CAVE: CAVE Automatic Virtual Environment
This study attempts to help define optimal operating and training procedures for conducting SNI routes with GPS by studying pilot's scan patterns and navigation performance based on data collected in flight. Further, this study attempted to compare pilot scan patterns and navigation performance in real and simulated environments to determine if simulation is a viable framework for future study of helicopter pilot's use of GPS on SNI routes.
This experiment was divided into two phases. The initial phase involved constructing a route, refining data collection equipment and procedures and collecting in-flight data. This phase involved flight data collection in Tullahoma Tennessee. The second phase of the experiment involved creating a virtual replica of the actual aircraft and test environment. "Virtual Tullahoma" was created at the Naval Postgraduate School in Monterey California.
The image display hardware system was a CAVE (CAVE Automatic Virtual Environment) consisting of three rear-projected screen and projector support structures arranged in a "U" shape. A mock cockpit is positioned on a raised platform facing the center screen. This results in 180 degree viewing angles for the virtual pilot.
- Built as part of a study to determine effect of GPS navigation by helos can reduce separation in high density airspaces.
- Put a simulated cockpit inside a 3 walled CAVE driven by Delta3D which provided the virtual pilot with 180 degree views.
- Virtual pilots flew the same route actual pilots flew over Tullahoma, Tennessee at the MOVES Institute at the Naval Postgraduate School in Monterey. Navigation performance was compared.
- Communicates to the Image Generators (IG) using the Common Image Generation Interface (CIGI).
- System consisted of five computers (one for flight characteristics, three to drive the three screens, and a host to coordinate everything and drive gauges) networked together using Delta3D.
- Terrain was generated from elevation data, satellite imagery and cultural data combined to produce an OpenFlight model, with approximately 64 polygons per square kilometer on the low level of detail (LOD), and 256 polygons on the high.
- Qualitative analysis of virtual Tullahoma track information suggests that navigation performance in the virtual configuration approximates navigation performance in flight. This indicates that results from virtual experiments can be used in lieu of actual flights.