Biomimetics and Dextrous Manipulation Lab

HandDesign

Hand Design

Project Summary

We are developing an underactuated robotic hand with the ability to sense what it is grasping, stably grasp a wide range of objects, and utilize different grasping strategies based on the task requirements. This is accomplished through the development of new technologies such as electrolaminates, electroadhesion, twisted cable actuators, and a capacitive sensing suite capable of sensing a wide range contact forces over the inner surface of the fingers.

About the Team

The team is comprised of several students from the BDML, the robotics group at SRI International, and Meka robotics, a robotics design firm located in San Francisco.

SRI is leading the development, focusing on integration, hand design, electrolaminates, and electroadhesives. Stanford is contributing to the project in the topics of hand design, sensing, and control. Students from the BDML are developing a sensor suite which will give the hand high resolution tactile feedback and joint position sensing using capacitive sensor technology. Stanford is also working with SRI to develop control strategeies that take advantage of the hand to lock leverageto lock in any position. Meka is developing a new actuator design which will give the hand backdriveability and high torque in a small package.

The Questions

The SRI hand is being developed with electrolaminates in each joint. Electrolaminates employ a method of electrostatic adhesion which allows the hands joints to lock in any position. This enables the hand to adopt postures which are unavailable to other underactuated hands. With this locking technology, the hand is to act as a fully actuated hand or an underactuated hand depening on the grasping requirements. There are several questions we wish to answer through development of the hand:

  • With a lockable underactuated hand, what control strategies are available that are unavailable to other underacutated hands?
  • Does underactuation with locking allow one to grasp a larger range of objects stably than with a highly underactuated hand?
  • Can

Hypothesis

  • Design Methods
    • Estimating grasp stability over a wide range of objects to optimize design
  • Several Control methods are proposed
    • Optimized trajectory control
    • Manipulation by with strategic locking.
    • Manipulation with

Research

Data

Observations

Conclusion

References

Acknowledgements

Resources

ForceTorqueSensing

Page last modified on June 23, 2014, at 11:18 AM