Michael Joyce
2024
This research addresses a significant limitation in current evacuation
simulation models by introducing an innovative method for representing movement
assistance devices, such as evacuation chairs, hospital beds, wheelchairs,
rescue sheets, and stretchers. These devices play a crucial role in evacuating
people with reduced mobility (PRM). Existing models often oversimplify spatial
constraints (size and shape) and kinematic constraints (movement speed and
manoeuvrability) inherent to these devices, leading to inaccuracies in
simulation outcomes. This research draws inspiration from other fields of study,
primarily autonomous robotics, to review methods for their potential
applicability in improving evacuation models. The research introduces a novel
model called HEPTAD (Hospital Evacuation Planning Tool for Assistance Devices),
designed to account for spatial and kinematic constraints of assistance
devices while route-finding. The model involves the integration of a network
within a 3D configuration space (C-Space), accounting for spatial and kinematic
constraints. An advanced model, called ObjectController, is developed which
extends HEPTAD to incorporate interactions between multiple devices and
pedestrians using a novel combination of social forces and velocity obstacles,
enabling realistic navigation. Object-Controller interfaces with established
evacuation simulation model EXODUS to control both device and pedestrian
movements. The developed model is tested through component and functional tests,
demonstrating its capability to represent spatial and kinematic constraints,
collision avoidance, and yielding behaviours. Techniques and software presented
in this thesis could amount to more efficient evacuations involving PRM. This
increases their safety as well as the safety of those evacuating around them,
potentially preventing injuries and saving lives. This advancement holds
particular significance in scenarios such as hospital evacuations, where
accurate simulation of numerous device interactions is paramount for safety
optimisation.