Skip to main content
Log in

Global stabilization of inherently non-linear systems using continuously differentiable controllers

  • Original Paper
  • Published:
Nonlinear Dynamics Aims and scope Submit manuscript

Abstract

This paper concerns the problem of constructing \(C^1\) (continuously differentiable) controllers to stabilize a class of uncertain non-linear systems whose linearization around the origin may contain uncontrollable modes. Based on a new definition of homogeneity with monotone degrees, a polynomial Lyapunov function and a \(C^1\) global stabilizer are constructed recursively. Moreover, several special cases are investigated to show the advantages of the proposed approaches using the generalized homogeneity compared to the existing approaches using the traditional homogeneity.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

Similar content being viewed by others

References

  1. Bhat, S.P., Bernstein, D.S.: Continuous finite-time stabilization of the translational and rotational double integrators. IEEE Trans. Autom. Control 43(5), 678–682 (1998)

    Article  MATH  MathSciNet  Google Scholar 

  2. Brockett, R.W.: Asymptotic stability and feedback stabilization. In: Differential geometric control theory (Houghton, Mich., 1982), volume 27 of Progr. Math., pp. 181–191. Birkhäuser Boston, Boston, MA, (1983).

  3. Byrnes, C.I., Isidori, A.: New results and examples in nonlinear feedback stabilization. Syst. Control Lett. 12(5), 437–442 (1989)

    Article  MATH  MathSciNet  Google Scholar 

  4. Cheong, S., Back, J., Shim, H., Seo, J.: Non-smooth feedback stabilizer for strict-feedback nonlinear systems not even linearizable at the origin. In Proceedings of 2005 American Control Conference, pp. 1907–1912, (2005).

  5. Dayawansa, W., Martin, C., Knowles, G.: Asymptotic stabilization of a class of smooth two dimensional systems. SIAM J. Control Optim. 28(6), 1321–1349 (1990)

    Article  MATH  MathSciNet  Google Scholar 

  6. Du, H., Li, S.: Finite-time attitude stabilization for a spacecraft using homogeneous method. J. Guid. Control Dyn. 35(3), 740–748 (2012)

    Article  Google Scholar 

  7. Du, H., Qian, C., Yang, S., Li, S.: Recursive design of finite-time convergent observers for a class of time-varying nonlinear systems. Automatica 49(2), 601–609 (2012)

    Article  MathSciNet  Google Scholar 

  8. Gendelman, O.V.: Nonlinear normal modes in homogeneous system with time delays. Nonlinear Dyn. 52(4), 367–376 (2008)

    Article  MATH  Google Scholar 

  9. Haimo, V.: Finite time controllers. SIAM J. Control Optim. 24, 760–770 (1986)

    Article  MATH  MathSciNet  Google Scholar 

  10. Hong, Y.: Finite-time stabilization and stabilizability of a class of controllable systems. Syst. Control Lett. 46, 231–236 (2002)

    Article  MATH  Google Scholar 

  11. Isidori, A.: Nonlinear Control Systems. Communications and Control Engineering Series, 3rd edn. Springer, Berlin (1995)

    Book  MATH  Google Scholar 

  12. Kawski, M.: Homogeneous stabilizing feedback laws. Control Theory Adv. Technol. 6(4), 497–516 (1990)

    MathSciNet  Google Scholar 

  13. Kokotovic, P.V., Freeman, R.: Robust Nonlinear Control Design: State-Space and Lyapunov Techniques. Springer, Berlin (1996)

    MATH  Google Scholar 

  14. Lin, W., Qian, C.: Adding one power integrator: a tool for global stabilization of high-order lower-triangular systems. Syst. Control Lett. 39(5), 339–351 (2000)

    Article  MATH  MathSciNet  Google Scholar 

  15. Lin, W., Qian, C.: Robust regulation of a chain of power integrators perturbed by a lower-triangular vector field. Int. J. Robust Nonlinear Control 10(5), 397–421 (2000)

    Article  MATH  MathSciNet  Google Scholar 

  16. Mahmoud, M.S.: Decentralized Systems with Design Constraints. Springer, Dordrecht (2011)

    Book  MATH  Google Scholar 

  17. Marek, S., Stanislaw, W.: Application of joint coordinates and homogeneous transformations to modeling of vehicle dynamics. Nonlinear Dyn. 52(4), 377–393 (2008)

    Article  MATH  Google Scholar 

  18. Polendo, J.: Global synthesis of highly nonlinear dynamic systems with limited and uncertain information. PhD thesis, University of Texas at San Antonio (2006).

