Landmark based corridor discrimination

For a mobile robot to perform some sort of useful function it usually must have some sort of global understanding of its environment. This is usually expressed in the form of map. Through real-world experiments, using a mobile robot inspired by insect visual guidance, we present results showing the performance of a mobile robot in recognising a previously encountered corridor environment and discriminating between various corridors. This is achieved through the building and refinement of maps based on the observation of simple landmarks en route.<br />