SlideShare a Scribd company logo
1 of 30
Download to read offline
Resilient State Estimation via Input and State
Interval Observer Synthesis
Mohammad Khajenejad, Zeyuan Jin, Thach Ngoc Dinh and
Sze Zheng Yong
Department of Mechanical and Industrial Engineering
Northeastern University, Boston, USA
Email: s.yong@northeastern.edu
62nd IEEE Conference on Decision and Control, Singapore
Dec. 13-15, 2023
Cyber-Physical Systems Under Attack
CPS are subject to attacks, malicious
behaviors, unknown inputs
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 2 / 21
Data Attack Resiliency
Research Question
Can we simultaneously obtain guaranteed estimates of states and
unknown inputs (adversarial signals)?
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 3 / 21
Resilient Observer Design; State and Input Estimation
xk+1 = f (xk ) + Wwk + Gdk ,
yk = h(xk ) + Vvk + H dk
|{z}
unknown input
no prior ‘useful’ knowledge or
assumption or known bounds
on the dynamics of dk
Problem (Simultaneous Input and State Observer)
Design stable and optimal interval-valued input and state estimator
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 4 / 21
Resilient Observer Design; State and Input Estimation
xk+1 = f (xk ) + Wwk + Gdk ,
yk = h(xk ) + Vvk + H dk
|{z}
unknown input
no prior ‘useful’ knowledge or
assumption or known bounds
on the dynamics of dk
Problem (Simultaneous Input and State Observer)
Design stable and optimal interval-valued input and state estimator
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 4 / 21
A Glance at Related Work
Resilient Estimation
Linear Systems
▶ [Yong.ea.2018]: residual-based estimation for stochastic systems
▶ [Nakahira.Mo.2018],[Khajenejad.Yong.2019]: set-membership
resilient estimation with multiple-model approach
Nonlinear Systems
▶ [Kim.ea.2018], [Chong.ea.2020]: only sensors are attacked
▶ [Khajenejad.Yong.2021]: both sensors and actuators are
compromised
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 5 / 21
Preliminaries: Decomposition Functions
Definition (DT Decomposition Functions (Yang.ea.2019))
x+
t = f (xt, wt): a DT system, f : Z → Rn
fd : Z × Z → Rn: a DT-MMDF with respect to f , if
▶ fd (z, z) = f (z)
▶ ẑ ≥ z ⇒ fd (ẑ, z′
) ≥ fd (z, z′
)
▶ ẑ ≥ z ⇒ fd (z′
, ẑ) ≤ fd (z′
, z)
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 6 / 21
Preliminaries: Mixed-Monotone Decompositions
f (x) = Hx
|{z}
linear remainder
+ g(x)
|{z}
JSS mapping
, Hij = J
f
ij ∨ Hij = Jf
ij
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-1
-0.5
0
0.5
1
1.5
2
2.5
3
H⊕
x − H⊖
x ≤ Hx ≤ H⊕
x − H⊖
x
g(xc ) ≤ g(x) ≤ g(xc )
max
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
≤ Hx + g(x)
| {z }
f (x)
≤ min
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
Preliminaries: Mixed-Monotone Decompositions
f (x) = Hx
|{z}
linear remainder
+ g(x)
|{z}
JSS mapping
, Hij = J
f
ij ∨ Hij = Jf
ij
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-1
-0.5
0
0.5
1
1.5
2
2.5
3
H⊕
x − H⊖
x ≤ Hx ≤ H⊕
x − H⊖
x
g(xc ) ≤ g(x) ≤ g(xc )
max
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
≤ Hx + g(x)
| {z }
f (x)
≤ min
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
Preliminaries: Mixed-Monotone Decompositions
f (x) = Hx
|{z}
linear remainder
+ g(x)
|{z}
JSS mapping
, Hij = J
f
ij ∨ Hij = Jf
ij
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-1
-0.5
0
0.5
1
1.5
2
2.5
3
H⊕
x − H⊖
x ≤ Hx ≤ H⊕
x − H⊖
x
g(xc ) ≤ g(x) ≤ g(xc )
max
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
≤ Hx + g(x)
| {z }
f (x)
≤ min
H∈H
{H⊕
x − H⊖
x + g(xc )}
| {z }
fd (x,x)
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
Preliminaries: Embedding Systems
x+
= f (x, w)
| {z }
original n-dimensional system

x+
x+

=

f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)
f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)

| {z }
lifted 2n-dimensional embedding system
Proposition (State Framer Property [Khajenejad.Yong.2021])
xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W.
Proposition (Tight Decompositions [Khajenejad.Yong.2021])
if f (z) = µ(z) + Hz such that µ is JSS, then:
µd,i (z1, z2) = µi (Di
z1 + (In − Di
)z2),
δµ
d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z)
Di = diag(max(sgn(J
µ
i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
Preliminaries: Embedding Systems
x+
= f (x, w)
| {z }
original n-dimensional system