  19. Polendo, J., Qian, C.: A generalized homogeneous approach for global stabilization of inherently nonlinear systems via output feedback. Int. J. Robust Nonlinear Control 17(7), 605–629 (2007)

    Google Scholar 

  20. Polendo, J., Qian, C., Schrader, C.: Homogeneous domination and decentralized control problem for nonlinear system stabilization. In: Won, C., Schrader, C.B., Michel, A.N. (eds.) Advances in Statistical Control, Algebraic Systems Theory, and Dynamic Systems Characteristics, pp. 257–280. Birkhauser, Cambridge (2008)

    Chapter  Google Scholar 

  21. Qian, C., Lin, W.: A continuous feedback approach to global strong stabilization of nonlinear systems. IEEE Trans. Autom. Control 46(7), 1061–1079 (2001)

    Article  MATH  MathSciNet  Google Scholar 

  22. Rosier, L.: Homogeneous lyapunov function for homogeneous continuous vector fields. Syst. Control Lett. 19(6), 467–473 (1992)

    Article  MATH  MathSciNet  Google Scholar 

  23. Rui, C., Reyhangolu, M., Kolmanovsky, I., Cho, S., McClamroch, H.N.: Non smooth stabilization of an underactuated unstable two degrees of freedom mechanical system. IEEE Conf. Control Decis. 4, 3998–4003 (1997)

    Google Scholar 

  24. Sun, H., Li, S., Sun, C.: Finite time integral sliding mode control of hypersonic vehicles. Nonlinear Dyn. 73(1), 229–244 (2013)

    Article  MATH  Google Scholar 

Download references

Acknowledgments

This work is supported in part by the U.S. National Science Foundation under Grant No. HRD-0932339, the Valero Research Excellence Award, the Natural Science Foundation of China (61074013), the Program for New Century Excellent Talents in University (NCET-10-0328), and the Science Foundation for Distinguished Young Scholars of Jiangsu Province (BK20130018).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Shihua Li.

Appendix

Appendix

This appendix collects the proofs of the propositions used in the proof.

Proof of Proposition 1

The proof of this proposition will be separated into two different cases.

Case 1

If \({\frac{r_{i}p_{i-1}}{r_n+k_n}}\le 1\), applying Lemma 1 gives

$$\begin{aligned}&|x_{i}^{p_{i-1}}-x_{i}^{*p_{i-1}}|\nonumber \\&\quad = \left| \left( {x_{i}}^{\frac{r_n+k_n}{r_{i}}}\right) ^{ \frac{r_{i}p_{i-1}}{r_n+k_n}}-\left( {x_{i}^*}^{\frac{r_n+k_n}{r_{i}}} \right) ^{\frac{r_{i}p_{i-1}}{r_n+k_n}}\right| \nonumber \\&\quad \le 2^{1-\frac{r_{i}p_{i-1}}{r_n+k_n}}\left| {x_{i}}^{ \frac{r_n+k_n}{r_{i}}}-{x_{i}^*}^{\frac{r_n+k_n}{r_{i}}} \right| ^{\frac{r_{i}p_{i-1}}{r_n+k_n}}\nonumber \\&\quad \le 2 \left| \xi _i\right| ^{\frac{r_{i}p_{i-1}}{r_n+k_n}}. \end{aligned}$$
(47)

