Visually Guided Robotic Hand / Eye Coordination

The proposed work leverages the prototypes developed in the Phase I project toward refinement of a machine vision system for visually guided manipulation. In Phase II, the machine vision system modules of image feature detection, object detection, pose estimation, and tracking will be generalized to handle a wide range of deformable objects. A new four fingered hydraulic end effector will be fabricated to physically test the effectiveness of a parameterized grasping approach. The Hybrid Mode Estimator, also designed in Phase I, will be implemented to allow for dynamic path planning with respect to a constantly changing environment. to key components of the machine vision system required for visual guided manipulation. The machine vision system is comprised of the following modules: image feature detection, object detection and recognition, pose estimation, and tracking. The focus of the Phase II effort is to optimize and integrate these components into a robust and generalizable system.The machine vision system is comprised of the following modules: image feature detection, object detection and recognition, pose estimation, and tracking.