Next: Robot Architecture and Multiagent Up: Future Work Previous: Future Work
The extension of Prescott's method, together with the algorithms to compute diverting targets, has been shown to successfully encode the environment into a map that permits navigating from a starting point to the target. However, we would like to explore other mapping methods, so that the combination of the different methods adds robustness to the Navigation system. With the current mapping method, the robot needs to see at least three landmarks in order to be able to use the information stored in the map. We would like to develop some other mapping methods to cope with the situations in which the robot has very little information (i.e. less than three landmarks). These methods would be even more qualitative than our fuzzy extension of Prescott's method. We could, for example, look at the field of Spatial Cognition, which works with spatial relationships such as ``landmark X is at the left hand side of the line connecting landmark Y and landmark Z''.