ME 234: Advanced Robotics
Winter 2022
Winter 2022
As a final project for ME 234a, I 3D printed and assembled a fractal vise. This design allows the vise to grip irregularly shaped objects, which is an interesting problem in the field of grasping.
The fractal gripper design has full immobilizing closure for most objects for the frictioned finger and soft finger models as long as the fingers can squeeze (either enough friction or fingers ‘enclose’ the object’. Furthermore, the grasping is quite gentle as it has as many as 8 contact points for each side of the gripper, each of which applies the same force. This disperses the force across many more points than traditional vises (except for objects which make use of parallel grips). This seems like a fairly good ‘universal’ gripper design as objects can be immobilized with a single DOF. If this gripper is attached at the end of a 6/7DOF arm, it would be able to apply wrenches in the same space as the arm’s motions allow since, relative to the gripper, the object should stay immobilized under closure.
The 3D printed pieces rotate nicely at the ends of the gripper. However, the friction gets worse as the pieces get bigger, so it ends up taking quite a bit of force sometimes for the large pieces to conform correctly. Something interesting about the gripper is that because of its symmetrical fractal design, the forces at each of its contact points will be the same. Therefore, the force vector is really just a force scalar if we give each contact point a column in the grasp map.
Since this gripper only has one DOF (the lead screw opening or closing the jaws), it doesn’t have enough actuation to have grasp closure, but in some cases, it may have ‘holding’ closure, where the gripped object cannot be moved while gripped.
Course description: The first quarter focuses on advanced robot kinematics and mechanisms. Topics include a Lie Algebraic viewpoint on kinematics and robot dynamics, a review of robotic mechanisms, and topics in robotic grasping and manipulation. The second quarter focuses on advanced topics in robotic motion planning and navigation, including planning with POMDPs, robotic coverage planning, and multi-robot coordination. Course work will consist of homeworks, programming projects, and labs.