A wearable gesture recognition system for natural navigation interfaces

Ambient intelligence applications pose the challenge of enabling real-time interaction with virtual environments in a natural manner. Gesture recognition based on body-mounted accelerometers has been proposed as a viable solution to translate patterns of movements that are associated with user commands, thus substituting point-and-click methods or other cumbersome input devices. These wearable systems are a promising approach to gesture recognition because of their low cost and size, high robustness and minimal constraints on user location and freedom of movements. In this work we present a gesture interface system based on low-cost integrated accelerometers for navigation in virtual spaces. Qualitative and quantitative assessments are presented using a 3D game application as a test-bed to evaluate the effectiveness of the interface