To make this a closed-loop mechanism, we needed to incorporate closed linkages into our gripper jaws, which would likely be mirrored in structure or attachment. To achieve the desired motion, we quickly settled on using a slider-crank mechanism with the slider translation being the input to the system. This was chosen because we could use a worm gear or lead screw with a motor rotating perpendicular to the direction of translation to achieve high force output with minimal back-drivability, meaning we could use a relatively small motor and the mechanism could passively maintain its position. The output of this first mechanism would be rotation of its most distal link, which we could use as the input to another mechanism.
We experimented with many different four-bar, six-bar, and slider crank configurations for this second mechanism, before eventually settling on an inverted slider crank which produced a favorable motion profile for the distal link of the finger. This profile allowed the distal link to rapidly invert near the mechanism’s maximum contraction, helping widen the stance of the gripper when in the ‘landing-gear’ configuration for maximum stability, and more gradually close with the first mechanism while grasping.
As the primary objects the gripper was intended to grasp were tree branches, which vary significantly in size depending on the specific tree within which a drone may need to perch, the mechanism needed to be able to grasp a range of differently-sized, roughly cylindrical objects. For the purposes of this project, we chose to design the gripper such that it could grasp branches between 3-8 inches in diameter, which would likely be strong enough to support the gripper and the drone to which it were attached but also small enough that a gripper designed to grasp them could be made small enough to reasonably be attached.
We chose to use a stepper motor for its fine positional control, which would be useful for later implementation of closed-loop control, and for its high output torque, which would maximize the achievable grasp force. For materials, most of our parts were 3D printed using PLA. This was desirable because all of our parts have irregular extrusions and cuts of varying depths associated with their pin joints and other features. We benefited significantly from having as many of these features being manufactured into the parts as possible–instead of having to attach many smaller pieces together–though a more rigid material, likely aluminum or some other metal, would be a better choice for a later prototype for its rigidity and strength.
