Research And Implementation Of High Immersion Experience On Smartphone HMD

The VR(Virtual Reality), as a more natural approach to human-computer interaction, has been greatly developed in the past a few years. Currently, the HMD(Head-mounted Display) is the most popular VR solution. At the same time, an entry-level VR solution, the smart-phone-based HMD, has already gained ground in recent years. The core concept of the smart-phone-based HMD is that to use the imbedded gyroscope of a smart phone to track the movement of user's head, and then present the real-time results in the virtual scene as view angle changes. Plus 3D display technology, users would therefore receive a good sense of immersion. As it is easy to be popularized and the nice experience it provides, the smart-phone-based HMD has shown a bright prospect in entertainment, education and tourism, etc.The professional HMD solutions often apply depth camera, infra-red location system and movement capture technology to track the movement of other parts of user's body, thereby achieving a much more complicated interaction. In the contrary, limited by hardware, the smart-phone-based HMD can only complete simple interactions based on the movement of user's head and individual real buttons, which could merely give us a limited sense of immersion. As a result, the VR applications of the smart-phone-based HMD inevitably lack variety in terms of content and the corresponding experience is relatively poor.In response to above problems, this paper will analyze the source of immersion feeling in VR interactions as well as the ways of improvement. According to the result, this paper will also design a smart-phone-based HMD solution, which could achieve a deeper sense of immersion with increased ways of interaction. In addition to the conventional head movement tracking, this solution adds voice recognition to the system and innovatively introduces an algorithm of movement recognition based on acceleration information, which enables users to control the state of movement in the virtual reality through changing their state of movement in the real world. In order to adapt to such low-delay requirement in VR applications, this paper improves the conventional algorithm of movement recognition based on acceleration information, and it gives us a real-time state of movement recognition algorithm applying threshold method, which could increase the speed of recognition with an acceptable accuracy. This new algorithm could recognize user's state of standing, walking, running and jumping, effectively.In the end, a VR game demo to which applies above solutions has be designed and made in the aim of testing the practical results of the solutions.