x+
x+

=

f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)
f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)

| {z }
lifted 2n-dimensional embedding system
Proposition (State Framer Property [Khajenejad.Yong.2021])
xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W.
Proposition (Tight Decompositions [Khajenejad.Yong.2021])
if f (z) = µ(z) + Hz such that µ is JSS, then:
µd,i (z1, z2) = µi (Di
z1 + (In − Di
)z2),
δµ
d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z)
Di = diag(max(sgn(J
µ
i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
Preliminaries: Embedding Systems
x+
= f (x, w)
| {z }
original n-dimensional system

x+
x+

=

f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)
f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤)

| {z }
lifted 2n-dimensional embedding system
Proposition (State Framer Property [Khajenejad.Yong.2021])
xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W.
Proposition (Tight Decompositions [Khajenejad.Yong.2021])
if f (z) = µ(z) + Hz such that µ is JSS, then:
µd,i (z1, z2) = µi (Di
z1 + (In − Di
)z2),
δµ
d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z)
Di = diag(max(sgn(J
µ
i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
Preliminaries: Affine Outer Approximation
Proposition (Affine Abstraction [Singh.ea.2018])
f (·) : B ≜ [x, x] ⊂ Rn → Rm, VB: set of
vertices.
(AB, AB, eB, eB, θB) ∈ argmin
θ,A,A,e,e
θ (1)
s.t Axs + e + σ ≤ f (xs) ≤ Axs + e − σ,
(A − A)xs + e − e − 2σ ≤ θ1m, ∀xs ∈ VB,
Then, Ax + e ≤ f (x) ≤ Ax + e, ∀x ∈ B.
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 9 / 21
Resilient Interval Observer Synthesis
G :
(
xk+1 = f (xk ) + Wwk + Gdk ,
yk = h(xk ) + Vvk + Hdk
Problem (Input and State Interval Observer)
Synthesize framers xk , xk , dk , dk such that
states  inputs are framed: xk ≤ xk ≤ xk , dk ≤ dk ≤ dk
framers are uniformly bounded
design is optimized
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 10 / 21
Resilient Interval Observer Synthesis
G :
(
xk+1 = f (xk ) + Wwk + Gdk ,
yk = h(xk ) + Vvk + Hdk
Problem (Input and State Interval Observer)
Synthesize framers xk , xk , dk , dk such that
states  inputs are framed: xk ≤ xk ≤ xk , dk ≤ dk ≤ dk
framers are uniformly bounded
design is optimized
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 10 / 21
Observer Overview
xk+1 = f (xk ) + Gdk + Wwk ,
yk = h(xk ) + Hdk + Vvk
Key Insights:
dk ⇔ d1,k  d2,k ,
yk ⇔ z1,k  z2,k
auxiliary state: γk ≜ Λ(I − NC2)xk
▶ unaffected by dk by choosing
N = G2(C2G2)†
▶ Λ ≜ A−1
g with Ag being
affine approximation slope of
xk + Nψ2(x)
Λ(I − NC2)(f (x) − G1Sh1(x)) =
Ax + ρ(x)
| {z }
mixed-monotone decomposition
L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
Observer Overview
xk+1 = f (xk ) + G1d1,k + G2d2,k + Wwk ,
z1,k = h1(xk ) + Ξd1,k + V1vk
z2,k = C2xk + ψ2(xk )
| {z }
mixed-monotone decomposition
+V2vk
Key Insights:
dk ⇔ d1,k  d2,k ,
yk ⇔ z1,k  z2,k
auxiliary state: γk ≜ Λ(I − NC2)xk
▶ unaffected by dk by choosing
N = G2(C2G2)†
▶ Λ ≜ A−1
g with Ag being
affine approximation slope of
xk + Nψ2(x)
Λ(I − NC2)(f (x) − G1Sh1(x)) =
Ax + ρ(x)
| {z }
mixed-monotone decomposition
L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
Observer Overview
xk+1 = f (xk ) + G1d1,k + G2d2,k + Wwk ,
z1,k = h1(xk ) + Ξd1,k + V1vk
z2,k = C2xk + ψ2(xk )
| {z }
mixed-monotone decomposition
+V2vk
Input
Framers
dk−1,dk−1
Auxiliary
State
Framers
γk
, γk
State
Framers
xk , xk
Start x0, x0
Measurement
yk
k ← k + 1
k = 1
Recursive
algorithm:
Key Insights:
dk ⇔ d1,k  d2,k ,
yk ⇔ z1,k  z2,k
auxiliary state: γk ≜ Λ(I − NC2)xk
▶ unaffected by dk by choosing
N = G2(C2G2)†
▶ Λ ≜ A−1
g with Ag being
affine approximation slope of
xk + Nψ2(x)
Λ(I − NC2)(f (x) − G1Sh1(x)) =
Ax + ρ(x)
| {z }
mixed-monotone decomposition
L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
3-Step Recursive Observer
assumption: C2G2 has a full column rank
d1,k = S(z1,k − h1(xk ) − V1vk ), S = Ξ−1
d2,k = (C2G2)†C2(xk+1 − f (xk ) + G1Sh1(xk ) + G1S(V1vk − z1,k ) − Wwk )
dk = E1d1,k + E2d2,k
⇓
Step 1: Input Framer Computation
dk−1 = Φ⊕xk − Φ⊖xk + κd (xk−1, xk−1) + Az z1,k−1
+A⊕
v v − A⊖
v v + Φ⊖w − Φ⊕w,
dk−1 = Φ⊕xk − Φ⊖xk + κd (xk−1, xk−1) + Az z1,k−1
+A⊕
v v − A⊖
v v + Φ⊖w − Φ⊕w,
Φ ≜ E2(C2G2)†C2, Av ≜ (ΦG1 − E1)SV1, Az ≜ (E1 − ΦG1)S
κ(x) ≜ (ΦG1 − E1)Sh1(x) − Φf (x)
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 12 / 21
3-Step Recursive Observer



























