Virtual reality is an attractive alternative to conventional settings in which human locomotion studies are conducted. It enables test subjects to be exposed to a wide variety of virtual environments which can be easily manipulated to suit a given experimental protocol. Therefore, it could be of great benefit to research on crowd dynamics in which current experimental approaches are rather limited in their scope. These approaches typically rely on top-view cameras recording video footage from which kinematic data on pedestrians’ behaviour are inferred. The main advantage of this technology is that it enables data on human gross motion in natural environments to be collected. However, this approach alone cannot provide information on pedestrians’ stepping behaviour, which is of great importance in research ranging from cognitive psychology, human movement science and, last but not least, structural stability. Moreover, repeated test trials can only be used in limited circumstances since the reproducibility of experimental conditions can only be achieved in a laboratory environment. Consequently, a relatively high volume of data needs to be collected in real-world settings to achieve a satisfactory level of data reliability.
To address this issue, as a part of a PhD project conducted by Artur Soczawa-Stronczyk, we set out to develop an experiment enabling pedestrian’s stepping behaviour to be recorded while walking overground with a virtual reality avatar.
To realise this ambition, we turned a sports hall at the University of Leicester into a warehouse-scale VR studio instrumented with a state-of-the-art optical motion capture system supplied by our friends at Target3D. A custom-built VR platform was developed by Artur for the purpose of this and other studies, including AI-driven humanoid avatar motion simulator, capable of generating auditory stepping cues, and a virtual environment closely resembling the physical space in which the experiment took place. The strength and directionality of movement coordination between a test subject and a pacer was quantified without and with the instruction to synchronise gait. The pacer was either a real human subject or their anatomically- and biomechanically-representative virtual reality avatar.
The paper reporting the results from our experimental campaign is freely available here:
Soczawa-Stronczyk, A.A. & Bocian, M., Gait coordination in overground walking with a virtual reality avatar, Royal Society Open Science 7, 2020, 200622.
Supplementary materials, including an interactive simulator showing the virtual environment used in our tests, are available here:
Soczawa-Stronczyk, A.A. & Bocian, M., Data from: Gait coordination in overground walking with a virtual reality avatar , v7, Dryad, Dataset.
The ongoing research seeks to understand interpersonal movement coordination in walking and its sensory determinants with the applications in structural engineering and rehabilitation.
You may also be interested in our recently published paper on gait coordination in overground walking groups, which is available here:
Soczawa-Stronczyk, A.A., Bocian, M., Wdowicka, H. & Malin, J., Topological assessment of gait synchronisation in overground walking groups, Human Movement Science 66, 2019, p.541-553.