9412637 Kovacevic The principal investigator will work with representatives of industry on the development of the prototype of an adaptive weld penetration controller based on vision sensing. The research is a novel approach for controlling weld bead penetration, specifically for the thin plates is proposed. The methodology is distinguished from past approaches by the use of (1) an improved method for sensing the weld pool and its surrounding area using an ultra-high shutter speed camera assisted with pulsating laser as an illumination source; (2) a more effective algorithm for image processing (the extraction of the weld pool boundary is done in less that 100 milliseconds); (3) a model-based relation of the weld pool geometry, joint penetration, and welding process parameters; and (4) a newly developed robust adaptive controller. The control system will overcome the influence of joint geometry, mismatching, and thickness variation on the weld quality of nickel-based and titanium alloys with thicknesses ranging from one millimeter to two millimeters.