The Sensor Based Manipulation of Irregularly Shaped Objects with Special Application to the Semiconductor Industry

A Robot Assisted Crucible charging System (RACS) has been proposed in the automation of the crucible packing process in the CZ semiconductor wafer production procedure, involving the delicate manipulation and placement of polycrystalline silicon nuggets, into a stiff and fragile fused silica crucible. Two of the key elements in this design have been described in this thesis, a non-intrusive nugget and crucible surface geometry acquisition module, and a nugget placement and packing algorithm. A non-contact 3-D surface geometry measuring system has been developed based on the principal of optoelectronic active laser triangulation after a thorough investigation of a range of non-intrusive range sensing methodologies. This system measures both the nugget geometry profile and the internal crucible geometry profile, with a resolution of 1mm. Additionally, a novel on-line, multidimensional, flexible packing algorithm has been developed based on a principal of "Virtual Trial and Error" and extensively tested by simulation for cost function optimization. Four general cost function principals are tested in six critical combinations, in a two dimensional simulation of the packing algorithm using random non-convex polygons. The final cost function choice of lowest fit is shown to have the best performance index. For general comparison, the algorithm with all six packing schemes, has been applied to random sized rectangular object packing. This final packing algorithm scheme has been applied to a simulation of the 3D case, that compares well with the expected performance of human packing. These key technology components and requirements have been successfully demonstrated in order to provide for a feasible solution for a Robot Assisted Crucible charging System. Thesis Supervisor: Dr. Steven Dubowsky Professor of Mechanical Engineering Acknowledgements I would like to thank my academic and research advisor, Prof. S. Dubowsky, for guiding me over the course of the past two years, for pushing me to reach farther, for teaching me what it takes, for believing in and trusting me, and for giving me my freedom to explore the fascinating world of machine vision. It is through his tutelage and his friendship that I have learnt how to be strong in the face of adversity and how to see the big picture. His drive and love for technology has helped us transform a concept into reality and see the fruits of our efforts come to life. Next, I would like to thank Shit-Etsu Handotai, Co. Ltd. of Japan, for giving us the opportunity and resources necessary in order to carry out the research presented here, and help explore and refine a new window of technology application. I would also like to thank my project colleagues Joe Calzaretta, Tony Leier and Melissa Tata, for providing the support and friendship during both the tough and facile times over the past two years. Many thanks to Mike Mulqueen for being the big brother that I never quite had, who understood and cared through times of both sadness and joy, who was always there when I needed him and for whom I will always be there. It is Prof. B. Andereck and Prof. L.T. Dillman, of Ohio Wesleyan University, Prof. Erik K. Antonsson and Prof. J Burdick, of Caltech, that helped guide and set me on the track of academic research. It is through their direction, patience and support that I have learned to believe in myself and in realizing the dreams that otherwise would have been unachievable. With their friendship and brilliance that, over the course of the past several years, I have learnt how never to give up and to continue on to the next level in search for excellence. And yet none of this could have ever been done without my kind yet strong mother, and my firm but loving father. It is through their love and optimism for life that I have learnt how never to feel weak, how never to feel down, how never to be boastful, and how never to accept defeat. It is they who have taught me love, patience, courage, discipline and honor. It is they who have taught me that the challenges of life are what makes it all worthwhile. It is they who have taught me how to be a man and that there is always another round. Finally, I would like to thank the Almighty, for blessing me with the opportunities to explore this wonderful world of ours. This thesis is dedicated to the memories of my grandmother Sati H. Sujan.