ACTIVE FAULT TOLERANT CONTROL :
APPLICATION TO A ROBOT MANIPULATOR
S. Aberkane * H. Noura ** D. Sauter *
* CRAN – CNRS UMR 7039
Universit´e Henri Poincar´e, Nancy 1, BP 239,
F-54506 Vandœuvre-l`es-Nancy Cedex
Tel.: +33 3 83 68 44 74, Fax: +33 3 83 68 44 61
e-mail: samir.aberkane@cran.uhp-nancy.fr
** LSIS – UMR CNRS 6168
Domaine Universitaire de Saint-J´erˆome
avenue Escadrille Normandie-Niemen
F-13397 Marseille cedex 20
Abstract: Fault tolerance is an important design criterion for robotic systems
operating in hazardous or remote environments. In this paper we firstly present a
new FDI method which deals with an actuator failures a®ecting a robot manipulator.
This method is based on the linearization of equations of motion (via a feedback
linearization control) and the use of a bank of unknown inputs observers which are
synthesised iteratively. Then, we introduce a fault tolerant control module which can
deals with critical and non critical failures.
Keywords: Robot manipulator - feedback linearization - fault detection and isolation
- fault tolerant control .
1. INTRODUCTION
Robots have found a niche working in dangerous environments.
Robots are used in deep-sea operations,
space missions, nuclear cleanup. This critical situations
demands robots because hostile environments
to be fault tolerant.
In this article, we present a fault tolerant control
scheme for a robot manipulator. This scheme can
be divided into 3 principal parts: first, an FDI block
which is based on the linearization of the equations
of motion (via a feedback linearization control)
and the synthesis of a bank of unknown inputs
observers which are computed iteratively, secondly
a first fault tolerant control module dedicated to
non critical failures and which is based on adaptive
feedback linearization control principals (Ortega
and Spong, 1989). This method has the drawback
of using informations about acceleration signals,
but in our fault tolerant control application this
informations are not required and we just exploit
the fact that this method uses an additional outer
loop control. And finally a second fault tolerant
control module dedicated to critical failures. In this
paper, a critical failure is considered as a total loss
of one actuator. The manipulator is then considered
as an under-actuated one. The second fault tolerant
control module is based on the control of such a
manipulators (Bergman and Xu, 1996).
This paper is organized as follows. The second section
is devoted to the modelling of the robot manipulator.
Section 3 will first treat about feedback
linearization control of the manipulator, secondly
about the adaptive feedback linearization control
and finally about the control of under-actuated
robots. In section 4 we will present the fault tolerant
control scheme adopted in our application. Section
5 concludes this work and gives some prospects.
.....