fastSLAM
fastSLAM with and without data association This project reports on the implementation and discussion of the FastSLAM algorithms. Three FastSLAM algorithms, (i) FastSLAM with known data association, and (ii) FastSLAM1.0 and (iii) FastSLAM2.0 with unknown data association, are implemented and tested in the landmark-based scenario. In addition, a mapping matrix norm induced error is proposed to estimate and analyze the SLAM performance with unknown data association. Finally, we tried to implement on the real dataset and provides some discussion and comments.
Video link: https://www.youtube.com/watch?v=l1vEpVRktko
Guideline for running code
FastSLAM with self-generated data: run.m
FastSLAM1.0 with self-generated data: run_unknown1.m
FastSLAM2.0 with self-generated data: run_unknown2.m
FastSLAM on Victoria Park Dataset: run_unknown2_vp.m