Humans interact with the built environment through multisensory modalities, presenting a great challenge to evaluate the performance of building structures for human comfort. This paper aims to develop a testbed facility for the study of human interaction with visual, acoustic and vibrational stimuli, by in-situ modelling of a lively footbridge in virtual reality (VR). The footbridge, known as the Warwick Footbridge (WF), is a 19.8 m long, 2.35 m wide, and 475 mm deep all fibre-reinforced polymer shallow truss structure, with a total mass of only 1,400 kg. Its four movable supports create a simply supported condition, allowing for the change of span for variable natural frequencies and vibration environment. To provide realistic visual and acoustic environment for human sensory interaction analysis, a VR simulator of the WF has been created and spatially anchored to the physical bridge, leading to a testbed that integrates real footbridge vibrations with controllable audiovisual stimuli in VR. The WF is sensitive to human excitation that can produce visible movement to humans. Vibration displacement of the physical WF is therefore collected to create visual displays of the WF movement in real time. In addition, the tracking system of the VR headset is evaluated to ensure the pedestrian’s position on the virtual WF aligns with their location on the physical WF. The tracking error, averaged over 30 test points distributed across the physical WF, is 2.4 ± 2.3 cm (mean ± standard deviation). Usability of the VR simulator is then successfully verified by human walking tests. Consequently, multisensory human-structure interactions can be evaluated under control over the visual, acoustic, and vibration stimuli. Understanding of these multisensory interactions will contribute to the construction of building structures that better meet human needs and comfort.