On the other hand, we have

$$\begin{aligned}&{x_{i-1}}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}-{x_{i-1}^*}^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\nonumber \\&\quad =\left( {x_{i-1}}^{\frac{r_n+k_n}{r_{i-1}}}\right) ^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}}\nonumber \\&\quad \quad -\left( {x_{i-1}^*}^{\frac{r_n+k_n}{r_{i-1}}}\right) ^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}}. \end{aligned}$$
(48)

By the definitions of \(\rho \) and \(k_j\)’s, \(j=1,2,\cdots ,i\), one can tell that the power \(\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}\ge 1\). Thus, we can apply the Mean Value Theorem to (48), then using Lemma 1 and by (9) one can obtain the following

$$\begin{aligned}&\left| \left( {x_{i-1}}^{\frac{r_n+_n}{r_{i-1}}}\right) ^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}}- \left( {x_{i-1}^*}^{ \frac{r_n+k_n}{r_{i-1}}}\right) ^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}} \right| \nonumber \\&\quad \le c'_i|\xi _{i-1}| \Big (\left| {x_{i-1}}^{\frac{r_n+k_n}{r_{i-1}}} \right| ^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\nonumber \\&\quad +\left| {x_{i-1}^*}^{\frac{r_n+k_n}{r_{i-1}}}\right| ^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\Big )\nonumber \\&\quad \le c'_i|\xi _{i-1}| \Big (\left| \xi _{i-1}+\beta _{i-2}^{ \frac{r_n+k_n}{r_{i-2}+k_{i-2}}}(\cdot )\xi _{i-2}\right| ^{ \frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\nonumber \\&\quad +\left| \beta _{i-2}^{\frac{r_n+k_n}{r_{i-2}+k_{i-2}}} (\cdot )\xi _{i-2}\right| ^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1} \Big )\nonumber \\&\quad \le \tilde{c}_i(\cdot )|\xi _{i-1}|\left( |\xi _{i-1}|^{\frac{2\rho -k_{i-1} -r_{i-1}}{r_n+k_n}-1}\right. \nonumber \\&\quad \left. +|\xi _{i-2}|^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\right) , \end{aligned}$$
(49)

where \(c'_i\) is a constant and \(\tilde{c}_i(x_1,\cdots ,x_i)\) is a positive function. With the help of (47) and (49), utilizing Lemma 2 and noting \(r_{i}p_{i-1}=r_{i-1}+k_{i-1}\), we have

$$\begin{aligned}&\left| \left( {x_{i-1}}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\! -\!{x_{i-1}^*}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\!\right) (x_{i}^{p_{i-1}}-x_{i}^{*p_{i-1}})\right| \nonumber \\&\quad \le 2\tilde{c}_i(\cdot )|\xi _{i-1}|\left( |\xi _{i-1}|^{\frac{2\rho -k_{i-1} -r_{i-1}}{r_n+k_n}-1}\right. \nonumber \\&~~~\left. +|\xi _{i-2}|^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\right) | \xi _{i}|^{\frac{r_{i-1}+k_{i-1}}{r_n+k_n}}&\nonumber \\&\quad \le \frac{1}{3}\left( \xi _{i-2}^{\frac{2\rho }{r_n+k_n}} +\xi _{i-1}^{\frac{2\rho }{r_n+k_n}}\right) +\underline{c}_{i}(x_1,\cdots ,x_{i-1})\xi _{i}^{\frac{2\rho }{r_n+k_n}}, \end{aligned}$$
(50)

where \(\underline{c}_{i}>0\) is a function.

Case 2

If \({\frac{r_{i}p_{i-1}}{r_n+k_n}}\ge 1\), applying the Mean Value Theorem yields and with a similar analysis as (49), we have

