Imaging through low-visibility fire smoke for ship-board navigation of robotic fire-fighting systems

The goals of this Phase II project are to develop a prototype vision system and demonstrate its capability of locating boundaries and obstructions, normally 15 feet away, within one second in scenarios containing smoke from a real fire. This vision system needs to be compact and lightweight. ASC will fabricate and modify a new TigerCub camera and take advantages of its weather resistant ability as well as the replaceable external laser design to accomplish the task. To evaluate the system performance quantitatively and record valuable information for further improvement, there is need to carry out tests in a well-controlled fire lab. ASC will work with Dr. Brian Lattimer at Virginia Tech to control smoke densities, flame size, and water suppression effects during data acquisition. ASC will also develop a calibration procedure to quantify the accuracy because profiling uncertainty levels will be essential for later developments in high-level mapping and path planning algorithms of the robotic firefighting system. Thermal tests for equipment survival in real fire environment are more suitable to be carried out in the subsequent projects after the system accuracy has been verified and achieves the expected goal. The deliverables of this project include a prototype of the testing unit, tailored control/display software installed on a selected laptop, operational manual, status reports and the final report.