γk ≜ Λ(I − NC2)xk = xk − Λ(N(z2,k − V2vk ) − ϵk )
N = G2(C2G2)† → cancel out the effect of dk
L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0
˜
f (x) ≜ Λ(I − NC2)(f (x) − G1Sh1(x)) = Ax + ρ(x)
| {z }
mixed-monotone decomposition
⇒
γk+1 = (A − LC2)γk + ρ(xk ) − Lψ2(xk ) − V̂ vk + Ŵ wk − Dϵk + ẑk
⇓
Step 2. Auxiliary State Propagation
γk+1
= (A − LC2)⊕γk
− (A − LC2)⊖γk + ρd (xk , xk )
+D⊖ϵ − D⊕ϵ + L⊖ψ2,d (xk , xk ) − L⊕ψ2,d (xk , xk )
+V̂ ⊖v − V̂ ⊕v + Ŵ ⊖w −Ŵ ⊖w + ẑk ,
γk+1 = (A − LC2)⊕γk − (A − LC2)⊖γk
+ ρd (xk , xk )
+D⊖ϵ − D⊕ϵ + L⊖ψ2,d (xk , xk ) − L⊕ψ2,d (xk , xk )
+V̂ ⊖v − V̂ ⊕v + Ŵ ⊕w − Ŵ ⊖w + ẑk ,
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 13 / 21
3-Step Recursive Observer
xk = γk + Λ(N(z2,k − V2vk ) − ϵk )
⇓
Step 3. State Framer Computation
xk = γk
+ ΛNz2,k + Λ⊖ϵ − Λ⊕ϵ + (ΛNV2)⊖v − (ΛNV2)⊕v,
xk = γk + ΛNz2,k + Λ⊖ϵ − Λ⊕ϵ + (ΛNV2)⊖v − (ΛNV2)⊕v,
by construction and properties of decomposition functions:
(
xk ≤ xk ≤ xk
dk ≤ dk ≤ dk
, for all L.
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 14 / 21
H∞-Optimal State and Input Observer Design
to design L to satisfy stability and optimality
Error Dynamics
ex
k+1 ≜ xk − xk = ≤ (|A − LC2| + Fρ + |L|Fψ2
)ex
k + |Ŵ |δw
+(|Va − LVb| − |A − LC2||ΛNV2| + |ΛNV2|)δv
+(|Λ| + |Da − LDb| − |A − LC2||Λ|)δϵ,
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 15 / 21
H∞-Optimal State and Input Observer Design
Theorem (ISS  H∞-Observer Design)
f , h have bounded Jacobians, C2G2 has full column rank
if ψ2(x) ̸= 0 then Ag is invertible
(P∗ ≻ 0, η∗  0, Γ∗ ≥ 0) is a solution to
min
{η,P,Γ}
η
s.t.




P PÃ − ΓC̃ PB̃ − ΓD̃ 0
∗ P 0 I
∗ ∗ ηI 0
∗ ∗ ∗ ηI



≻0, (P, Γ) ∈ C, −P ∈ Mn
then the observer is ISS and optimal with L = (P∗)−1Γ∗
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 16 / 21
H∞-Optimal State and Input Observer Design
Theorem (ISS  H∞-Observer Design)
f , h have bounded Jacobians, C2G2 has full column rank
if ψ2(x) ̸= 0 then Ag is invertible
(P∗ ≻ 0, η∗  0, Γ∗ ≥ 0) is a solution to
min
{η,P,Γ}
η
s.t.




