Vision assisted SCARA (selective compliance assembly robot arm)

• Main Authors: Lim, Phyllis, Lim, Sam Benedict, Luna, Edmund, Sotto, Jinky, Syling, Christopher A.
• Format: text
• Language: English
• Published: Animo Repository 1997
• Subjects: Visual programming (computer science); Eye, visual pathways; Robotics; Information display systems industry; Computer-aided software engineering; Assembling (Electronic computers)
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/8632
• Institution: De La Salle University

The study presents a Vision Assisted SCARA. The vision system uses a back propagation neural network to identify and locate seven geometric objects. This information is then used to direct the SCARA robot to pick the objects and place the objects in pre-assigned bins. The neural net was trained to recognize with invariance to location and orientation. Recognition accuracy is 100 percent and the point accuracy of the robot is also 100 percent within a 1 mm tolerance.