This course focuses on the role of visual sensing in robotic manipulation. It covers fundamental manipulation concepts such as mathematical grasp formulations, grasp taxonomies, and grasp stability metrics. Various grasp planning strategies in the literature are studied. 2D and 3D vision-based control algorithms are covered. Point cloud processing techniques that allow object detection, segmentation, and feature extraction are studied and implemented. Students will integrate all of these aspects to design the whole vision-based robotic manipulation pipeline.
Students cannot receive credit for both 450X and 4540.
Knowledge of robot kinematics, wrench spaces, and rigid body transformations as presented in RBE 3001. Familiarity with robotic simulation software as presented in RBE 3002.