Locomotion by Natural Gestures for Immersive Virtual Environments

In this paper we evaluate methods to move 'naturally' in an Immersive Virtual Environment (IVE) visualised through an Head Mounted Display (HMD). Natural interaction is provided through gesture recognition on depth sensors' data. Gestural input solutions in the literature to provide locomotion are discussed. Two new methods for locomotion are proposed, implemented in a framework used for comparative evaluation. Perceived naturalness and effectiveness of locomotion methods are assessed through qualitative and quantitative measures. Extensive tests are conducted on the locomotion considering also: 1) obstacles in navigation; 2) interaction with virtual objects during locomotion. This is done with the aim to identify methods capable to provide a full body experience in an IVE. Results show that one of the methods for locomotion we propose has a performance comparable to established techniques in literature. Outcomes may be exploited to improve the naturalness of users' movements in IVEs and help to unlock new strategies in providing IVEs for learning, training, collaboration and entertainment, also with respect to users with disabilities.