P PÃ − ΓC̃ PB̃ − ΓD̃ 0
∗ P 0 I
∗ ∗ ηI 0
∗ ∗ ∗ ηI



≻0, (P, Γ) ∈ C, −P ∈ Mn
then the observer is ISS and optimal with L = (P∗)−1Γ∗
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 16 / 21
Different Choices for the Set C
(i) C={(P, Γ) | P

A Va Da

−Γ

C2 Vb Db

≥ 0}, if:
à = A + Fρ, C̃ = C2 − Fψ2
,
B̃ =

Va + (I − A)|ΛNV2| |Ŵ | Da + (I − A)|Λ|

,
D̃ =

Vb − C2|ΛNV2| 0 Db − C2|Λ|

.
(ii) C = {(P, Γ) | Γ

C2 Vb Db

≥ 0}, if
à = |A| + Fρ, C̃ = −C2 − Fψ2
,
B̃ =

|Va|+(I−|A|)|ΛNV2| |Ŵ | (I−|A|)|Λ|+|Da|

,
D̃ =

C2|ΛNV2| − Vb 0 C2|Λ| − Db

.
(iii) C = {(P, Γ) | PA − ΓC2 ≥ 0}, if:
à = A + Fρ, C̃ = C2 − Fψ2
, D̃ =

−V2 0 0

,
B̃ =

|Λ(I−NC2)G1SV1|+|ΛNV2| |Ŵ | |Λ|

.
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 17 / 21
Simulation Results: A Three-Area Power Station
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 18 / 21
Takeaway
leveraged mixed-monotone decompositions and affine
outer-approximations for resilient estimation
designed state and input interval observers
the design is correct-by-construction
derived sufficient conditions for stability and optimality
future work:
▶ alternative designs for minimizing L1 gain
▶ continuous-time and hybrid systems
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 19 / 21
References
L. Yang, O. Mickelin, and N. Ozay. “On sufficient conditions for mixed monotonicity”. IEEE Transactions
on Automatic Control, 64(12):5080–5085, 2019
M. Khajenejad, Z. Jin, and S.Z. Yong. “Interval observers for simultaneous state and model estimation of
partially known nonlinear systems”. In American Control Conference, pages 2848–2854. IEEE, 2021.
M. Khajenejad and S. Z. Yong. “Simultaneous mode, input and state set-valued observers with
applications to resilient estimation against sparse attacks”. In IEEE Conference on Decision and Control
(CDC), pages 1544–1550, 2019.
S.Z. Yong, M. Zhu, and E. Frazzoli. “Switching and data injection attacks on stochastic cyber-physical
systems: Modeling, resilient esti- mation, and attack mitigation”. ACM Transactions on Cyber-Physical
Systems, 2(2):9, 2018.
Y. Bar-Shalom, T. Kirubarajan, and X.R. Li. “Estimation with Applications to Tracking and Navigation”.
John Wiley  Sons, Inc., New York, NY, USA, 2002.
Y. Nakahira and Y. Mo. “Attack-resilient H2 , H∞ , and ℓ11 state estimator”. IEEE Transactions on
Automatic Control, 63(12):4353–4360, 2018.
J. Kim, C. Lee, H. Shim, Y. Eun, and J.H. Seo. “Detection of sensor attack and resilient state estimation
for uniformly observable nonlinear systems having redundant sensors”. IEEE Transactions on Automatic
Control, 64(3):1162–1169, 2018.
M.S. Chong, H. Sandberg, and J.P. Hespanha. “A secure state estimation algorithm for nonlinear systems
under sensor attacks”. In IEEE Conference on Decision and Control, pages 5743–5748, 2020.
M. Milanese and C. Novara. “Set membership identification of nonlinear systems”. Automatica,
40:957–975, 2004.
Z.B. Zabinsky, R.L. Smith, and B.P. Kristinsdottir. “Optimal estimation of univariate black-box Lipschitz
functions with upper and lower error bounds”. Computers  Operations Res., 30(10):1539–1553, 2003.
Z. Jin, M. Khajenejad, and S.Z. Yong. “Data-driven model invalidation for unknown Lipschitz continuous
systems via abstraction”. In American Control Conference (ACC), pages 2975–2980. IEEE, 2020.
K.R. Singh, Q. Shen, and S.Z. Yong. “Mesh-based affine abstraction of nonlinear systems with tighter
bounds”. In IEEE Conference on Decision and Control (CDC), pages 3056–3061, 2018.
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 20 / 21
Thank you!
Questions?
M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 21 / 21

