Department of Computer Science
CB #3175, Sitterson Hall
University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3175

(919) 962-1884
smithfd@cs.unc.edu
https://www.cs.unc.edu/~smithfd

Distributed Virtual Laboratory Environments

My research is concerned with network and operating system support for distributed virtual environments (VEs) in general and with the distributed nanoManipulator system in particular. The nanoManipulator is a virtual reality interface to a scanned probe microscope (SPM) constructed by UNC-CH computer scientists. It allows chemists, biologists, and physicists to "see" the surface of a material at atomic (nanometer) scale and "feel" the properties of the surface through the use of a haptic, force-feedback device. This is accomplished by integrating an SPM with high-performance 3D graphics and a tracking/force-feedback haptic device. This interface has enabled new scientific investigations that were otherwise not possible to be conceived and performed.

The current system structure couples a graphics supercomputer (a high-end SGI), an Intel-processor based microscope/controller (a TopoMetrix, Inc. Explorer) and an Intel-processor based hand-tracking/force-feedback controller (a SensAble Devices, Inc. Phantom) via a network (a dedicated switched Ethernet). We are now exploring designs that will allow interconnection of these components over the Internet to create a virtual microscopy laboratory. Highly interactive virtual environments that provide true telepresence require low-latency, real-time communication between system components and real-time computation within components. Realizing these requirements on commodity computing engines, and commonly deployed packet-switched networks such as the Internet, is the essence of this research. Unfortunately, current networks are subject to highly variable latency, available bandwidth, and rates of packet loss, all of which present significant research issues in distributed immersive environments.

Only one of the Intel processors in this system currently runs Windows NT (the Phantom controller). We are now considering design options that would allow us to migrate the entire system to NT and use that as our base for the distributed version. The key issues are (a) high performance for the 3D graphics, (b) high-speed network support, and (c) real-time support.