TELE-IMMERSIVE VIRTUAL ENVIRONMENTS FOR COLLABORATIVE KNOWLEDGE DISCOVERY

This paper describes the design and implementation of two tele-immersive applications, CVD and Cave6D, both designed to support collaborative knowledge discovery from large multidimensional datasets. CVD integrates the capabilities of two existing VR applications, Cave5D and Virtual Director, in order to provide immersive experiences of distributed data using high performance networks and interactive hardware and software. Cave6D is similar in function yet is tightly integrated with the CAVERNSoft toolkit so as to provide access to remote computational platforms and databases. INTRODUCTION In recent years there has been a rapid increase in the capability of environmental observation and modeling systems to provide high resolution spatial and temporal data about the world around us. It is not enough, however, to be able to collect, generate or share large amounts of data. Scientists, educators, students and managers must have the ability to collaboratively view, analyze and interact with the data in a way that is understandable, repeatable and intuitive. Concurrent advances in visualization and computational capabilities have spurred the creation and use of realistic, three-dimensional virtual environments for interaction with and analysis of this data. Viewing and interacting with these large multivariate datasets in a georeferenced virtual environment provides an actual sense of presence that inherently changes the way the data is analyzed (Wheless et al., 1995) thereby aiding in the mental process of assimilating complex information. Users are able to navigate through, view and interact with the data in a fully three-dimensional context, thus preserving necessary geospatial relationships crucial for intuitive analysis. High performance networks, such as the NSFfunded very High Bandwidth Network Service (vBNS), now provide an infrastructure for low latency multi-user access to these virtual environments (DeFanti et al., 1996). The use of these Collaborative Virtual Environments (CVEs) enable users at many distributed sites to interact with each other and with the data in a many-to-many session from within a common virtual world. Linking remote users into a shared virtual space enables a high level mode of collaboration that is conducive to knowledge discovery (NSTC, 1998). We describe in this paper our efforts to integrate this collaborative capability into an existing VR-based scientific visualization application, Cave5D (Wheless et al., 1996). Our objective is to enable multi-user immersive visualization of large multi-dimensional datasets in support of oceanographic, atmospheric and terrestrial scientific studies. Two separate implementation paths were followed, each resulting in a distinct prototype. In one case, Cave5D was augmented with remote interaction techniques and camera choreography capabilities provided by the VR application Virtual Director. In the other case, Cave5D was retrofitted with collaborative features provided by the CAVERNSoft toolkit (Leigh et al., 1997), a software library that is designed to allow for easy development of similar applications.