skip to main content
Robot arm control exploiting natural dynamics
Publisher:
  • Massachusetts Institute of Technology
  • 201 Vassar Street, W59-200 Cambridge, MA
  • United States
Order Number:AAI0800843
Pages:
1
Bibliometrics
Skip Abstract Section
Abstract

This thesis presents an approach to robot arm control exploiting natural dynamics. The approach consists of using a compliant arm whose joints are controlled with simple non-linear oscillators. The arm has special actuators which makes it robust to collisions and gives it a smooth compliant motion. The oscillators produce rhythmic commands of the joints of the arm, and feedback of the joint motions is used to modify the oscillator behavior. The oscillators enable the resonant properties of the arm to be exploited to perform a variety of rhythmic and discrete tasks. These tasks include tuning into the resonant frequencies of the arm itself, juggling, turning cranks, playing with a Slinky toy, sawing wood, throwing balls, hammering nails and drumming.

For most of these tasks, the controllers at each joint are completely independent, being coupled by mechanical coupling through the physical arm of the robot. The thesis shows that this mechanical coupling allows the oscillators to automatically adjust their commands to be appropriate for the arm dynamics and the task. This coordination is robust to large changes in the oscillator parameters, and large changes in the dynamic properties of the arm.

As well as providing a wealth of experimental data to support this approach, the thesis also provides a range of analysis tools, both approximate and exact. These can be used to understand and predict the behavior of current implementations, and design new ones. These analysis techniques improve the value of oscillator solutions.

The results in the thesis suggest that the general approach of exploiting natural dynamics is a powerful method for obtaining coordinated dynamic behavior of robot arms. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.)

Cited By

  1. Fang K, Zhu Y, Garg A, Kurenkov A, Mehta V, Fei-Fei L and Savarese S (2020). Learning task-oriented grasping for tool manipulation from simulated self-supervision, International Journal of Robotics Research, 39:2-3, (202-216), Online publication date: 1-Mar-2020.
  2. ACM
    Bretan M and Weinberg G (2016). A survey of robotic musicianship, Communications of the ACM, 59:5, (100-109), Online publication date: 26-Apr-2016.
  3. Jain A, Killpack M, Edsinger A and Kemp C (2013). Reaching in clutter with whole-arm tactile sensing, International Journal of Robotics Research, 32:4, (458-482), Online publication date: 1-Apr-2013.
  4. Al-Busaidi A, Zaier R and Al-Yahmadi A Control of biped robot joints' angles using coordinated matsuoka oscillators Proceedings of the 22nd international conference on Artificial Neural Networks and Machine Learning - Volume Part I, (304-312)
  5. Xiao J, Song X, Su J and Xu X Gait planning research for biped robot with heterogeneous legs Proceedings of the 11th Pacific Rim international conference on Trends in artificial intelligence, (693-698)
  6. Boblan I, Bannasch R, Schulz A and Schwenk H A human-like robot torso ZAR5 with fluidic muscles 50 years of artificial intelligence, (347-357)
  7. Clark M, Anderson G and Skinner R (2019). Coupled Oscillator Control of Autonomous Mobile Robots, Autonomous Robots, 9:2, (189-198), Online publication date: 1-Sep-2000.
  8. Adams B, Breazeal C, Brooks R and Scassellati B (2000). Humanoid Robots, IEEE Intelligent Systems, 15:4, (25-31), Online publication date: 1-Jul-2000.
Contributors
  • MIT Computer Science & Artificial Intelligence Laboratory
  • Massachusetts Institute of Technology

Recommendations