Introduction
Human hands have been highlighted in robotics, and researchers and engineers constantly try to replicate anthropomorphic hands with various actuation means. However, due to technological constraints and current knowledge, the human hand is extremely difficult to replicate due to its complexity with all the tendons, bones, and muscles required to move the 27 degrees of freedom. In the attempts to replicate the human hands, the cost has increased, and the designs haven’t become any less complex. Our project will attempt to go down another path: lower the price as much as possible, reduce the number of actuators to a singular motor, and maintain as much functionality as possible.
Problem Statement
Currently, robotic hands, such as prosthetic hands, which sparked the idea for this project, are very expensive, and this high cost comes primarily from the difficulty with manufacturing. They also require a very One of the most striking examples of the gulf between human ingenuity and the designs of nature is the great degree of effort and technical knowledge an engineer needs to mimic even the simplest of natural mechanisms. There is perhaps no better example of this fact than the human hand: an assemblage of skin, muscles, tendons, ligaments, cartilage and bone that is capable of remarkably complex and precise motion. When looking at the vast majority of robotic hands, one of the first design realities that becomes apparent is the need for many different actuators to accomplish a product that resembles the real thing. The human hand has 27 degrees of freedom, so it is not hard to see why the amount of actuators required can easily become unwieldy when trying to replicate it. In this conundrum, the simplicity of flesh is lost, and the manufacturing cost can easily become prohibitive. With our project, we aim to explore another path: a robotic hand with as few actuators as possible.
Problem Statement
Many robotic hands designed today make use of numerous small actuators in combination with tensioning cables, which must be precisely toleranced and calibrated. The highly time- and labor-intensive development cycle in which the hardware and software must be optimized to perform their intended function. This primary pain in the development is from the design of the hand enclosure, which requires very high levels of tolerancing and a ton of research to fit the tiny drive motors and tensioning cables. Another issue is with the mechanism itself. Due to the use of tensioning cables, frequent re-tensioning and calibrations are required, adding to the cost and complexity of the hand.We aim to design and manufacture a cheap robotic hand based purely on linkages and compliant/dampening designs. It will have one actuator and be able to grasp small objects, such as a small ball or a pencilrequired is one of the main reasons for their high cost. As part of our goal of minimizing the amount of actuators we use, motors and pistons must be replaced with mechanically interconnected linkages. Even with only one motor, we would like to be able to flex and extend our robotic fingers fully. If time permits, we would also like to animate the wrist.
Additionally, reducing the amount of actuators used is a goal that is intrinsically at odds with the goals of precision and delicacy. There is no purpose in simplifying your design if it cannot accomplish its intended function. We would like our product to have the required force to firmly grasp a small object without using so much force that the object is damaged. As in a real human hand, we want our robotic hand to be able to grasp objects of differing shapes and orientations. To do so, we will need to build in some inherent compliance to our mechanism so that our fingers can close tight around our object no matter what direction it is oriented in.
Ultimately, the problem we face is to find a middle ground between design simplicity and operational precision.
Mechanism
The primary mechanism that will actuate the project will be the crossing 4-bar linkage, which will act as a crank, giving us complex motion. Each finger will be constructed by two colocated crossing 4-bar linkages, enabling the finger to open and close all three joints, covering the flexion and extension movement of the finger. However, with this design, we are planning to forfeit abduction and adduction movements of the fingers to minimize the complexity.
...
Our plan for producing this robotic hand includes researching the anatomy of the human hand, existing linkage designs for tension-driven robotic hands, experimenting with various linkage designs, software optimization, prototyping, testing, and improvements. We aim to complete our set objective of creating a pure linkage-based hand that can at least grasp some objectsfor our hand to be able to repeatably hold small objects in varying orientations, with built-in compliance to allow for flexibility in what we test it on. This will likely include a lot of test designing, kinematic simulations, and prototyping. If possible, we would like to perfect the design and functionality of the hand by adding more range of motion and sensor-driven features in the future. Additionally, we would like to put effort into the form factor of our hand. We want it to be sleek and aesthetically pleasing, and for it to connivingly model the human hand.
Preliminary Design
Preliminary design of the fingers
We will be referring to the finger joints by their anatomical name. Figure 1 will show what the names of the joints are. Also, the crossing 4-bar mechanisms are 1 DOF based on the Gruebler equation: M = 3(4-1) - 2(4) = 1
Figure 1: Graph showing and labeling finger joints with their anatomical name.
Crossing 4-bar mechanism from MCP to PIP Joints
Figure 2 shows the planned preliminary design for the mechanism between the MCP and PIP joint, with an initial length. The actual lengths are still to be determined.
Figure 2: Preliminary design of the MCP - PIP crossing 4-bar mechanism
| Link Name | Link Length (mm) |
|---|---|
| L1 | 6.3 |
| L2 | 28.2 |
| L3 | 9.5 |
| L4 | 27.3 |
Grashof Condition
6.3 + 28.2 < 9.5 + 27.3
34.5 < 36.8
Class I Grashof: Crank-rocker
Crossing 4-bar mechanism from PIP to DIP Joints
Figure 3 shows the planned preliminary design for the mechanism between the PIP and DIP joint, with an initial length. The actual lengths are still to be determined.
Figure 3: Preliminary design of the PIP - DIP crossing 4-bar mechanism
| Link Name | Link Length (mm) |
|---|---|
| L5 | 9.4 |
| L6 | 24.8 |
| L7 | 6.0 |
| L8 | 20.1 |
Grashof Condition
6.0 + 24.8 > 9.4 + 20.1
30.8 > 29.5
Class II Grashof
Various ideas for achieving compliance within the hand.
Figure
...
4:
...
shows the side profile of the pawm and finger mechanism. This is one potential design where the finger mechanism will close to a percentage of fully grasping then the compliant link will snap the finger close.
Figure
...
5: shows another potential configuration where the figure mechanism stays the same without the added compliant part. (4-bar kinematic chain) The palm shown has the linear actuator linked to a dampening kinematic chain that drives the finger mechanism.