Evaluation of use of 3D Maps in virtual navigation

The use of cartographic representations in 3D views accumulates the advantages of technological advances, but faces a lack of cartographic knowledge needed to build these representations. Consequences include misconceptions in map design that can inhibit or even impair the understanding of the features represented and their spatial relations. A common task in 3D map-use is virtual navigation supported by 3D topographical maps. This geographical task gathers different knowledge schema, whereby someone uses it to determine their position and orientation related to other elements in the landscape and uses these elements to move to other locations. However, the determination of relative orientation and the selection of reference points are not trivial tasks in virtual environments. This problem is exacerbated by the lack of knowledge of how users react to the new perspective and perceptive differences proposed by it. There is a need for the adaptation of cartographic representations into 3D to improve the communication offered by this kind of map. However, the internal representation from the map-user does not match cartographic maps, but some features must have a visual correlation to allow the information present in both representations to be related. The strategy adopted in this research, therefore, is to compare the 3D sketch maps with conventional 2D topographic maps, and identify which features can be used for doing it. To achieve this goal, we developed a testing methodology which combines a qualitative think-aloud protocol and evaluation by questionnaires in a community of expert users, i.e. cartographers and geographers. By identifying the number of reference points, the topology between them and the orientation for the whole sketch map we were able to find evidence of the success in the proposal to navigate with 3D topographical maps. In this research 43 volunteers were selected and their 179 sketch maps were evaluated by the qualitative methods described above as well as by the Kruskall-Wallis statistical method.