Enhancing telepresence with mobile virtual proxies

Abstract Traditional telepresence systems are comprised of a person remotely controlling a robot in a hostile environment while receiving visual feedback from a camera mounted on the robot. While useful for a number of applications, this model is not particularly useful for everyday work applications. The size of the robot is intrusive, the robot needs to be designed for specific interactions and only one person can use the robot at any one time. This work seeks to address this problem by replacing the physical proxy with a virtual proxy. The purpose of this virtual proxy is to enhance the system by providing improved support for multiple users, interaction and navigation in the remote environment. This enhanced, or improved, version of telepresence is termed TeleReality as it combines elements of virtual reality with traditional telepresence technologies. To achieve this goal, the basic building blocks, or constructs, of traditional telepresence systems need to be changed. This thesis identifies and evaluates the base constructs needed to build any TeleReality system. These constructs include the need to support navigation of the remote environment and this is achieved by using a network of cameras and image processing software to calculate the various perspective viewpoints of the users. These constructs govern the means in which this collection of cameras are organised and connected. Each user receives a common set of video images from which they calculate their own perspective viewpoints, and consequently supporting a multi-user system. Interaction within the remote environment is promoted using ad-hoc networks and augmented reality technologies. Constructs also cover security and privacy issues that arise from using multiple cameras by adopting both an organisational and technological viewpoint. The focus is on establishing trust within the system by divesting control to the user. An example of these constructs is given by the implementation of a TeleReality model called a 'Visual Cell' system. The conclusion of this work identifies the constructs that are needed to support a telepresence system using a virtual proxy where the primary interaction framework is informational exchange, although physical interaction can also be supported depending upon the environmental support. This work also identifies the technical issues that require additional research for the implementation of a TeleReality system, from the need for improved image processing, video codec's, broadcast and ad-hoc security protocols, software architecture, registration and the availability of suitable head mounted displays