The faculty members of the graphics and image groups share over 3,500 sq. ft. of laboratory space in Sitterson Hall. This lab is equipped with a large number of high-performance graphic workstations.
A large portion of the laboratory space is dedicated to the group’s work in effective virtual environments. Specialized VE equipment includes head-mounted displays from NVIS and Virtual Research that are used in a 32 ft. by 34 ft. tracked space. This wide-area tracking is done using the UNC-developed HiBall™ optical tracker (now a commercial product from the 3rdTech, Inc.). A PhaseSpace IMPULSE capture system is also used for motion capture research.
An 800 sq. ft. portion of the laboratory is dedicated to robotics research. The laboratory houses a variety of robots, including a Rethink Robotics Baxter robot with two 7 degree-of-freedom arms, an Aldebaran Nao small humanoid robot, a differential drive Pioneer P3-DX mobile robot from Mobile Robots, Inc., iRobot Roomba platforms for multi-robot coordination experiments, and prototypes of custom-designed robotic devices for minimally invasive surgery. The laboratory features an open space for mobile robot research and includes ceiling mounted cameras for robot tracking.
Two additional lab spaces are dedicated to multi-user, auto-stereo display research and advanced telepresence research. The latter currently includes a full-duplex video wall testbed built with eight 64″ HDTV flat-panel displays, 20+ Point Grey Research high-resolution digital cameras, a low-resolution PMD[vision] S3 3D-camera from PMD Technologies, and five multi-core Dell T7400 workstations. The team also enjoys dedicated access to a Cisco TelePresence System 1000 installed and funded by Cisco as part of their national research and education network (NREN).
The laboratory’s scene capture resources include two DeltaSphere™ rangefinders, one with a visible light laser and another infrared. The DeltaSphere has a range of up to 50 feet, with typical error of less than 0.1 inches at 30 feet. The absolute error depends on the surface scanned. Angular resolution is variable; we typically set it at 15-30 samples per degree.
The expanded graphics/vision lab in the department’s new building addition is approximately 2,500 sq. ft. that is configurable into multiple curtained zones, each with its own lighting, power, and A/C controls. All of the spaces have a Unistrut® grid below the ceiling for mounting equipment (projectors, cameras and lighting) with power and data connections.
This space is currently configured into three research zones – one for projective display research, one for computer vision and optical tracking research, and one for physical simulation and audio synthesis research with a focus on GPU processing.
We have an extensive inventory of high-quality digital cameras (e.g., over 100 from Point Grey Research and Sony, and 10 progressive-scan PTZ networked cameras, Canon EOS 5D Mark II, Canon EOS 60D) and projectors (high-end units including Projection Design models F1, F2, F20 and F22 and numerous LCD and DLP projectors from Proxima, InFocus, Mitsubishi, Epson and NEC) that we can use for this project. We also have 10 pan-tilt-units and controllers from Directed Perception, and full-duplex audio equipment. This lab space has three high-end human tracking and motion capture systems: a 3rd Tech HiBall-3100™ Wide-Area, High-Precision Tracker and 3D Digitizer, a Northern Digital Optotrak Certus® motion capture system, two NaturalPoint OptiTrack motion capture systems, and an Xsens Moven Suit motion capture system. We have enough lab space to stage both static and mobile platforms. To aid in research on avatar inhabitation by people we have both a Point Grey Research Ladybug2® and Ladybug3® spherical digital video camera system, and a custom built 7-projector surround workspace/console.
The Applied Engineering Laboratory (AEL) provides facilities and expertise for building prototypes of a variety of microtechnology-based systems. It enables research in the use of information processing technology in a multidisciplinary context. Custom-designed and/or off-the-shelf VLSI and conventional components are used to build systems from small-scale proof-of-concept prototypes through systems of significant size and complexity, such as the Pixel-Planes family of graphics supercomputers and the HiBall tracker for virtual environments. A technical staff of research faculty and state-funded engineers provides continuity of know-how and economically leverages new research thrusts. Sophisticated facilities and equipment are maintained to support work in electronic, optical, mechanical, and other relevant technologies.
The AEL is an exceptional academic prototyping laboratory with an established, world-class track record. The facility contains a complete machine shop (with milling machine, lathe, etc.), measurement and analysis equipment (oscilloscopes, spectrum analyzers, signal generators, etc.), chip test and rework gear, and chemical handling equipment.
The lab maintains an extensive suite of design automation (CAD) tools. These tools include industry-standard Cadence and Synopsys tool suites, as well as industrial-strength tools for the design of asynchronous circuits and systems (Haste/TiDE tools from Handshake Systems, a subsidiary of Philips Semiconductors). These tools are used for design and simulation, and for preparing VLSI layouts for sending to foundries for fabrication through the MOSIS fabrication service (www.mosis.com).
The department’s networking group also maintains a large lab for network research that currently contains over 150 PC systems running a wide range of production and experimental operating systems, interconnected via dynamically configurable and partitionable networks running at 10, 100, and 1,000 Mbps. The laboratory is also directly connected to the North Carolina Networking Initiative’s GigaPOP, a Cisco DPT fiber ring spanning the Research Triangle Park region of North Carolina, operating at speeds of up to OC-48 (2.4 Gbps), and interconnecting the Triangle with the National LambdaRail. This flexible facility, which has been used to investigate technology for realizing differentiated services based on combinations of active router-queue management and packet scheduling, will be used for virtual presence research.
Over 5,000 sq. ft. space is dedicated to the bioinformatics and computational biology research cluster, which includes office space for faculty, staff, and students, computer laboratories, classrooms, and conference rooms. The laboratory is also equipped with a newly installed Dell cluster server with over 200 CPUs and over 50 TB storage capacity. Ten computing nodes on the cluster each has 32GB memory while the rest of the nodes each has 8GB memory. The cluster runs Ubuntu linux and Sun’s Grid Engine. The faculty, staff, and students in this facility develop cutting edge computational methods for large scale data management, analysis, and visualization, which are critical for many biology research projects. This facility supports current and new research collaborations between computer scientists and biologists, as well as educational and training activities for interdisciplinary students. The laboratory will provide a dedicated environment to host and train students and postdoctoral researchers from outside the CS department, and provide new computational tools benefiting biology research.
A space dedicated to computer security faculty, students and associated postdocs and staff members exists on the third floor of the Frederick P. Brooks, Jr. Computer Science Building. This space accommodates three faculty members focused on computer security, and up to twenty graduate students. It also includes meeting rooms and offices to accommodate short- and long-term visitors to the group in support of collaborations with other institutions. The Security Laboratory leverages the networking and computing facilities in our department, but also has a number of specialized computing clusters dedicated solely to the group. Many of the machines in these clusters have strict policies on how sensitive data is collected and stored (e.g., anonymized and encrypted while resident on disk), as well as restriction on what principals can access such data. Critical data is automatically backed up for archival purposes.