Ashish Kirori
2018
The ability to use signage information to wayfind and determine the location of
facilities within buildings is an important component in the successful use of
the space. In reality, there are various types of signs (routes, services,
location, etc.) for general circulation and evacuation procedures, which
normally form a chain along the intended route that leads to the desired target
location within the premises or a place of safety; the signs do not work in
isolation. Despite the importance of signage information in helping occupant
identify and follow the intended route, the effectiveness of signage, depending
on the design of the signage system, the environmental conditions and the viewer
attributes, etc. have been generally ignored in most evacuation/pedestrian
models. A few evacuation models such as PEDROUTE, buildingEXODUS and MASSEgress
do have a representation of emergency exit signs allowing agents to detect signs
and use this information to find a way out of the structure. However, this is
mostly based on the detection and interaction with a single sign. Representing
the interaction between agents and a series of signs is crucial to properly
simulate people’s wayfinding behaviour, especially in an unfamiliar environment.
The work presented in this thesis is about a new signage-based
navigation model developed specifically to improve the representation of the
interaction between agents and series of signs in evacuation modelling (and
potentially circulation modelling). The enhancement to evacuation modelling in
terms of the agent wayfinding through this work includes: combining signage
(with direction) and navigational graph to expand agent’s visual perception of
the environment and sense of direction, introducing a preliminary form of
cognitive understanding of the building layout through memory and providing
individual level decision-making capability for wayfinding in both familiar and
unfamiliar environments. The new model allows the simulation of the agent’s
active wayfinding behaviour through detecting the signs in a chain to follow the
intended route. The model also allows the agents to build up and use individual
navigational experiences to search a way out when there is imperfect signage
information (e.g. an incomplete signage chain) or even a lack of signage
information.
The new signage-based navigation model was implemented
within the buildingEXODUS evacuation simulation tool using C++ programming
language. The model can also potentially be implemented within other evacuation
and circulation simulation tools to allow the study of the effectiveness of
signage systems in a built environment. The enhanced capability of the new model
has been verified through a series of verification cases and the improvement
over the existing signage model within buildingEXODUS has been demonstrated
through evacuation analysis performed over a hypothetical evacuation scenario.
.