Search

A Robotic Command Center, is to be developed to allow a commander and two drivers (three people) to control four remote vehicles. The drivers will be attempting to control two vehicles each--a difficult task. Giving vehicles a measure of autonomy is one way to simplify that task. On the other hand, deployed vehicles must not be hampered by software with limited competence or reliability. The only solution available soon is a vehicle control interface that allows for the entire range of possibilities between autonomous behavior and continuous teleoperation. Ideas are fused from telerobotics and from the Reaction Planning subfield of Artificial Intelligence, to develop a technology that permits vehicles to react autonomously to expected circumstances, while also permitting the human operator to instruct vehicles about what to do and how to do it. This study will program some examples of behaviors for the vehicles to perform semi-autonomously, and those behaviors by running them in the Team Works I environment. An interface was designed and built that allows the vehicle driver to communicate with the intelligent software controlling any and all of the vehicles. A single person is capable of controlling four simulated vehicles at the same time. The competence of the vehicle is to be expanded to give the operator new ways of instructing the vehicles, and to push those capabilities toward deployment in the real world.

The goal of this effort is to develop and demonstrate prototype processing capabilities for a knowledge-based system to automatically extract and analyze features from Synthetic Aperture Radar (SAR) imagery. This effort constitutes Phase II funding through the Defense Small Business Innovative Research (SBIR) Program. Previous work examined the feasibility of and technology issues involved in the development of an automated linear feature extraction system. This final report documents this examination and the technologies involved in automating this image understanding task. In particular, it reports on a major software delivery containing an image processing algorithmic base, a perceptual structures manipulation package, a preliminary hypothesis management framework and an enhanced user interface., Original contains color plates: All DTIC and NTIS reproductions will be in black and white.

The goal of this effort is to develop and demonstrate prototype processing capabilities for a knowledge based system to automatically extract and analyze linear features from synthetic aperture radar (SAR) imagery. This effort constitutes Phase II funding through the Defense Small Business Innovative Research (SBIR) Program. Previous work examined the feasibility of and technology issues involved in the development of an automated linear feature extraction system. The current effort continues this examination and is developing the technologies involved in automating this image understanding task. Keywords: Feature extraction, Edge detection, Terrain analysis, Image understanding.