More Related Content

Similar to Slides_Resilient_State_Estimation_CDC23.pdf

Modeling the Dynamics of SGD by Stochastic Differential Equation
Modeling the Dynamics of SGD by Stochastic Differential EquationModeling the Dynamics of SGD by Stochastic Differential Equation
Modeling the Dynamics of SGD by Stochastic Differential EquationMark Chang
 
Neutrosophic Soft Topological Spaces on New Operations
Neutrosophic Soft Topological Spaces on New OperationsNeutrosophic Soft Topological Spaces on New Operations
Neutrosophic Soft Topological Spaces on New OperationsIJSRED
 
1.1_The_Definite_Integral.pdf odjoqwddoio
1.1_The_Definite_Integral.pdf odjoqwddoio1.1_The_Definite_Integral.pdf odjoqwddoio
1.1_The_Definite_Integral.pdf odjoqwddoioNoorYassinHJamel
 
Flexural analysis of thick beams using single
Flexural analysis of thick beams using singleFlexural analysis of thick beams using single
Flexural analysis of thick beams using singleiaemedu
 
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Frank Nielsen
 
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...Gabriel Peyré
 
QMC Error SAMSI Tutorial Aug 2017
QMC Error SAMSI Tutorial Aug 2017QMC Error SAMSI Tutorial Aug 2017
QMC Error SAMSI Tutorial Aug 2017Fred J. Hickernell
 
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...Tomoya Murata
 
Point Collocation Method used in the solving of Differential Equations, parti...
Point Collocation Method used in the solving of Differential Equations, parti...Point Collocation Method used in the solving of Differential Equations, parti...
Point Collocation Method used in the solving of Differential Equations, parti...Suddhasheel GHOSH, PhD
 
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flop
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flopOn Optimization of Manufacturing of a Sense-amplifier Based Flip-flop
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flopBRNSS Publication Hub
 
離散値ベクトル再構成手法とその通信応用
離散値ベクトル再構成手法とその通信応用離散値ベクトル再構成手法とその通信応用
離散値ベクトル再構成手法とその通信応用Ryo Hayakawa
 
slides_nuclear_norm_regularization_david_mateos
slides_nuclear_norm_regularization_david_mateosslides_nuclear_norm_regularization_david_mateos
slides_nuclear_norm_regularization_david_mateosDavid Mateos
 
Low Complexity Regularization of Inverse Problems
Low Complexity Regularization of Inverse ProblemsLow Complexity Regularization of Inverse Problems
Low Complexity Regularization of Inverse ProblemsGabriel Peyré
 

Similar to Slides_Resilient_State_Estimation_CDC23.pdf (20)

Modeling the Dynamics of SGD by Stochastic Differential Equation
Modeling the Dynamics of SGD by Stochastic Differential EquationModeling the Dynamics of SGD by Stochastic Differential Equation
Modeling the Dynamics of SGD by Stochastic Differential Equation
 
Ece3075 a 8
Ece3075 a 8Ece3075 a 8
Ece3075 a 8
 
Neutrosophic Soft Topological Spaces on New Operations
Neutrosophic Soft Topological Spaces on New OperationsNeutrosophic Soft Topological Spaces on New Operations
Neutrosophic Soft Topological Spaces on New Operations
 
WITMSE 2013
WITMSE 2013WITMSE 2013
WITMSE 2013
 
One sided z transform
One sided z transformOne sided z transform
One sided z transform
 
1.1_The_Definite_Integral.pdf odjoqwddoio
1.1_The_Definite_Integral.pdf odjoqwddoio1.1_The_Definite_Integral.pdf odjoqwddoio
1.1_The_Definite_Integral.pdf odjoqwddoio
 
main
mainmain
main
 
Flexural analysis of thick beams using single
Flexural analysis of thick beams using singleFlexural analysis of thick beams using single
Flexural analysis of thick beams using single
 
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)Computational Information Geometry on Matrix Manifolds (ICTP 2013)
Computational Information Geometry on Matrix Manifolds (ICTP 2013)
 
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...
Low Complexity Regularization of Inverse Problems - Course #2 Recovery Guaran...
 
QMC Error SAMSI Tutorial Aug 2017
QMC Error SAMSI Tutorial Aug 2017QMC Error SAMSI Tutorial Aug 2017
QMC Error SAMSI Tutorial Aug 2017
 
41-50
41-5041-50
41-50
 
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...
Doubly Accelerated Stochastic Variance Reduced Gradient Methods for Regulariz...
 
Point Collocation Method used in the solving of Differential Equations, parti...
Point Collocation Method used in the solving of Differential Equations, parti...Point Collocation Method used in the solving of Differential Equations, parti...
Point Collocation Method used in the solving of Differential Equations, parti...
 
