MISD-SLAM: Multimodal Semantic SLAM for Dynamic Environments

Table 2

Results of absolute trajectory error (ATE) and improvements of processed PL-SVO compared to original PL-SVO in dynamic sequences.

Sequence	PL-SVO [37]			Processed PL-SVO			Improvements
	RMSE	Mean	S.D.	RMSE	Mean	S.D.	RMSE	Mean	S.D.
Sitting_static	0.0084	0.0077	0.0033	0.0064	0.0056	0.0031	23.81%	27.27%	6.06%
Walking_static	0.0075	0.0065	0.0037	0.0057	0.0048	0.0030	24.00%	26.15%	18.92%
Walking_xyz	0.2811	0.2409	0.1363	0.0230	0.0199	0.0116	90.96%	90.74%	91.49%
Walking_rpy	0.1770	0.1541	0.0872	0.0357	0.0275	0.0228	79.83%	82.15%	73.85%
Walking_halfsphere	0.1149	0.1112	0.0289	0.0401	0.0320	0.0242	65.10%	71.22%	16.26%