Abstract:
The implementation of a particle filter (PF)
for vision-based simultaneous localisation and
mapping (SLAM) for a mobile robot in an unstructured
indoor environment is presented in
this paper. Variations to standard PF are proposed
to remedy the sample impoverishment
problem in bearing-only SLAM. A CCD camera
mounted on the robot is used as the measuring
device and image quality is incorporated into
data association, PF update and map management.
A passive path control strategy to maintain
the accuracy of the SLAM process is also
illustrated. Experimental results from an implementation
using real-life data acquired from
a Pioneer robot are included to demonstrate
the effectiveness of our approach.
