ALS Ice Bucket Challenge
Remember the ALS Ice Bucket Challenge? It happened right around back to school time, so we decided to go big, the only way to do it, and have the dean of the engineering school do the Challenge publicly at the Lyle Ice Cream Bash back-to-school event for engineers. We built a box that held 40 cubic feet or about 2500 pounds of cold water, and dropped it from 20 feet up on the Director of the Innovation Gym, the Dean of the Lyle School of Engineering, and the Assistant Dean of Recruitment at the end of the Ice Cream Bash. A spectacular way to start the year!
In collaboration with the US Marine Corps Warfighting Lab through its Commercial Hunter Program, SMU engineers worked on a non-classified project with the intent of finding ways to use commercial off-the-shelf (COTS) technologies in systems that could potentially be used to harm Marines and other soldiers. We focused on specifically unmanned systems, which include aerial, land, and sea-based drones. The use of drones to keep soldiers out of harms way and to allow us to go places humans can't has become popular in warfare in recent years. However, we are not the only ones researching this technology and using it to our advantage. Insurgents and our adversaries have begun to look into low-cost drones that can be used to either counter our drone technology or directly attack us. A growing hobbyist movement behind flying advanced consumer quadcopters and fixed-wing aircraft means more and more unregulated parts become available for purchase. We researched methods that adversaries could most effectively use to do the most damage, and then created an aerial object recognition platform that could be used to identify suspicious aircraft. The idea behind it was that enemies wouldn't be flying the large 50-foot wingspan drones that the US military uses, but they could easily fly a $200 hobby plane with a 5-foot wingspan packed with explosives right into a convoy of American soldiers. They are cheap and dispensable, so we decided to look for small to medium-sized hobby aircraft in our detection algorithms because these would most likely go unidentified by recognition systems the military uses to find larger foreign aircraft. Our system uses an HD CCTV camera mounted on a pan-tilt system. It was trained using a series of reference images with test aircraft and backgrounds in an algorithm written in OpenCV. We were able to use the system to find aircraft flying around in the general proximity and track it when identified. Had we upgraded to a telescopic lens camera, we could definitely increase the detection range of the system. My role in the project was primarily research and documentation. Although I did learn an immense amount about the state of drone technologies and the how computer vision systems work, I did not have the prerequisite knowledge to grasp the advanced computer science concepts used to create the operating system behind drone tracking system.