Skip to main content

InfoTech e-conference

32nd International Conference on Information Technologies

Evolutionary Graph - Based SLAM

Ventseslav Shopov
In stitute of Robotics ‚Äď BAS, Plovdiv

Abstract: The simultaneous localization and mapping (SLAM) with evolutionary graph based approach is studied. The impact of re-sampling population on performance of evolutionary SLAM is experimentally studied. Two algorithms are compared: one with random re-sampling and one with greedy population re-sampling. With augmentation of number of nodes the greedy re-sampling crossing-over approach demonstrated significantly better performance.

Key words:
graph-based SLAM
evolutionary approach