Real Walking In Large Scale Scenes For Virtual Reality

Ivan Sutherland,the father of computer graphics,regards virtual reality as an ulti-mate display mode.Virtual reality provides users with multisensory information sim-ulation and stimuli such as visual,auditory,and haptic,so that they can be completely immersed in the virtual environment generated by computer,perceive and act as they are still in reality.For virtual reality technology,an important and popular research is to provide users with a realistic experience in virtual environment and allow users to interact with the virtual environment naturally.Moving from one place to another is a common activity in daily life,so it is also one of the most common and universal form of interaction in virtual reality.Roaming the virtual environment by real walking in real workspace can provide users with highly immersive presence and a good user experi-ence.However,enabling users to do so is not trivial.There are two major challenges of providing users with natural walking experiences in VR applications:(1)How to enable users to walk freely in virtual world that are larger than the real workspace;(2)How to avoid collisions between users to ensure users safety when multiple users roam in the same real workspace.Aiming at these two major challenges in real walking in virtual reality,this paper proposes two mapping based methods based for real walking and an automatic collision avoidance method that supports simultaneous real walking and physical interaction for multiple users.Mapping based techniques are very effective for supporting real walking in small physical workspaces while exploring large virtual scene.However,the existing meth-ods for computing real walking maps will distort virtual scene severely when the virtual scene is much larger than the real workspace,thus affects users' walking experience and visual experience.In this dissertation,we present a novel divide-and-conquer method,called Smooth Assembly Mapping(SAM),to compute real walking maps with low isometric distortion for large-scale virtual scenes.First,the input virtual scene is de-composed into a set of smaller local patches.Then,a group of local patches is mapped together into a real workspace by minimizing a low isometric distortion energy with smoothness constraints between the adjacent patches.All local patches are mapped and assembled one by one to obtain a complete map.Finally,a global optimization is adopted to further reduce the distortion throughout the entire map.A large number of experiments,including formative user studies and comparisons,have shown that our method succeeds in generating high-quality real walking maps from large-scale virtual scenes to small real workspaces and is demonstrably superior to state-of-the-art meth-ods.Although the isometric distortion of Smooth Assembly Mapping can be at a rela-tively low level in value,the road in the virtual scene becomes very curved after being mapped,which affects the user's visual experience while roaming the virtual scene.In this dissertation,we present another scene mapping method called Redirected Smooth Mapping.The key idea of this method is to combine the redirected walking technique and the smooth mapping method,take advantages of both.First,the road in the virtual scene is mapped into real workspace using Smooth Assembly Mapping as a guide path for users.Then,we optimize the Smooth Assembly Mapping by using the threshold of redirected walking operation as a constraint,and obtain a mapping with lower dis-tortion.Users possess a wide field of view to explore the mapped virtual environment which has low distortion while being redirected in the real workspace.Compared with state-of-the-art methods,Redirected Smooth Mapping can provide a better roaming ex-perience.In multi-user VR applications,if users are in the same real workspace,they are at risk of colliding with others.In this dissertation,we present a novel automatic colli-sion avoidance technique to provide multi-user real walking experiences with physical interactions in VR applications.In our system,multiple users walk freely while navi-gating a large,virtual environment within a smaller,physical workspace.These users can interact with other users or physical props in the workspace.In order to enable the interaction between users,we need to utilize the mapping based method to support users' real walking,which makes the roads in the virtual scene overlap and makes users at the risk of collision while roaming.To avoid the risk of collision,we present an auto-matic collision avoidance technique based on dynamic avatars.These avatars naturally appear,move,and disappear,producing as little influence as possible on users' walk-ing experiences.We demonstrate the capability and practicability of our multi-user real walking system through formative user studies.