RESEARCH ACTIVITIES


Main Activities and Interests Journal Publications
Templates and Help Files
Projects
Conference Publications
Downloads
Theses
Notes & Technical Reports
Links


Main Activities and Interests


Projects


Theses

Abstract. This work addresses the problem of stabilizing feedback design for strongly nonlinear systems, i.e. systems whose linearization about their  equilibria is uncontrollable and for which there does not exist a smooth or even continuous stabilizer.  The construction of stabilizing controls for these systems is often further complicated by the presence of a drift term in the differential equation describing their dynamics.  This research also considers the development of stabilizing feedback laws for bilinear systems.

The proposed methodologies yield time-varying feedback controls whose construction is based on Lie algebraic techniques.

The contributions of this thesis can be summarized in the development of:
Conditions under which the constructed feedback laws render the corresponding systems asymptotically stable are analyzed.  The applicability and effectiveness of the proposed approaches is demonstrated through computer simulations of several nonlinear systems, including well known nonholonomic driftless systems, such as the kinematic models of a unicycle and a front-wheel drive car, and systems with drift like the dynamic model of a satellite in the challenging actuator failure condition.


Journal Publications
Conference Publications Notes & Technical Reports

Templates and Help Files


Downloads

  • Philip Hall Basis Constructor and Other Routines for Symbolic Lie Algebraic Computations: See the LTP Package.
  • MSTAR Patchwork:  The zip compressed file contains a 512x512 16 bits file with a mosaic created from MSTAR (Public) Clutter collection.  The file has no information header, the data is in raw binary Big Endian format and row-wise written.
  • Numerik Tools version Alpha 0.0:  Zip compressed file containing C routines for ODE integration and least squares problems optimization.  The methods implemented are the Runge-Kutta (RK45) and the Levenberg-Marquardt modification to the Newton-Raphson algorithm, respectively.  These routines have been adapted to handle control system problems in a better way.  In these problems one usually has some function f(x,u) of the state x and the control variable u, instead of the typical autonomous version f(x).  Details and examples have been included.  The routines have been successfully compiled with gcc both under Linux and DOS/djgpp.  (Keywords: ODE RK45 C routine, Levenberg-Marquardt Optimization C routine).

  • Links

    My links to www pages about: can be found here.

    Copyright Miguel Torres-Torriti ©MCMXCIX
    Last modified: 29.Jul.2003 7:38:14 PM