Getting inspired from bees to perform large scale visual precise navigation

In this paper we present a mapping paradigm for large scale precise navigation that takes inspiration from the bees´ large scale navigation behavior. Navigation is performed through the use of two distinct sets of landmarks: global landmarks guide roughly the agent to a place, with local landmarks the agent performs very precise motion to the final destination. The paradigm is equivalent to a map composed of two distinct levels: with the first level, the agent navigates from place to place (topological navigation) following the global potential function. When the agent is close to the place of interest then it switches to a finer map of the specific place represented by a local potential function (second level). The general idea behind the article (2 levels hierarchical mapping and navigation) has been successfully used so far in the robotics literature. By the way, a robot that uses only one sensor, a color camera, and operates in real time with global and local self extracted natural unprocessed visual landmarks has not been presented yet. A prototype of the motion algorithm has been developed and diverse experiments and tests operating with a nomad 200 have been performed to prove the performance of the topological algorithm.