$$\begin{aligned}&|x_{i}^{p_{i-1}}-x_{i}^{*p_{i-1}}|\nonumber \\&=\left| \left( {x_{i}}^{\frac{r_n+k_n}{r_{i}}}\right) ^{ \frac{r_{i}p_{i-1}}{r_n+k_n}}-\left( {x_{i}^*}^{\frac{r_n+k_n}{r_{i}}} \right) ^{\frac{r_{i}p_{i-1}}{r_n+k_n}}\right| \nonumber \\&\le \hat{c}_i\left| {x_{i}}^{\frac{r_n+k_n}{r_{i}}}-{x_{i}^*}^{ \frac{r_n+k_n}{r_{i}}}\right| \Big |\left( {x_{i}}^{\frac{r_n+k_n}{r_{i}}} \right) ^{\frac{r_{i}p_{i-1}}{r_n+k_n}-1}\nonumber \\&~~~+\left( {x_{i}^*}^{\frac{r_n+k_n}{r_{i}}}\right) ^{\frac{r_{i} p_{i-1}}{r_n+k_n}-1}\Big |\nonumber \\&\le \check{c}_i(x_1,\cdots ,x_{i-1})|\xi _{i}|\left( |\xi _{i-1}|^{ \frac{r_{i}p_{i-1}}{r_n+k_n}-1}+|\xi _{i}|^{\frac{r_{i}p_{i-1}}{r_n+k_n} -1}\right) , \end{aligned}$$
(51)

where \(\hat{c}_i\) is a positive constant and \(\check{c}_i(\cdot )\) is a positive function. With the relation \(r_{i}p_{i-1}=r_{i-1}+k_{i-1}\) in mind, putting (49) with (51) together and applying Lemma 2 yields

$$\begin{aligned}&\left| \left( {x_{i-1}}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\! -\!{x_{i-1}^*}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\!\right) (x_{i}^{p_{i-1}}-x_{i}^{*p_{i-1}})\right| \nonumber \\&\le \check{c}_i(\cdot )\tilde{c}_i(\cdot )|\xi _{i-1}| \left( |\xi _{i-1}|^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\right. \nonumber \\&~~~\left. +|\xi _{i-2}|^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_n+k_n}-1}\right) |\xi _{i}|\Big (|\xi _{i-1}|^{\frac{r_{i}p_{i-1}}{r_n+k_n}-1}\nonumber \\&~~~+|\xi _{i}|^{\frac{r_{i}p_{i-1}}{r_n+k_n}-1}\Big )\nonumber \\&\le \frac{1}{3}\left( \xi _{i-2}^{\frac{2\rho }{r_n+k_n}} +\xi _{i-1}^{\frac{2\rho }{r_n+k_n}}\right) +\bar{c}_{i}(x_1,\cdots ,x_{i-1})\xi _{i}^{\frac{2\rho }{r_n+k_n}}, \end{aligned}$$
(52)

where \(\bar{c}_i(\cdot )>0\) is a function.

In conclusion, combining (50) and (52), we finish the proof of Proposition 1 as

$$\begin{aligned}&\left| \left( {x_{i\!-1}}^{\frac{2\rho -k_{i-1}\!-r_{i-1}}{r_{i-1}}}\!-\!{x_{i-1}^*}^{\frac{2\rho -k_{i-1}-r_{i-1}}{r_{i-1}}}\!\right) \left( x_{i}^{p_{i-1}}\!-\!x_{i}^{*p_{i-1}}\right) \right| \\&\quad \le \frac{1}{3}\left( \xi _{i-2}^{\frac{2\rho }{r_n+k_n}} +\xi _{i-1}^{\frac{2\rho }{r_n+k_n}}\right) +c_{i}(x_1,\cdots ,x_{i-1})\xi _{i}^{\frac{2\rho }{r_n+k_n}} \end{aligned}$$

by choosing \(c_i(\cdot )=\max \{\underline{c}_i(\cdot ),\bar{c}_i(\cdot )\}\).

