Distributed Spatial Reasoning

Abstract

Reasoning about perceived context information is fundamental for autonomous systems to draw conclusions about their spatial contexts and adapt to changing situations at runtime. According to our focus on qualitative spatial relations, this chapter builds on the qualitative abstractions presented in Chapter 3 and is concerned with inferring new from known relations. In Section 4.1, we first give an overview of approaches for qualitative spatial reasoning, with a focus on binary relations and their properties as well as compositional reasoning about them. In Section 4.2, we discuss reasoning about positional and directional relations in more detail and present composition tables for distance and orientation relations. The main part of this chapter is about an algorithm for inferring and distributing spatial relationships among autonomous digital artifacts over multiple hops, which is presented in Section 4.3 and evaluated by simulation means in Section 4.4. The general idea is that artifacts recognize spatial relations to others in their vicinity by sharing spatial context information, and infer relations to artifacts out of communication range (i.e. data cannot be exchanged directly, but just via other artifacts) by means of qualitative spatial reasoning techniques which are discussed in this chapter. In particular, the algorithm makes use of composition tables or the properties of binary relations in order to provide artifacts, which are out of communication range, with an awareness about their spatial relationships to each other. Section 4.5 eventually discusses benefits and drawbacks of the presented qualitative spatial reasoning approaches and elaborates on their use for autonomous embedded systems.