Comparing indoor and outdoor network models for automatically calculating turns

The goal of this paper is to compare several indoor and outdoor network models for wayfinding, on their suitability for automatically calculating turns. Automatic turn calculations are of relevance in providing improved cognitive algorithms for route guidance, as it has been widely recognised that routes with minimal angular deviations are easier to follow. It is demonstrated that the currently available indoor network models do not allow accurate calculation of the number of turns along a path, while the common outdoor route networks do. This discrepancy is found to be rooted in an inconsistent definition of indoor decision nodes, which in turn is linked to the inherent differences in space structure between indoor and outdoor environments. Additionally, it is proven that these also have a major influence on the generation of accurate indoor route instructions. Recommendations for future research within the context of both turn calculations and verbalisations of directional changes are made, as well as in the broader context of indoor spatial analyses.