Integrating fuzzy topological maps and fuzzy geometric maps for behavior-based robots

摘要: In behavior-based robots, planning is necessary to elaborate abstract plans that resolve complex navigational tasks. Usually maps of the environment are used plan robot motion and Two types have been mainly used: metric topological maps. Both present advantages weakness so several integration approaches proposed in literature. However, many conducted build a global representation model, techniques not fitted profit from both kinds information. We propose models into hybrid deliberative-reactive architecture through path algorithm based on A ∗ hierarchical map with two levels abstraction. The contains required information take advantage modeling. On one hand, model fuzzy perceptual allows classify distinguished places, other built using regions possibility shape segments, which later grid-based approach decide use most appropriate navigate world depending minimum-cost safety criteria. Experiments simulation real officelike shown for validating integrated architecture. © 2002 Wiley Periodicals, Inc.