Real-Time Step Detection Using The Integrated Sensors Of A Head-Mounted Display

Recent improvements in virtual reality technology and head-mounted displays have led to a number of novel and innovative applications in entertainment, education, science and healthcare. The primary goal in most of these applications is to give the user a sensation of being part of the virtual reality. This focus on presence or immersion requires to create a connection between the user and the virtual environment but also between the user and his virtual avatar. Synchronizing the body movement of the user and his avatar can help to improve the feeling of presence by enhancing the experience of agency and body ownership over the avatar. A typical example is the combination of a head-mounted display and a treadmill to create a realistic walking or running simulation in a virtual environment. In this scenario, the synchronization of leg movement improves the feeling of presence and allows to further enhance the virtual experience, for example by adjusting gait parameters of the avatar or triggering customized stepping sounds.This paper presents a robust real-time step detector that uses the integrated sensors of a state-of-the-art head-mounted display and allows to recognize the pattern of individual steps. No additional sensors on the trunk or lower body are required. By applying a coordinate transformation and straightforward signal processing, it is possible to discriminate between left and right steps and detect the current walking speed. This information is used to animate a virtual avatar by scaling a predefined walking trajectory and to control the walking speed within a virtual environment.