Pedestrian-pedestrian interaction (PPI) is one of the fundamental mechanisms
purported to influence the amplitudes of structural response under the action of a walking
crowd. This is because a pedestrian is likely to alter their gait due to the presence of other pedestrians, which in turns alters the magnitude of structural loading. However, little empirical
data are currently available to assess the effect of PPI in the context of vibration serviceability. This is mainly due to logistical challenges in assembling and instrumenting a crowd of
walking pedestrians, and the associated cost. To this end, a novel virtual reality platform is
developed for experimental investigation of pedestrian-pedestrian interaction. In comparison
to real-world crowd testing, the platform enables experimental protocols to be implemented
repeatedly in a highly controlled environment while collecting a rich set of data on pedestrian
behaviour. The platform incorporates state-of-the-art technology for motion capture, artificial intelligence and three-dimensional computer modelling, and comprises of three core
modules: (i) the environment, (ii) the crowd and (iii) the user interface enabling real walking
behaviour. To assess the validity of the platform for investigating PPI, tests were conducted to
quantify gait synchronisation between a pair of walking pedestrians. The pair of pedestrians
consisted of either two real humans or a real human and an avatar generated within a fully
immersive VR environment. The test subject was either not explicitly asked to or specifically
asked to synchronise their gait while walking side-by-side or front-to-back. It was found that
walking with an avatar yields qualitatively the same results as walking with a real person,
whether that is with or without the instruction to synchronise gait. However, the results differ
quantitatively in terms of the synchronisation strength and the directionality.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)