Gripper

Problem Identification

This year, major changes have been made to the grabbing object, where instead of the cube (left image), it has changed to a cylindrical object with uneven surfaces. This therefore forces us to change the end effector of the engineer robot.

Below are in detail specification, right shows the retrieval area for the energy cylinder, left shows the interactable area of the energy bars during deposit. Due to DJI being slow as always, they did not publish the specifications for the new deposited area. Therefore, based on the image and the additional textual hints that I will not go into here, we suspect a depositing sequence of similar nature to lock picking. Where it is likely required for the end effector to insert the energy cylinder to a tight space, then perform action to complete depositing.

Design Criterias

There are only two mainstream end effector designs that are created for object grabbing, that is, suction and grippers. With the first being the choice for almost all teams in the previous season. However, this year this option will no longer be viable, since the side of the cylinder is not a surface. Therefore, we opt for develop a gripper for this year's design, more specifically a two-piece gripper.

Based on given information, this new gripper design needs to meet the following criteria.

Needs to be able to handle 500g of load that is perpendicular to the direction of gripper opening
Each side of the gripper needs to handle 500g of load individually
Needs to be able to fit within a 200 x 95 x 75mm area after gripping a 65mm object
Needs to have a minimum 95mm distance between grippers when fully extended
Needs to be able to hold on to the uneven surface on the side surface of the cylinder.

Design

Gripper

After some discussion with my teammates, I put my eyes on soft gripper structures, specifically - Fin-Ray Effector Grippers. These grippers become popular for their excellent adaptability to the shape of the object that are gripping. These gripper utilizes soft materials, and strucutures that collapse under load to curve aroud the target gripping object. This allows mulitple point of contact and a larger surface area to allow a more secure grip.

A example I create, functioned effectively and was able to grip the object with success. The problem with this was however, that it does not fit with 200 x 95 x 75mm requirement. The triangular structure reaches almost8 80mm at the base.

With some calculation, I found to meet the requirement: Needs to be able to fit within a 200 x 95 x 75mm area after gripping a 65mm object. The gripper can have a maximum thickness of 15mm, which in practice is actuallally 14mm because space is needed for the power cylinder to be released from the gripper.

With that I designed and tested a variety slimmer gripper option that I designed. With the bottom most being the first iteration, the top being the second and the middle two being the forth and third accordingly.

Improvements over the iterations includes; Softer TPU material; Additon of a piece that catches the cube when pushed in, allowing for better positioning and more tip deformation (Top Right); Adding support beams on the bottom, which betters the collapse structures, also enhancing the deformation, with the latest being the best (Bot Left).

The improve is also evident when we look at the FEA results. Under the same load conditions, the latest version demonstrated significantly more deformation (2x) then the second iteration.

NOTE!!: These numbers are used for parallel comparison only, as the material information's are not 100% incorrect.

Gripper Mount

This an equally import part of the gripper, for this I referenced a paper by (Mohammad), seen below. Where they experimented with a support structure instead of conventional mounting, which they found to have increased the vertical load of the grippers by at least 120%.

Mohammad Sheikh Sofla, Hanita Golshanian, Elizabeth I. Sklar, Marcello Calisti, Modeling and modification of fin-ray effect grippers to improve their load capacity and grasp stability Sensors and Actuators A: Physical, Volume 392, 2025, 116711, ISSN 0924-4247, https://doi.org/10.1016/j.sna.2025.116711.

This is the first prototype, and had several mechanical flaws, specfically:

Attachment to linear rail to thing, cracks
Needs to support the all the way at the tip of the gripper (for level one exchange)

This is the final version of the gripper mount, which now adapts to the thinner gripper.

In addition to solving the problems with the previous version. I also placed more emphasis on controlling the deformation of the structure in this version, where the ribs on the sides are designed to resist as much bending force outwards as possible while still preserving enough space for the deformation of soft gripper.

Mechanics

We utilized a rack and pinion structure to control the gripping motion of the gripper. This is because the tight spaces, means only linearly actuated mechanisms are viable options, and rack and pinion is among the more reliable and easy ones to design.

My specific design utilizes two linear rails, which originally was because it was material we had in the workshop, but it actually ended up working very neatly as end stoppers for each other's.

In this design the two linear rails and the overarching structural piece creates a very rigid structure double layer structure, for force coming from all directions.

For the motor, I choose a DJI GM2006, which had a max output of 1Nm. This when translated through our mechanisms is about 70N of normal force per finger, and assuming coefficient of friction 0.2, is 14N of frictional force per finger. This is plenty to hold an object of 0.5kg.

Lastly is the attachment piece for the linear carriage and the gripper mount, while also responsible for attaching the linear racks. Not much special about these pieces, just made extra rigid to try and prevent any deformation

Manufacturing

All the pieces, apart from the soft gripper will be cnced from 6601 Aluminium Alloy.

Page updated

Google Sites

Report abuse