30
Bounded Backstepping for Nonlinear Control Systems Frederic Mazenc Michael Malisoff Laurent Burlion 0/6

Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

  • Upload
    vudien

  • View
    252

  • Download
    1

Embed Size (px)

Citation preview

Page 1: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Bounded Backstepping forNonlinear Control Systems

Frederic MazencMichael MalisoffLaurent Burlion

0/6

Page 2: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 3: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 4: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 5: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 6: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 7: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Background on Backstepping

B We study control systems, i.e., systems of ODEs having astate- and time-dependent parameter called a control.

B We choose the control to ensure global asymptotic stability ofthe systems and to depend on current and past state values.

B We require the control to satisfy certain boundednessconditions and certain partially linear cascade structures.

B Many engineering applications yield the dynamic structuresfrom our theorem, resulting in a large literature.

B R. Freeman, H. Khalil, M. Krstic, F. Mazenc, J. Tsinias, ....

1/6

Page 8: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Class of Dynamics and 3 Assumptions

x(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

(1)

Assumption 1: There is a bounded locally Lipschitz ϑ such thatzi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = ϑ(z(t)) +k∑

i=1vizi(t)

(2)

is globally asymptotically and locally exponentially stable to 0.

Assumption 2: The function F in (1) is continuous in t andglobally Lipschitz in (x , z1) and F(t ,0,0) = 0 for all t ≥ 0.

2/6

Page 9: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Class of Dynamics and 3 Assumptionsx(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

(1)

Assumption 1: There is a bounded locally Lipschitz ϑ such thatzi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = ϑ(z(t)) +k∑

i=1vizi(t)

(2)

is globally asymptotically and locally exponentially stable to 0.

Assumption 2: The function F in (1) is continuous in t andglobally Lipschitz in (x , z1) and F(t ,0,0) = 0 for all t ≥ 0.

2/6

Page 10: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Class of Dynamics and 3 Assumptionsx(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

(1)

Assumption 1: There is a bounded locally Lipschitz ϑ such thatzi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = ϑ(z(t)) +k∑

i=1vizi(t)

(2)

is globally asymptotically and locally exponentially stable to 0.

Assumption 2: The function F in (1) is continuous in t andglobally Lipschitz in (x , z1) and F(t ,0,0) = 0 for all t ≥ 0.

2/6

Page 11: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Class of Dynamics and 3 Assumptionsx(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

(1)

Assumption 1: There is a bounded locally Lipschitz ϑ such thatzi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = ϑ(z(t)) +k∑

i=1vizi(t)

(2)

is globally asymptotically and locally exponentially stable to 0.

Assumption 2: The function F in (1) is continuous in t andglobally Lipschitz in (x , z1) and F(t ,0,0) = 0 for all t ≥ 0.

2/6

Page 12: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Class of Dynamics and 3 Assumptionsx(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

(1)

Assumption 3: (CICS) There is a globally Lipschitz bounded ωsuch that ω(0) = 0, and constants T > 0 and q > 0, such that foreach continuous δ : [0,+∞)→ R that exponentially convergesto 0, the following is true: All solutions ξ : [0,+∞)→ Rn of

ξ(t) = F(

t , ξ(t),∫ t

t−T

eq(`−t)Q(t ,`,`+T )ω(ξ(`))∫ 0−T eqr r k−1(r+T )k−1dr

d`+ δ(t))

(C)

satisfy limt→+∞ ξ(t) = 0, where Q(t ,a,b) = (t − a)k−1(t − b)k−1.

2/6

Page 13: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Dynamic Extension

x(t) = F(t , x(t), z1(t))

zi(t) = zi+1(t), i ∈ {1, . . . , k − 1}

zk (t) = u(t) +k∑

j=1vjzj(t)

Y (t) = J2k−1Y (t) +e2k−1

Tω(x(t))

bT, where

(AUG)

J2k−1 =

−q 1 0 . . . 00 −q 1 . . . 0...

. . . . . . . . ....