Proof of Proposition 2

Denote \(\tilde{\beta }_j(x_1,\cdots ,x_j)\) \(= \beta _j^{\frac{r_n+k_n}{r_j+k_j}}\) \((x_1,\ldots ,x_j),~j=1,\ldots ,i-1\). By applying Lemma 1, Assumption 1 can be rewritten as

$$\begin{aligned} |\phi _i(x_1,\cdots ,x_i)|&\le b_i(\cdot )\Big (|\xi _1|^{\frac{r_i+k_i}{r_n+k_n}}+ |\xi _2-\tilde{\beta }_1(\cdot )\xi _1|^{\frac{r_i+k_i}{r_n+k_n}}\nonumber \\&~~~+\cdots +|\xi _i-\tilde{\beta }_{i-1}(\cdot )\xi _{i-1}|^{ \frac{r_i+k_i}{r_n+k_n}}\Big )\nonumber \\&\le \tilde{b}_i(\cdot )\left( |\xi _1|^{\frac{r_i+k_i}{r_n+k_n}}+ \cdots +|\xi _i|^{\frac{r_i+k_i}{r_n+k_n}}\right) \end{aligned}$$
(53)

with a smooth function \(\tilde{b}_i(x_1,\cdots ,x_i)>0\).

With the help of (49) and Lemma , we know

$$\begin{aligned}&\left| \left( ({x_i}^{\frac{r_n+k_n}{r_i}})^{\frac{2\rho -k_i -r_i}{r_n+k_n}}-({x_i^*}^{\frac{r_n+k_n}{r_i}})^{\frac{2\rho -k_i-r_i}{r_n+k_n}}\right) \phi _i(t,x,u)\right| \\&\le \tilde{b}_i(\cdot )\tilde{c}_{i+1}\Big (|\xi _1|^{ \frac{r_i+k_i}{r_n+k_n}}+\cdots \\&~~~+|\xi _i|^{\frac{r_i+k_i}{r_n+k_n}}\Big ) (|\xi _i|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1} +|\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1})|\xi _i|\\&\le \check{b}_i(\cdot )\left( |\xi _1|^{\frac{2\rho }{r_n+k_n}-1} +\cdots +|\xi _i|^{\frac{2\rho }{r_n+k_n}-1}\right) |\xi _i|\\&\le \frac{1}{3}\left( |\xi _1|^{\frac{2\rho }{r_n+k_n}}+ \cdots +|\xi _{i-1}|^{\frac{2\rho }{r_n+k_n}}\right) +\hat{b}_i(\cdot )|\xi _{i}|^{\frac{2\rho }{r_n+k_n}}, \end{aligned}$$

where \(\check{b}_i(x_1,\cdots ,x_i)>0,~\hat{b}_i(x_1,\cdots ,x_i)>0\) are smooth functions.

Proof of Proposition 3

First, we will show that there exists a smooth function \(a_{i,l}(x_1,\cdots ,x_{i-1})>0\), such that the following holds

$$\begin{aligned}&\left| \frac{\partial \xi _{i-1}}{\partial x_l}\dot{x}_l\right| \le a_{i,l}(\cdot )\Big (|\xi _{1}|^{ \frac{r_n+k_n+k_l}{r_n+k_n}}+\cdots \\&\quad +\,|\xi _{i-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\Big ),~l=1,\cdots ,i-1. \end{aligned}$$

The proof is done by utilizing an inductive argument.

Initial step First we consider the case \(i=2\) and \(l=1\). With \(r_n+k_n\ge r_l\) in mind, by (9), (53), and applying Lemma 2, it yields

$$\begin{aligned}&\left| \frac{\partial \xi _2}{\partial x_1}\dot{x}_1\right| \\&=\left| \frac{\partial {\tilde{\beta }_{1}(x_1)}}{\partial x_1}\xi _{1}+\tilde{\beta }_1(x_1)\frac{\partial \xi _{1}}{\partial x_1}\right| \big |x_2^{p_1}+\phi _1(x_1)\big |\\&\quad \le \left( \left| \frac{\partial {\tilde{\beta }_{1}(x_1)}}{\partial x_1}\right| |\xi _{1}|+\tilde{\beta }_1(x_1)\frac{r_n+k_n}{r_1}|\xi _1|^{ \frac{r_n+k_n-r_1}{r_n+k_n}}\right) \\&~~~\check{a}_{2,1}(x_1)\left( |\xi _1|^{\frac{r_1+k_1}{r_n+k_n}} +|\xi _2|^{\frac{r_1+k_1}{r_n+k_n}}\right) \\&\quad \le \bar{a}_{2,1}(x_1)|\xi _1|^{\frac{r_n+k_n-r_1}{r_n+k_n}} \left( |\xi _1|^{\frac{r_1+k_1}{r_n+k_n}}+|\xi _2|^{\frac{r_1+k_1}{r_n+k_n}}\right) \\&~~~+\hat{a}_{2,1}(x_1)|\xi _1|^{\frac{r_n+k_n-r_1}{r_n+k_n}} \left( |\xi _1|^{\frac{r_1+k_1}{r_n+k_n}}+|\xi _2|^{\frac{r_1+k_1}{r_n+k_n}}\right) \\&\quad \le a_{2,1}(x_1)\left( |\xi _1|^{\frac{r_n+k_n+k_1}{r_n+k_n}} +|\xi _2|^{\frac{r_n+k_n+k_1}{r_n+k_n}}\right) \end{aligned}$$

