Unbiased Model-Agnostic Metalearning Algorithm for Learning Target-Driven Visual Navigation Policy

Table 2

Comparison results of standard deviation for SPL and SR in the different scene types.

Scenario	Method	Scene instance								SD
		bedroom		kitchen		livingroom		bathroom
		SPL	SR	SPL	SR	SPL	SR	SPL	SR	SPL	SR
Trained target in unseen environment	TDVG	1.42	9.14	9.81	17.7	15.6	30.8	5.45	13.9	6.07	9.29
	MPSL	4.26	12.3	10.6	22.8	2.47	9.34	10.5	25.8	4.21	7.97
	GCN	11.4	25.8	15.1	32.5	5.43	12.1	14.3	28.9	4.38	8.91
	Our(glo)	9.25	20.5	17.4	36.8	8.56	18.4	12.1	31.3	4.01	8.76
	Our(GC-glo)	8.34	15.4	11.4	28.6	12.1	20.4	17.7	36.6	3.90	9.32
	Our(VL-glo)	7.2	24.7	14.5	31.4	9.81	19.1	14.8	27.2	3.70	5.14
	Our(UM-glo)	10.1	33.6	15.5	35.7	13.5	26.8	17.2	34.7	3.05	3.85