A prehensile robotic hand with a floating body for underactuation

The lab's first prosthetic hand shared the force of one motor among the fingers with pulleys in series, and lost a good part of that force to friction. This project redesigned the mechanics. A floating body in the palm replaces the intermediate pulleys, the thumb gets its own motor, and the hand's dimensions are parameterized to fit each user. The prototype holds cylindrical and spherical objects well and has difficulty with fine pinch.
Context
The cost target was ambitious: to come in below the price of a hook prosthesis, which is the only one many users have access to, and offer in exchange a five-finger hand with adaptive grasp.
Design
Parametric dimensions. Since there is no generic hand size, the measurements of the palm and phalanges were defined as model parameters, so the prosthesis can be scaled to the user's age and build.
Floating body. A moving part inside the palm receives the motor tension at its center and delivers it to the finger tendons at its ends. Because the part can rotate and translate, when one finger stops against the object the others keep closing. Force is shared with fewer changes of cord direction than in a pulley train.
Floating body in the palm for finger underactuation
Layout of the cords transmitting force to each finger
Fingers. Each finger closes with a tendon and opens with elastic elements at the joints. The stiffness of each joint determines the order in which the phalanges flex, and it was chosen so that closure wraps around the object.
Thumb. It has an independent motor and can act as a support or close against the other fingers.
Cylindrical and spherical power grasp in the model
Tests
The hand was 3D printed and mounted on a bench with DC motors.
Hand mounted on the test bench
Cylindrical power grasp test
Tests with irregularly shaped objects
Phalanx motion was recorded on video to verify the closing sequence.
Angular position of the index proximal phalanx during closure
Results
- Good performance in power grasp, cylindrical and spherical, and with irregular objects.
- Poor performance in fine, strong pinch. It is possible with some skill, but the thumb design limits it. Increasing the stiffness of the distal interphalangeal joints improves it.
- Cost came in below that of a hook prosthesis.
- The floating body reduced friction losses and distributed motor force well.
What is missing
The hand was tested on a bench, with no user, no sensors and direct motor command. There are no grasp force measurements. The socket and attachment to the residual limb are not part of the work.
How it fits in Robiolab
This is the second project in the prosthetic hand line, and the design inherited by the next ones: the force sensor with its controller and the integrated version tested with a user. The floating body is a differential mechanism, a solution the human hand also uses through its interconnected tendons, and a good example of how an idea from biomechanics simplifies a design.
