental or pMapulated study; sAlg: Other algorithms used in the study; AP: APs used; mError: Mean Error; rPoint: Reference Points used in offline phase; oError: Other metrics reported in the study; fMap: Size of experimental room or radio-map used; sType: Signal type used; fmRoom: Rooms used in exp/pMap.