Appearance-based mapping and localisation using feature stability histograms

Proposed in this article is an appearance-based mapping and localisation method based on the human memory model, which is used to build a feature stability histogram (FSH) at each node in the robot topological map. FSH registers local feature stability over time through a voting scheme, and most stable features are considered for mapping and Bayesian localisation. Experimental results are presented, using omnidirectional images acquired through long-term acquisition considering: illumination changes, occlusions, random removal of features and perceptual aliasing. This method is able to adapt the internal node's representation through time to achieve global and local robot localisation.