with smooth functions \(\check{a}_{2,1}(\cdot )>0,~\bar{a}_{2,1}(\cdot )>0\), \(\hat{a}_{2,1}(\cdot )>0\), and \(a_{2,1}(\cdot )>0\).

Inductive step Assume in step \(j-1\) that, for \(l=1,\cdots ,j-2\), there exists a smooth function \(a_{j-1,l}\) (\(x_l\), \(\cdots \), \(x_{j-2}\)) \(>0\) such that

$$\begin{aligned}&\left| \frac{\partial \xi _{j-1}}{\partial x_l}\dot{x}_l\right| \nonumber \\&\quad \le a_{j-1,l}(\cdot )\left( |\xi _1|^{\frac{r_n+k_n+k_l}{r_n+k_n}} +\cdots +|\xi _{j-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\right) . \end{aligned}$$
(54)

In what follows, we show (54) will also hold for step \(j\), \(l=1,2,\cdots ,j-1\).

First for \(l=1,2,\cdots , j-2\), from (9) and (53), we have

$$\begin{aligned}&\left| \frac{\partial {\xi _j}}{\partial {x_l}}\dot{x}_l\right| \\&=\left| \frac{\partial \big (x_j^{\frac{r_n+k_n}{r_j}}+\beta _{j-1}^{ \frac{r_n+k_n}{r_jp_{j-1}}}(x_1,x_2,\cdots ,x_{j-1})\xi _{j-1}\big )}{ \partial x_l}\dot{x}_l\right| \\&=\left| \frac{\partial \big (\tilde{\beta }_{j-1}(x_1,\cdots ,x_{j-1}) \xi _{j-1}\big )}{\partial {x_l}}\dot{x}_l\right| \\&\le \left| \frac{\partial \tilde{\beta }_{j-1}(x_1,\cdots ,x_{j-1})}{ \partial x_l}\xi _{j-1}\dot{x}_l\right| \\&~~~+\left| \tilde{\beta }_{j-1}(x_1,\cdots ,x_{j-1})\frac{\partial \xi _{j-1}}{\partial x_l}\dot{x}_l\right| ,\\&\quad ~l=1,2,\cdots ,j-2, \end{aligned}$$

where \(\tilde{\beta }_{j-1}(\cdot )=\beta _{j-1}^{\frac{r_n+k_n}{r_jp_{j-1}}}(\cdot )\). Noting that \(r_n+k_n\ge r_l\), using Lemma 2 together with (54), yields

