Deep Learning-Based Localization with Urban Electromagnetic and Geographic Information

Table 2

Comparative results of different algorithms.

Algorithms	Positioning errors (m) with varied number of sensors
	48	12	8	4
DeepMTL	19.00	203.08	223.53	222.59
U-net	5.48	35.95	100.69	140.13
VGG-based proposed method	2.36	9.92	16.40	50.87
ResNet-based proposed method	2.28	7.34	10.16	37.81