Yu-Hung Hsu


The machine vision system (MVS) has been heavily publicized because of its potentially crucial contribution to robotics. It is essential to alignment and inspection tasks in integrated circuit components' assembly equipment. The objective of this thesis was to examine the capability of a MVS to assist in the automated robot assembly of the dual-in line device (DIL) and surface-mounted-device (SMD). The DIL, with 8x2-pin, and the SMD, with 24x 16-pin, were used as models, while an Adept SCARA robot was used as the robot system. A magnet-type end effector, with a force sensor, was also designed to be incorporated with the vision system to measure applied force during assembly. A vision accuracy of 0.12mm/pixel and 0.12 was obtained, and an overall reliability of 91% for chip recognition and inspection was achieved. A complete assembly cycle time of 75 seconds, operating at a low robot speed of 50mm/sec (2in/sec), was recorded. The end effector detected assembly force with an average of 5. 1% tolerance from the preset value. Polynomial fitting functions, calculated by using the Least Squares method, were employed to interpolate the experimental calibration data and to improve the force sensor accuracy of the end effector. Descriptions of the Adept robot system, vision system, and force sensing end effector are included.

Library of Congress Subject Headings

Integrated circuits--Design and construction; Computer vision--Industrial applications; Robots, Industrial; Process control

Publication Date


Document Type


Department, Program, or Center

Mechanical Engineering (KGCOE)


Walter, Wayne

Advisor/Committee Member

Kempski, Mark

Advisor/Committee Member

Hennessey, Michael


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TK7874 .H799 1996


RIT – Main Campus