The Mercator Representation of Spatial Knowledge

Ernest Davis
1983 International Joint Conference on Artificial Intelligence  
The MERCATOR program constructs a cognitive map from a sequence of scene descriptions. A new representation of two-dimensional geography was developed for this program. Objects are represented by sets of polygons; their boundaries, by sets of directed edges. The relative positions of objects are determined by connecting edges. A truthconditional semantics for this representation is presented, its strengths and weaknesses are evaluated, and it is compared to other AI representations of shape and position.
dblp:conf/ijcai/Davis83 fatcat:jwr6lep6czeylgk5m744yw2ivm