$$\begin{aligned}&\left| \frac{\partial {\xi _j}}{\partial {x_l}}\dot{x}_l\right| \nonumber \\&\le \left| \frac{\partial \tilde{\beta }_{j-1}(\cdot )}{\partial x_l}\xi _{j-1}\right| \Big (|\xi _{1}|^{\frac{r_{l}+k_l}{r_n+k_n}}+\cdots +|\xi _{l+1}|^{\frac{r_{l}+k_l}{r_n+k_n}}\Big )\nonumber \\&~~~+\hat{a}_{j-1,l}(\cdot )\left( |\xi _1|^{\frac{r_n+k_n+k_l}{r_n+k_n}} +\cdots +|\xi _{j-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\right) \nonumber \\&\le \bar{a}_{j,l}(\cdot )|\xi _{j-1}|^{\frac{r_n+k_n-r_l}{r_n+k_n}} \big (|\xi _{1}|^{\frac{r_{l}+k_l}{r_n+k_n}}+\cdots +|\xi _{l+1}|^{ \frac{r_{l}+k_l}{r_n+k_n}}\big )\nonumber \\&~~~+\hat{a}_{j,l}(\cdot )\left( |\xi _1|^{\frac{r_n+k_n+k_l}{r_n+k_n}} +\cdots +|\xi _{j-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\right) \nonumber \\&\le a_{j,l}(\cdot )\left( |\xi _1|^{\frac{r_n+k_n+k_l}{r_n+k_n}} +\cdots +|\xi _{j-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\right) ,\nonumber \\&~~~~l=1,2,\cdots ,j-2, \end{aligned}$$
(55)

where \(\bar{a}_{j,l}(x_1,\cdots ,x_{j-1})>0,~\hat{a}_{j,l} (x_1,\cdots ,x_{j-1})>0\), and \(a_{j,l}(x_1,\cdots ,x_{j-1})>0\) are smooth functions.

Next we shall prove (55) also holds in the case of \(l=j-1\). Similarly, using Lemma 2 together with (9) and (53), one can also obtain the following

$$\begin{aligned}&\left| \frac{\partial {\xi _j}}{\partial {x_{j-1}}}\dot{x}_{j-1} \right| \\&=\left| \frac{\partial \Big (x_j^{\frac{r_n+k_n}{r_j}} +\beta _{j-1}^{\frac{r_n+k_n}{r_jp_{j-1}}}(\cdot )\big (x_{j-1}^{ \frac{r_n+k_n}{r_{j-1}}}-x_{j-1}^{*\frac{r_n+k_n}{r_{j-1}}}\big ) \Big )}{\partial x_{j-1}}\dot{x}_{j-1}\right| \\&\le \left| \frac{\partial \tilde{\beta }_{j-1}(\cdot )}{\partial x_{j-1}}\xi _{j-1}\dot{x}_{j-1}\right| +\left| \tilde{\beta }_{j-1}(\cdot ) \frac{\partial x_{j-1}^{\frac{r_n+k_n}{r_{j-1}}}}{\partial x_{j-1}} \dot{x}_{j-1}\right| \\&\le \bar{a}_{j,j-1}(\cdot )\big (|\xi _{j-1}|^{\frac{r_n+k_n-r_{j-1}}{r_n +k_n}}+|\xi _{j-2}|^{\frac{r_n+k_n-r_{j-1}}{r_n+k_n}}\big )\\&~~~\big (|\xi _{1}|^{\frac{r_{j-1}+k_{j-1}}{r_n+k_n}}+\cdots +| \xi _{j}|^{\frac{r_{j-1}+k_{j-1}}{r_n+k_n}}\big )\\&\le a_{j,j-1}(\cdot )\left( |\xi _1|^{\frac{r_n+k_n+k_{j-1}}{r_n+k_n}} +\cdots +|\xi _{j}|^{\frac{r_n+k_n+k_{j-1}}{r_n+k_n}}\right) , \end{aligned}$$

where \(\bar{a}_{j,j-1}\) \((x_1,\cdots ,x_{j-1})\) \( >0\), \(\hat{a}_{j,j-1}\) \((x_1,\!\cdots \!,x_{j-1})\) \(>0\), and \(a_{j,j-1}(x_1,\cdots ,x_{j-1})>0\) are smooth functions.

In summary, the following holds for step \(j\) and \(l=1,2,\cdots ,j-1\)

$$\begin{aligned} \left| \frac{\partial \xi _j}{\partial x_l}\dot{x}_l\right|&\le a_{j,l}(x_1,\cdots ,x_{j-1})\Big (|\xi _{1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\nonumber \\&~~~+\cdots +|\xi _{j}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\Big ),~l=1,\cdots ,j-1. \end{aligned}$$
(56)

This completes the inductive proof.

