Tong et al. / J Zhejiang Univ Sci A 2008 9(6):833-839
833
Control synthesis for polynomial nonlinear systems
and application in attitude control
*
Chang-fei TONG
†
, Hui ZHANG
†‡
, You-xian SUN
(Institute of Industrial Process Control, Zhejiang University, Hangzhou 310027, China)
†
E-mail: cftong@iipc.zju.edu.cn; zhanghui@iipc.zju.edu.cn
Received Sept. 11, 2007; revision accepted Mar. 10, 2008
Abstract: A method for positive polynomial validation based on polynomial decomposition is proposed to deal with control
synthesis problems. Detailed algorithms for decomposition are given which mainly consider how to convert coefficients of a
polynomial to a matrix with free variables. Then, the positivity of a polynomial is checked by the decomposed matrix with
semidefinite programming solvers. A nonlinear control law is presented for single input polynomial systems based on the
Lyapunov stability theorem. The control synthesis method is advanced to multi-input systems further. An application in attitude
control is finally presented. The proposed control law achieves effective performance as illustrated by the numerical example.
Key words: Nonlinear control, Attitude control, Polynomial systems
doi:10.1631/jzus.A071486 Document code: A CLC number: TP273
INTRODUCTION
In recent years, considerable attention has been
devoted to the study of polynomial nonlinear systems.
Significant progress has been made in the stability
analysis of those systems by sum of squares (SOS)
decomposition approach (Parrilo, 2000; Papachris-
todoulou and Prajna, 2005; Fisher and Bhattacharya,
2007). Stability analysis with this methodology is
mainly based on the Lyapunov stability theorem.
Lyapunov functions are constructed by SOSTOOLS
(Parrilo, 2000) which converts the problems into
semidefinite programs (SDPs) with the SDP problems
being further solved by SeDuMi (Sturm, 1999). Fur-
thermore, Chesi et al. considered the stability of un-
certain polynomial systems (Chesi et al., 2005; Chesi,
2007).
Though stability analysis is solved effectively in
SOS approach, control synthesis for nonlinear sys-
tems still remains a stubborn problem since the
nonlinear components of variables in the SOS terms
cannot be solved directly by SOSTOOLS. To solve
the synthesis problem, an iterative algorithm was
proposed in (Jarvis-Wloszek, 2003) to obtain a stabi-
lized controller. However, the controller is not glob-
ally optimal; furthermore, the iterative algorithm may
fail to get a solution in some cases. Given the diffi-
culties of control synthesis based on Lyapunov sta-
bility theorem, it is most striking to find that the new
convergence criterion presented in (Rantzer, 2001)
based on the so-called
density function ρ has much
better convexity properties. Then, Prajna et al.(2004)
exploited this criterion to solve the control synthesis
problems, and Ataei-Esfahani and Wang (2007) ap-
plied it to the control design of a hypersonic aircraft.
Unfortunately, the convergence criterion via the den-
sity function does not involve any information about
the convergence rate, so the controller designed by
this scheme may perform slow convergence in some
instances. Moreover, the criterion by Rantzer (2001)
Journal of Zhejiang University SCIENCE A
ISSN 1673-565X (Print); ISSN 1862-1775 (Online)
www.zju.edu.cn/jzus; www.springerlink.com
E-mail: jzus@zju.edu.cn
‡
Corresponding author
*
Project supported by the National Natural Science Foundation of
China (Nos. 60674028 and 60736021), the Hi-Tech Research and
Development Program (863) of China (Nos. 2006AA04Z184 and
2007AA041406), the Key Technologies R&D Program of Zhejiang
Province, China (No. 2006C11066), and the Joint Funds of
SFC-Guangdong Province of China (No. U0735003)