Advanced Robot-System Programming

Robot-System Programming, whether performed in teach mode or off-line, is characterized by the need to specify geometric data explicitly and the need to tie up production equipment for significant periods of time for training and debugging. In teach mode, it becomes tedious or difficult for repetitive or complex motions, and nearly impossible if sensor inputs such as vision or force are involved. In off-line programming, the explicit entry of geometric data becomes a severe problem unless coupled to a teach mode. This paper proposes the Advanced Robot System Programming. It is divided into three parts 1) automatically determining gripping positions of objects from their, CAD model. 2) extracting the training information for certain vision systems from the CAD models of the parts to be recognized. 3) developing a graphics-based programming system that can be used on widely available inexpensive hardware.