Design of Nonlinear State Feedback Control
Law for Underactuated TORA System:
A Block Backstepping Approach
Shubhobrata Rudra, Ranjit Kumar Barai, Madhubanti Maitra, Dharmadas Mandai, Shimul Dam, Somsubhra Ghosh,
Prarthna Bhattacharyya and Arka Dutta
Electrical Engineering Department, Jadavpur University, Kolkata-700032, India
Abstract-This paper presents the formulation of a novel
block-backstepping based control algorithm to address the
stabilization problem for the well-known nonlinear
benchmark TOA system. The ideas behind the method are as
follows. At first, state model of the TOA system has been
converted into block-strict feedback form. Then the control
Lyapunov function has been desined for each cascaded
ynamic block to derive the expression of the control input
for the overall nonlinear system. The overall asymptotic
stabiliy of the TOA system has been analyzed using
Lyapunov Stabiliy Criteria. Finally, the efectiveness of the
proposed control algorithm has been verfied in the
simulation environment.
Keywords: Underactuated System, TORA System, Block
Backstepping Control, Lyapunov Stability Criteria.
I. INTRODUCTION
Stabilization of a TORA system (Translational Oscillator with
a Rotational Actuator) has been considered as an active
research area for control system engineers [1-4]. The TORA
system has been considered as a benchmark nonlinear
system for the class of underactuated mechanical systems.
Nowadays, TORA system has received a conspicuous
amount of research attention [1]-[4]. Early research
activities on TORA were initiated by the need of the design
of the controllers to address the complex control problem of
a dual-spin aircrat, where the interaction between spin nd
nutation complicates the algorithm design task for the
conrol system engineers. Out-of-the-way, controlling the
TORA is of independent interest as a benchmark problem in
nonlinear control design [4]. Consequently, devising an
eicient control algorithm for a TORA system remains as
an active area of research. However, the control design of a
TORA system is more complicated than that of a ully
actuated system. Moreover, complicated coupling action
between translational motion of the cart and angular rotation
of the eccentric mass make the control design task more
complicated.
Backstepping is Lyapunov method based versatile control
design approach for nonlinear systems that ensures the
convergence of the regulated variables to zero [5-7].
Backstepping relies on the fact that the system under
978-1-4673-4603-0/12/$31.00 © 2012 IEEE
consideration should be in a strict feedback form. Generally,
single input single ouput systems satisy this condition
under some simpliying assumptions [5-9]. Nevertheless, in
case of multivariable control problem, quite oten the system
sucture is not in he lower riangulr fom (srict feedback
form or semi strict feedback form). In fact, ordinary
backstepping eventually fails to generate a conrol algorithm
for MIMO systems. However, if it is possible to represent the
state space structure of such system in block strict feedback
form by means of some state transformation, it is possible to
address the conrol problems of MIMO system using
backstepping technique. This technique is also known as
block backstepping [8-10].
The seminal contribution of Reza Olfati Saber in the ield of
underactuated system's reserch mkes it possible to fomulate
different types of control algorithm for stabilization problem
of the underactuated system [11-19]. Moreover, it is an
inevitable fact that the proposed transformation [21] of an
underactuated system's state model in the normal form
simpliies the design task for the conrol system engineers.
However, his proposed backstepping based conrol
law [described in 11-13] is somewhat complicated and
design of conrol algorithm for a TORA system using nomal
form is not a very convenient way to design a backstepping
conroller for the TORA system.
The main objective of the paper is to fomulate a block
backstepping based conrol algorithm for the TORA system.
To accomplish the mentioned objective the paper is
organized as follows: section II describes the mathematical
model of the TORA system and precisely formulates the
conrol objective in an analytic manner. Section III describes
the derivation of the proposed algorithm. Section IV Analyzes
the stability of the proposed conrol law. Section V describes
the implantation procedure of the proposed algorithm on
TORA system and analyzes the results. Finally, section V
concludes the work.
II. PROBLEM FORMULATION
The nonlinear TORA system considered in this section was
introduced by Wan, Benstein and Coppola in [4]. The
picture in Fig. 1 illustrates the top view of this nonlinear
benchmark mechanical system in which the rotational