Approximated Environment Features With Application to Trajectory Annotation


Indoor location-based services is an ever emerging field of research and application. In particular indoor navigation, i.e. positioning and route planning in freely walkable space, gets much attention. The philosophy of this paper is to incorporate isovists as a numerical representation of the local environment to enable new types of indoor LBS. This paper has got two main contributions. First, we define discrete isovists as an approximation of exact isovists using a simple ray casting approach. Second, we define, create, and use isovist feature cubes for semantic evaluations of floor plans as well as trajectory annotation.

6th IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2016)