CLIM Fall 2017 Course: Statistics for Climate Research, Statistics of Climate...
CLIM Fall 2017 Course: Statistics for Climate Research, Statistics of Climate...CLIM Fall 2017 Course: Statistics for Climate Research, Statistics of Climate...
CLIM Fall 2017 Course: Statistics for Climate Research, Statistics of Climate...
 
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flop
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flopOn Optimization of Manufacturing of a Sense-amplifier Based Flip-flop
On Optimization of Manufacturing of a Sense-amplifier Based Flip-flop
 
離散値ベクトル再構成手法とその通信応用
離散値ベクトル再構成手法とその通信応用離散値ベクトル再構成手法とその通信応用
離散値ベクトル再構成手法とその通信応用
 
slides_nuclear_norm_regularization_david_mateos
slides_nuclear_norm_regularization_david_mateosslides_nuclear_norm_regularization_david_mateos
slides_nuclear_norm_regularization_david_mateos
 
Low Complexity Regularization of Inverse Problems
Low Complexity Regularization of Inverse ProblemsLow Complexity Regularization of Inverse Problems
Low Complexity Regularization of Inverse Problems
 
Krishna
KrishnaKrishna
Krishna
 

Recently uploaded

SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )Tsuyoshi Horigome
 
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Dr.Costas Sachpazis
 
Introduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxIntroduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxupamatechverse
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSKurinjimalarL3
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)Suman Mia
 
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...srsj9000
 
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxthe ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxhumanexperienceaaa
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learningmisbanausheenparvam
 
Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxpurnimasatapathy1234
 
Porous Ceramics seminar and technical writing
Porous Ceramics seminar and technical writingPorous Ceramics seminar and technical writing
Porous Ceramics seminar and technical writingrakeshbaidya232001
 
Introduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxIntroduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxupamatechverse
 
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...ranjana rawat
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur High Profile
 
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escortsranjana rawat
 
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Serviceranjana rawat
 
Extrusion Processes and Their Limitations
Extrusion Processes and Their LimitationsExtrusion Processes and Their Limitations
Extrusion Processes and Their Limitations120cr0395
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidNikhilNagaraju
 

Recently uploaded (20)

SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )SPICE PARK APR2024 ( 6,793 SPICE Models )
SPICE PARK APR2024 ( 6,793 SPICE Models )
 
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
 
Introduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptxIntroduction and different types of Ethernet.pptx
Introduction and different types of Ethernet.pptx
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
 
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
Gfe Mayur Vihar Call Girls Service WhatsApp -> 9999965857 Available 24x7 ^ De...
 
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
 
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptxthe ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
the ladakh protest in leh ladakh 2024 sonam wangchuk.pptx
 
chaitra-1.pptx fake news detection using machine learning
chaitra-1.pptx  fake news detection using machine learningchaitra-1.pptx  fake news detection using machine learning
chaitra-1.pptx fake news detection using machine learning
 
Microscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptxMicroscopic Analysis of Ceramic Materials.pptx
Microscopic Analysis of Ceramic Materials.pptx
 
Porous Ceramics seminar and technical writing
Porous Ceramics seminar and technical writingPorous Ceramics seminar and technical writing
Porous Ceramics seminar and technical writing
 
Introduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxIntroduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptx
 
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
(SHREYA) Chakan Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Esc...
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
 
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
(MEERA) Dapodi Call Girls Just Call 7001035870 [ Cash on Delivery ] Pune Escorts
 
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
 
Extrusion Processes and Their Limitations
Extrusion Processes and Their LimitationsExtrusion Processes and Their Limitations
Extrusion Processes and Their Limitations
 
main PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfidmain PPT.pptx of girls hostel security using rfid
main PPT.pptx of girls hostel security using rfid
 
