Loading...

Design of Globally Convergent Observers for Robotic Systems

Lotfi Yagin, Nima | 2010

579 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40145 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Namvar, Mehrzad; Mobed, Mohammad
  7. Abstract:
  8. In a typical rigid robot manipulator, joint angels and velocities are often considered as system states. Joint angels are usually measured by precise Shaft Encoders. However, lack of velocity measurement sensors “Tachometers” and their undesired characteristics in most of existing robotic systems make it difficult to have access to full system states. As full access to the system states in various applications is inevitable, velocity observers can be used to reconstruct or estimate the unmeasured velocity signals. This thesis presents an observer for global estimation of joint velocities in robot manipulators. A non-minimal model of a robotic manipulator is introduced to design an adaptive observer based on this model. Stability analysis of the proposed observer is presented using Lyapanov stability analysis. Smoothness of the dynamics of the proposed observer allows its easy implementation in comparison with non-smooth observers. Dimension of the proposed observer is shown to be at least where stands for the manipulator degrees of freedom. This number is less than the dimension of most existing globally convergent adaptive observers. The novelty of this observer is that, in addition to its global asymptotic convergence, it can also handle uncertainties in the entire robot dynamics, a problem which had not been previously adressed in literature. Smoothness of the dynamics of this observer is another advantage of it in comparison with existing observers. With negligible sensor noise, global asymptotic convergence of state estimates to their true values is achieved under no persistency of excitation condition. Besides, it has been shown that, in the presence of measurement noises, estimation errors of the proposed observer will be bounded and its ultimate bound can be easily computed. In order to evaluate the performance of the proposed observer, it has been implemented on a 2-DOF planar robot manipulator and its effectiveness is illustrated through simulations. Besides, simulation results imply low noise sensitivity of the proposed observer in comparison with non-smooth observers.
  9. Keywords:
  10. Robot Nonminimal Model ; Robot Manipulator ; Global Estimation ; Adaptive Velocity Observer

 Digital Object List

 Bookmark

No TOC