Final step Repeatedly using the inductive arguments, we can conclude that for the final step \(j=i\), there exists a smooth function \(a_{i,l}(x_1,\cdots ,x_{i-1})>0\), such that

$$\begin{aligned} \left| \frac{\partial \xi _i}{\partial x_l}\dot{x}_l\right|&\le a_{i,l}(x_1,\cdots ,x_{i-1})\Big (|\xi _{1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\nonumber \\&~~~+\cdots +|\xi _{i}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\Big ),~l=1,\cdots ,i-1. \end{aligned}$$
(57)

With the help of (57), now we arrive at

$$\begin{aligned}&|x_i+x_i^*||\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1}\left| {\displaystyle \frac{\partial \xi _{i-1}}{\partial x_l}}\dot{x_l}\right| \nonumber \\&=|x_i-x_i^*+2x_i^*||\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1} \left| {\displaystyle \frac{\partial \xi _{i-1}}{\partial x_l}}\dot{ x_l}\right| \nonumber \\&\le a_{i,l}(\cdot )\left( |x_i-x_i^*|+2\beta _{i-1}^{\frac{1}{p_i}} (\cdot )|\xi _{i-1}|^{\frac{r_i}{r_n+k_n}}\right) \nonumber \\&~~~|\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1}\left( |\xi _{1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\!+\!\cdots \!+\!|\xi _{i-1}|^{ \frac{r_n+k_n+k_l}{r_n+k_n}}\right) \nonumber \\&\le \check{a}_{i,l}(\cdot )|\xi _i|^{\frac{r_i}{r_n+k_n}}| \xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1}\Big (| \xi _{1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\nonumber \\&~~~+\cdots +|\xi _{i-1}|^{\frac{r_n+k_n+k_l}{r_n+k_n}}\Big ) \end{aligned}$$
(58)

with a smooth function \(\check{a}_{i,l}(\cdot )>0\).

Under the assumption of \(k_l\ge k_i\) \((\forall l=1,2,\cdots ,i-1)\), by utilizing Lemma 2, we have

$$\begin{aligned}&|x_i+x_i^*||\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1}\left| {\displaystyle \frac{\partial \xi _{i-1}}{\partial x_l}}\dot{x_l}\right| \\&\quad \le \frac{1}{3(i-1)}\left( \xi _1^{\frac{2\rho }{r_n+k_n}}+\cdots +\xi _{i-1}^{\frac{2\rho }{r_n+k_n}}\right) \\&~~~+\tilde{a}_{i,l}(x_1,\cdots ,x_{i})\xi _i^{\frac{2\rho }{r_n+k_n}}, \end{aligned}$$

where \(\tilde{a}_{i,l}(\cdot )>0\) is a smooth function. Finally, we can always find a smooth function \(\bar{b}_i(\cdot )>0\), such that Proposition 3 holds

$$\begin{aligned}&\bar{\beta }(\cdot ) |x_i+x_i^*||\xi _{i-1}|^{\frac{2\rho -k_i-r_i}{r_n+k_n}-1} \sum _{l=1}^{i-1}\left| {\displaystyle \frac{\partial \xi _{i-1}}{\partial x_l}}\dot{x}_l\right| \\&\quad \le \frac{1}{3}\left( \xi _1^{\frac{2\rho }{r_n+k_n}}+\cdots +\xi _{i-1}^{\frac{2\rho }{r_n+k_n}}\right) +\bar{b}_i(x_1,\cdots ,x_i) \xi _i^{\frac{2\rho }{r_n+k_n}}. \end{aligned}$$

Rights and permissions

Reprints and permissions

About this article

Cite this article

Tian, W., Zhang, C., Qian, C. et al. Global stabilization of inherently non-linear systems using continuously differentiable controllers. Nonlinear Dyn 77, 739–752 (2014). https://doi.org/10.1007/s11071-014-1336-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11071-014-1336-y

Keywords

Navigation