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
 

Slides_Resilient_State_Estimation_CDC23.pdf

  • 1. Resilient State Estimation via Input and State Interval Observer Synthesis Mohammad Khajenejad, Zeyuan Jin, Thach Ngoc Dinh and Sze Zheng Yong Department of Mechanical and Industrial Engineering Northeastern University, Boston, USA Email: s.yong@northeastern.edu 62nd IEEE Conference on Decision and Control, Singapore Dec. 13-15, 2023
  • 2. Cyber-Physical Systems Under Attack CPS are subject to attacks, malicious behaviors, unknown inputs M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 2 / 21
  • 3. Data Attack Resiliency Research Question Can we simultaneously obtain guaranteed estimates of states and unknown inputs (adversarial signals)? M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 3 / 21
  • 4. Resilient Observer Design; State and Input Estimation xk+1 = f (xk ) + Wwk + Gdk , yk = h(xk ) + Vvk + H dk |{z} unknown input no prior ‘useful’ knowledge or assumption or known bounds on the dynamics of dk Problem (Simultaneous Input and State Observer) Design stable and optimal interval-valued input and state estimator M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 4 / 21
  • 5. Resilient Observer Design; State and Input Estimation xk+1 = f (xk ) + Wwk + Gdk , yk = h(xk ) + Vvk + H dk |{z} unknown input no prior ‘useful’ knowledge or assumption or known bounds on the dynamics of dk Problem (Simultaneous Input and State Observer) Design stable and optimal interval-valued input and state estimator M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 4 / 21
  • 6. A Glance at Related Work Resilient Estimation Linear Systems ▶ [Yong.ea.2018]: residual-based estimation for stochastic systems ▶ [Nakahira.Mo.2018],[Khajenejad.Yong.2019]: set-membership resilient estimation with multiple-model approach Nonlinear Systems ▶ [Kim.ea.2018], [Chong.ea.2020]: only sensors are attacked ▶ [Khajenejad.Yong.2021]: both sensors and actuators are compromised M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 5 / 21
  • 7. Preliminaries: Decomposition Functions Definition (DT Decomposition Functions (Yang.ea.2019)) x+ t = f (xt, wt): a DT system, f : Z → Rn fd : Z × Z → Rn: a DT-MMDF with respect to f , if ▶ fd (z, z) = f (z) ▶ ẑ ≥ z ⇒ fd (ẑ, z′ ) ≥ fd (z, z′ ) ▶ ẑ ≥ z ⇒ fd (z′ , ẑ) ≤ fd (z′ , z) M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 6 / 21
  • 8. Preliminaries: Mixed-Monotone Decompositions f (x) = Hx |{z} linear remainder + g(x) |{z} JSS mapping , Hij = J f ij ∨ Hij = Jf ij -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -1 -0.5 0 0.5 1 1.5 2 2.5 3 H⊕ x − H⊖ x ≤ Hx ≤ H⊕ x − H⊖ x g(xc ) ≤ g(x) ≤ g(xc ) max H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) ≤ Hx + g(x) | {z } f (x) ≤ min H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
  • 9. Preliminaries: Mixed-Monotone Decompositions f (x) = Hx |{z} linear remainder + g(x) |{z} JSS mapping , Hij = J f ij ∨ Hij = Jf ij -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -1 -0.5 0 0.5 1 1.5 2 2.5 3 H⊕ x − H⊖ x ≤ Hx ≤ H⊕ x − H⊖ x g(xc ) ≤ g(x) ≤ g(xc ) max H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) ≤ Hx + g(x) | {z } f (x) ≤ min H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
  • 10. Preliminaries: Mixed-Monotone Decompositions f (x) = Hx |{z} linear remainder + g(x) |{z} JSS mapping , Hij = J f ij ∨ Hij = Jf ij -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -1 -0.5 0 0.5 1 1.5 2 2.5 3 H⊕ x − H⊖ x ≤ Hx ≤ H⊕ x − H⊖ x g(xc ) ≤ g(x) ≤ g(xc ) max H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) ≤ Hx + g(x) | {z } f (x) ≤ min H∈H {H⊕ x − H⊖ x + g(xc )} | {z } fd (x,x) M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 7 / 21
  • 11. Preliminaries: Embedding Systems x+ = f (x, w) | {z } original n-dimensional system x+ x+ = f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) | {z } lifted 2n-dimensional embedding system Proposition (State Framer Property [Khajenejad.Yong.2021]) xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W. Proposition (Tight Decompositions [Khajenejad.Yong.2021]) if f (z) = µ(z) + Hz such that µ is JSS, then: µd,i (z1, z2) = µi (Di z1 + (In − Di )z2), δµ d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z) Di = diag(max(sgn(J µ i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
