Geoscience Reference
In-Depth Information
in outdoor environments. Examples are hierarchical paths (Fu et al.
2006
), paths
minimizing route complexity (Duckham and Kulik
2003
; Richter and Duckham
2008
) or optimizing risks along the described routes (Grum
2005
). The major
advantage of those algorithms is their more qualitative description of routes and
their changed embedded cost function, simplifying the use and understanding
of the calculated routes and as such improving the entire act of navigation and
wayfinding.
Algorithms for 3D indoor navigation are currently restricted to Dijkstra or
derived algorithms. The results of those shortest path algorithms not necessar-
ily return realistic paths in terms of what an unfamiliar indoor wayfinder would
need to navigate this building, i.e. using complex intersections, avoiding main
walking areas etc. To date, only few researchers have attempted to approach algo-
rithms for indoor navigation differently, for example incorporating dynamic events
(Musliman et al.
2008
), or modelling evacuation situations (Atila et al.
2013
;
Vanclooster et al.
2012
). However, the need for more cognitively rich algorithms
is even more pronounced in indoor space than outdoors. This has its origin in the
explicit distinctiveness in structure, constraints and usage between indoor and out-
door environments. Outdoor environments are commonly described as continuous
with little constraints, while the perception of buildings is strongly influenced by
the architectural enclosures (Li
2008
; Walton and Worboys
2009
). Also, wayfind-
ing tasks in multi-level buildings have proven to be more challenging than out-
doors, for reasons of disorientation (due to multiple floor levels and staircases),
and less visual aid (e.g. landmarks are less obviously recognizable; corners and
narrow corridors prevent a complete overview) (Hölscher et al.
2007
). As such,
building occupants are faced with a deficient perspective on the building structure,
influencing their movement behaviour (Hölscher et al.
2007
). Algorithms devel-
oped to support a smooth navigation will have to consider these intricacies.
The main goal of this chapter is to translate existing outdoor cognitive algo-
rithms to an indoor environment to provide in indoor route calculations that are
more aligned to indoor human wayfinding behaviour. In a first phase, the original
algorithm is implemented in indoor environments and tested in terms of its effi-
ciency to reduce navigational complexity of the routes and as such improve the
cognitive route instructions. The tests consist of comparing the paths suggested by
the cognitive algorithm with those of the shortest path variant in indoor spaces.
Also, the results indoor will be compared to the results obtained by the original
algorithm. In this chapter, we currently focus on the implementation and applica-
bility testing of the least risk path algorithm as described by Grum (
2005
). Later
on, we are also planning to integrate the simplest path algorithm in indoor environ-
ments and develop a more general cognitive algorithm.
The remainder of the chapter is organized as follows.
Section 2
elaborates on
the definition of the least risk path algorithm and its relationship to other cogni-
tive algorithms and the shortest path algorithm. In
Sect. 3
, the indoor dataset is pre-
sented in combination with the choices and assumptions made when developing
the indoor network model. In the case study in
Sect. 4
, the outdoor least risk path
algorithm is duplicated and implemented in an indoor setting with multiple analyses
