TABLE 2.

Ablation study: influence of the amount of training data, number of sensors and type of model on the performance. We bolded results with the best mean error. Baseline models are trained using 20 datapoints and 8 sensors.

	Position error (mm)	Orientation error (deg)
Model-Based	1.1 ± 1.0	0.9 ± 0.8
Linear baseline	0.3 ± 0.3	0.3 ± 0.3
Polynomial baseline	0.2 ± 0.2	0.1 ± 0.1
Neural network (2 layers)	0.8 ± 0.6	0.8 ± 0.7
Linear (3 sensors)	1.0 ± 0.7	1.2 ± 1.0
Linear (4 sensors)	0.4 ± 0.3	0.5 ± 0.6
Polynomial (3 sensors)	0.9 ± 0.6	1.2 ± 0.9
Polynomial (4 sensors)	0.1 ± 0.1	0.1 ± 0.1
Polynomial (5 datapoints)	0.9 ± 0.9	1.1 ± 0.9
Polynomial (10 datapoints)	0.5 ± 0.5	0.6 ± 0.5
Polynomial (40 datapoints)	0.2 ± 0.3	0.3 ± 0.2