.... . . −q 1

0 . . . . . . 0 −q

∈ R(2k−1)×(2k−1)

and bT =∫ 0−T eq``k−1(`+ T )k−1d` with q and T as above.

3/6

Page 14: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Our Theorem

If Assumptions 1-3 hold, then we can construct positiveconstants a, b, and c and constant row vectors Ri such that allmaximal solutions (x , z,Y ) of (AUG) with the choices

u(t) = sata (R0Ψ(Yt ))

+bω(x(t))bT+ c ω(x(t−T ))

bT+ ϑ(z?(t)),

Ψ(Yt ) = Y (t)− eTJ2k−1Y (t − T )

z∗(t) = (z1(t) + R1Ψ(Yt ), . . . , zk (t) + Rk Ψ(Yt ))

(2)

satisfy limt→+∞(x , z,Y )(t) = 0. �

Here sata(r) = r if |r | ≤ a and a sign(r) otherwise.

Benefit: Bounded u(t) under mild checkable (CICS) condition.

4/6

Page 15: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Our Theorem

If Assumptions 1-3 hold, then we can construct positiveconstants a, b, and c and constant row vectors Ri such that allmaximal solutions (x , z,Y ) of (AUG) with the choices

u(t) = sata (R0Ψ(Yt ))

+bω(x(t))bT+ c ω(x(t−T ))

bT+ ϑ(z?(t)),

Ψ(Yt ) = Y (t)− eTJ2k−1Y (t − T )

z∗(t) = (z1(t) + R1Ψ(Yt ), . . . , zk (t) + Rk Ψ(Yt ))

(2)

satisfy limt→+∞(x , z,Y )(t) = 0. �

Here sata(r) = r if |r | ≤ a and a sign(r) otherwise.

Benefit: Bounded u(t) under mild checkable (CICS) condition.

4/6

Page 16: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Our Theorem

If Assumptions 1-3 hold, then we can construct positiveconstants a, b, and c and constant row vectors Ri such that allmaximal solutions (x , z,Y ) of (AUG) with the choices

u(t) = sata (R0Ψ(Yt ))

+bω(x(t))bT+ c ω(x(t−T ))

bT+ ϑ(z?(t)),

Ψ(Yt ) = Y (t)− eTJ2k−1Y (t − T )

z∗(t) = (z1(t) + R1Ψ(Yt ), . . . , zk (t) + Rk Ψ(Yt ))

(2)

satisfy limt→+∞(x , z,Y )(t) = 0. �

Here sata(r) = r if |r | ≤ a and a sign(r) otherwise.

Benefit: Bounded u(t) under mild checkable (CICS) condition.

4/6

Page 17: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Sufficient Conditions for (CICS) for Small T

There exist f and g that are uniformly globally Lipschitz in x andcontinuous such that F(t , x ,p) = f (t , x) + g(t , x)p holds for allt ≥ 0, x ∈ Rn, and p ∈ R.

Also, there exist a C1 uniformly properand positive definite function V ; a uniformly continuous positivedefinite function W ; positive constants r0, r1, and r3; and aconstant r2 ≥ 0 such that for all (t , x) ∈ [0,+∞)× Rn, we have

Vt (t , x) + Vx (t , x)(f (t , x) + g(t , x)ω(x)

)≤−W (x),

|Vx (t , x)g(t , x)| ≤ r0√

W (x), |ω(x)| ≤ r1√

W (x),

|f (t , x)| ≤ r2√

W (x), and |g(t , x)| ≤ r3,

(DB)

where ω is bounded, globally Lipschitz, and satisfies ω(0) = 0.

