19.1 - Project Proposal
Introduction
Robotic hands and fingers have advanced significantly in recent years and are becoming increasingly important in applications such as prosthetics and service robotics. Millions of people around the world rely on prosthetic or orthotic devices to perform everyday tasks. This creates a strong motivation to improve hand and finger designs in order to enhance accessibility and overall quality of life.
Current robotic hand designs generally fall into two extremes. On one end, simple grippers are reliable and easy to control but often lack the ability to adapt to objects with different shapes and sizes. On the other end, highly dexterous robotic hands can perform complex motions but require many actuators, making them more expensive, mechanically complex, and less accessible.
This gap motivates the focus of our project: designing a robotic hand that balances dexterity with simple actuation. To address this challenge, we aim to design and test compliant fingers that can adaptively grasp a variety of objects without requiring a motor at every joint.
Problem Statement
The goal of this project is to design a finger mechanism capable of producing complex grasping motions using limited actuation. Specifically, the mechanism must have at least two degrees of freedom while being driven by a single actuator input. This creates the challenge of creating a complex mechanical linkage system to produce adaptive position profiles during grasping.
The mechanism must be able to grasp objects of different shapes and sizes, which requires the finger to conform to object geometry and maintain stable contact throughout the grasp. Achieving this adaptive motion cannot be accomplished using simple joints with fixed trajectories.
Additional mechanical complexities must also be considered. The mechanism should maintain favorable transmission angles to ensure efficient force transfer from the actuator to the grasping surfaces. Toggle point configurations that limit motion should also be avoided. Furthermore, link lengths and actuator placement significantly influence the resulting motion profile, making their selection critical to the mechanism’s performance.
Mechanism
The proposed mechanism is based on the design presented in the 2017 paper “A Novel Under-Actuated Bionic Hand and Its Grasping Stability Analysis” [1]. The finger uses an under-actuated design to produce a human-like closing motion while requiring only a single actuator. This allows the finger to avoid simple gripper-like closing paths, and instead, it mimics the sequential flexion seen in human fingers.
Although the finger is structurally a five-bar linkage, the presence of a torsional spring allows the motion to be analyzed as two separate four-bar mechanisms. Before object contact, the mechanism behaves as the first four-bar system and drives the initial finger motion.
When the finger contacts an object, the torsional spring begins to deflect, effectively changing which links are grounded. This shifts the mechanism into a second four-bar configuration, allowing the distal portion of the finger to continue actuating using the same motor input and wrap around the object.
For this project, the design is simplified to two finger segments rather than three while preserving the key under-actuated behavior of the mechanism. The figure below comes directly from the paper, showing the under-actuated finger design at different stages.
Proposed Scope
The scope of this project is to design and test an under-actuated finger mechanism capable of adaptive grasping with a single actuator. Considering the time constraints of the semester, the final project deliverable will be a simplified robotic hand with two compliant fingers. For the prototype, we plan to fabricate one working finger to demonstrate the grasping behavior of the mechanism.
Position analysis will be performed on the two four-bar mechanisms that describe the finger motion before and after object contact. This analysis will focus on selecting appropriate link lengths and avoiding unfavorable configurations such as toggle points. Position analysis will also allow us to evaluate different points along the linkages themselves and not just the endpoints of the links. It will be important to consider the geometric shape of the finger links as they will influence grasping performance.
Additionally, sizing the torsional spring and analyzing its behavior will be highly critical. This will ensure the mechanism transitions properly between the two motion phases and produces the intended sequential joint actuation.
If extended beyond the current scope, the design could be expanded by adding a third finger, incorporating circumduction motion for a more flexible thumb, and integrating force sensing for more precise grasping. This project is directly related to the research efforts performed at the Human Centered Robotics Lab (HCRL) at UT Austin.
Preliminary Design
The preliminary design of the compliant finger mechanism is shown below. Structurally, the finger can be modeled as a five-bar mechanism as seen in Figure 2. However, due to the presence of a torsional spring at the joint connecting links 4 and 5, the mechanism will behave as two four-bar mechanisms that operate sequentially. The four-bars will be analyzed separately.
In the first stage, before the finger contacts an object, the system can be simplified as the following four-bar mechanism (Figure 3). During this phase, the torsional spring does not deflect, so the actuator drives the entire finger to rotate toward the object as a collective unit. Using the Gruebler equation, the mechanism is confirmed to have one degree of freedom (M=1).
Once the finger makes contact with the object, the torsional spring begins to flex. This changes the effective grounding of the links and transitions the system into a second four-bar configuration (Figure 4). As the actuator continues to drive link 2, the distal finger segment rotates relative to the proximal segment, allowing the finger to wrap around the object and increase contact area. The Gruebler equation again yields one degree of freedom, demonstrating that the sequential motion is achieved with a single actuator input.
The final visualization shows the proposed design using two identical, under-actuated fingers to produce an adaptive grasping motion. All schematics in this section are not drawn or modeled to scale.
References
[1] Li, Xin & Huang, Qiang & Chen, Xuechao & Yu, Zhangguo & Zhu, Jinying & Han, Jianda. (2017). A novel under-actuated bionic hand and its grasping stability analysis. Advances in Mechanical Engineering. 9. 168781401668885. 10.1177/1687814016688859.