In an era of rapid technological advancement, robots have become increasingly commonplace in settings from factories to households. Notably, robotic manipulators designed for pick and place applications increase efficiency and productivity. This report presents our team’s development and performance of an autonomous 5 DOF robotic manipulator that senses, manipulates, and stacks various sized blocks. This project explores computer vision to detect, classify, and differentiate block and non-block like objects, as well as recording their sizes, colors, and orientations. In addition, it documents our methodology of camera calibration that maps pixels to world coordinates utilizing intrinsic and extrinsic matrices, as well as camera homography, utilizing OpenCV packages. Finally, it explores robotic kinematics (both forward and inverse) to plan joint motions across various end effector waypoints.

These techniques are combined and validated in a competitive setting to test handling of blocks of various colors, sizes, and shapes through sensing, manipulating, and stacking exercises