5/6

Page 18: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Sufficient Conditions for (CICS) for Small T

There exist f and g that are uniformly globally Lipschitz in x andcontinuous such that F(t , x ,p) = f (t , x) + g(t , x)p holds for allt ≥ 0, x ∈ Rn, and p ∈ R. Also, there exist a C1 uniformly properand positive definite function V ; a uniformly continuous positivedefinite function W ; positive constants r0, r1, and r3; and aconstant r2 ≥ 0 such that for all (t , x) ∈ [0,+∞)× Rn, we have

Vt (t , x) + Vx (t , x)(f (t , x) + g(t , x)ω(x)

)≤−W (x),

|Vx (t , x)g(t , x)| ≤ r0√

W (x), |ω(x)| ≤ r1√

W (x),

|f (t , x)| ≤ r2√

W (x), and |g(t , x)| ≤ r3,

(DB)

where ω is bounded, globally Lipschitz, and satisfies ω(0) = 0.

5/6

Page 19: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Sufficient Conditions for (CICS) for Small T

There exist f and g that are uniformly globally Lipschitz in x andcontinuous such that F(t , x ,p) = f (t , x) + g(t , x)p holds for allt ≥ 0, x ∈ Rn, and p ∈ R. Also, there exist a C1 uniformly properand positive definite function V ; a uniformly continuous positivedefinite function W ; positive constants r0, r1, and r3; and aconstant r2 ≥ 0 such that for all (t , x) ∈ [0,+∞)× Rn, we have

Vt (t , x) + Vx (t , x)(f (t , x) + g(t , x)ω(x)

)≤−W (x),

|Vx (t , x)g(t , x)| ≤ r0√

W (x), |ω(x)| ≤ r1√

W (x),

|f (t , x)| ≤ r2√

W (x), and |g(t , x)| ≤ r3,

(DB)

where ω is bounded, globally Lipschitz, and satisfies ω(0) = 0.

5/6

Page 20: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Sufficient Conditions for (CICS) for Small T

There exist f and g that are uniformly globally Lipschitz in x andcontinuous such that F(t , x ,p) = f (t , x) + g(t , x)p holds for allt ≥ 0, x ∈ Rn, and p ∈ R. Also, there exist a C1 uniformly properand positive definite function V ; a uniformly continuous positivedefinite function W ; positive constants r0, r1, and r3; and aconstant r2 ≥ 0 such that for all (t , x) ∈ [0,+∞)× Rn, we have

Vt (t , x) + Vx (t , x)(f (t , x) + g(t , x)ω(x)

)≤−W (x),

|Vx (t , x)g(t , x)| ≤ r0√

W (x), |ω(x)| ≤ r1√

W (x),

|f (t , x)| ≤ r2√

W (x), and |g(t , x)| ≤ r3,

(DB)

where ω is bounded, globally Lipschitz, and satisfies ω(0) = 0.

5/6

Page 21: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 22: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 23: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 24: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 25: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 26: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 27: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, and L. Burlion. Bounded backsteppingthrough a dynamic extension with delays. In Proceedings of the56th IEEE Conference on Decision and Control (Melbourne,Australia, 12-15 December 2017).

Thank you for your attention!

6/6

Page 28: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., and M. Malisoff. New control design for boundedbackstepping under input delays. Automatica, 2016.

Thank you for your attention!Thank you for your attention!Thank you for your attention!

6/6

Page 29: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, L. Burlion, and J. Weston. Boundedbackstepping control and robustness analysis for time-varyingsystems under converging-input-converging-state conditions.European Journal of Control, 2017.

Thank you for your attention!

6/6

Page 30: Bounded Backstepping for Nonlinear Control Systemsmalisoff/papers/24Sep... · Bounded Backstepping for Nonlinear Control Systems ... H. Khalil, M. Krstic, F. Mazenc, ... Bounded Backstepping

Conclusions

B Backstepping is an essential technique in control theory.

B It is challenging when there are control bounds.

BWe overcame this under our general (CICS) condition.

B Sufficient conditions tell us how to choose artificial delay T .

BWe can cover F(t , x , z), measurement delay and uncertainty.

Mazenc, F., M. Malisoff, L. Burlion, and J. Weston. Boundedbackstepping control and robustness analysis for time-varyingsystems under converging-input-converging-state conditions.European Journal of Control, 2017.

Thank you for your attention!

6/6