  • 12. Preliminaries: Embedding Systems x+ = f (x, w) | {z } original n-dimensional system x+ x+ = f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) | {z } lifted 2n-dimensional embedding system Proposition (State Framer Property [Khajenejad.Yong.2021]) xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W. Proposition (Tight Decompositions [Khajenejad.Yong.2021]) if f (z) = µ(z) + Hz such that µ is JSS, then: µd,i (z1, z2) = µi (Di z1 + (In − Di )z2), δµ d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z) Di = diag(max(sgn(J µ i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
  • 13. Preliminaries: Embedding Systems x+ = f (x, w) | {z } original n-dimensional system x+ x+ = f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) f d ([x⊤ w⊤]⊤, [x⊤ w⊤]⊤) | {z } lifted 2n-dimensional embedding system Proposition (State Framer Property [Khajenejad.Yong.2021]) xt ≤ xt ≤ xt, ∀t ≥ 0, ∀w ∈ W. Proposition (Tight Decompositions [Khajenejad.Yong.2021]) if f (z) = µ(z) + Hz such that µ is JSS, then: µd,i (z1, z2) = µi (Di z1 + (In − Di )z2), δµ d ≜ µd (z, z) − µd (z, z) ≤ Fµ(z − z) Di = diag(max(sgn(J µ i ), 01,nz )), Fµ ≜ 2 max(Jf − H, 0p,nz )−Jf +H M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 8 / 21
  • 14. Preliminaries: Affine Outer Approximation Proposition (Affine Abstraction [Singh.ea.2018]) f (·) : B ≜ [x, x] ⊂ Rn → Rm, VB: set of vertices. (AB, AB, eB, eB, θB) ∈ argmin θ,A,A,e,e θ (1) s.t Axs + e + σ ≤ f (xs) ≤ Axs + e − σ, (A − A)xs + e − e − 2σ ≤ θ1m, ∀xs ∈ VB, Then, Ax + e ≤ f (x) ≤ Ax + e, ∀x ∈ B. M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 9 / 21
  • 15. Resilient Interval Observer Synthesis G : ( xk+1 = f (xk ) + Wwk + Gdk , yk = h(xk ) + Vvk + Hdk Problem (Input and State Interval Observer) Synthesize framers xk , xk , dk , dk such that states inputs are framed: xk ≤ xk ≤ xk , dk ≤ dk ≤ dk framers are uniformly bounded design is optimized M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 10 / 21
  • 16. Resilient Interval Observer Synthesis G : ( xk+1 = f (xk ) + Wwk + Gdk , yk = h(xk ) + Vvk + Hdk Problem (Input and State Interval Observer) Synthesize framers xk , xk , dk , dk such that states inputs are framed: xk ≤ xk ≤ xk , dk ≤ dk ≤ dk framers are uniformly bounded design is optimized M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 10 / 21
  • 17. Observer Overview xk+1 = f (xk ) + Gdk + Wwk , yk = h(xk ) + Hdk + Vvk Key Insights: dk ⇔ d1,k d2,k , yk ⇔ z1,k z2,k auxiliary state: γk ≜ Λ(I − NC2)xk ▶ unaffected by dk by choosing N = G2(C2G2)† ▶ Λ ≜ A−1 g with Ag being affine approximation slope of xk + Nψ2(x) Λ(I − NC2)(f (x) − G1Sh1(x)) = Ax + ρ(x) | {z } mixed-monotone decomposition L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0 M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
  • 18. Observer Overview xk+1 = f (xk ) + G1d1,k + G2d2,k + Wwk , z1,k = h1(xk ) + Ξd1,k + V1vk z2,k = C2xk + ψ2(xk ) | {z } mixed-monotone decomposition +V2vk Key Insights: dk ⇔ d1,k d2,k , yk ⇔ z1,k z2,k auxiliary state: γk ≜ Λ(I − NC2)xk ▶ unaffected by dk by choosing N = G2(C2G2)† ▶ Λ ≜ A−1 g with Ag being affine approximation slope of xk + Nψ2(x) Λ(I − NC2)(f (x) − G1Sh1(x)) = Ax + ρ(x) | {z } mixed-monotone decomposition L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0 M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
  • 19. Observer Overview xk+1 = f (xk ) + G1d1,k + G2d2,k + Wwk , z1,k = h1(xk ) + Ξd1,k + V1vk z2,k = C2xk + ψ2(xk ) | {z } mixed-monotone decomposition +V2vk Input Framers dk−1,dk−1 Auxiliary State Framers γk , γk State Framers xk , xk Start x0, x0 Measurement yk k ← k + 1 k = 1 Recursive algorithm: Key Insights: dk ⇔ d1,k d2,k , yk ⇔ z1,k z2,k auxiliary state: γk ≜ Λ(I − NC2)xk ▶ unaffected by dk by choosing N = G2(C2G2)† ▶ Λ ≜ A−1 g with Ag being affine approximation slope of xk + Nψ2(x) Λ(I − NC2)(f (x) − G1Sh1(x)) = Ax + ρ(x) | {z } mixed-monotone decomposition L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0 M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 11 / 21
  • 20. 3-Step Recursive Observer assumption: C2G2 has a full column rank d1,k = S(z1,k − h1(xk ) − V1vk ), S = Ξ−1 d2,k = (C2G2)†C2(xk+1 − f (xk ) + G1Sh1(xk ) + G1S(V1vk − z1,k ) − Wwk ) dk = E1d1,k + E2d2,k ⇓ Step 1: Input Framer Computation dk−1 = Φ⊕xk − Φ⊖xk + κd (xk−1, xk−1) + Az z1,k−1 +A⊕ v v − A⊖ v v + Φ⊖w − Φ⊕w, dk−1 = Φ⊕xk − Φ⊖xk + κd (xk−1, xk−1) + Az z1,k−1 +A⊕ v v − A⊖ v v + Φ⊖w − Φ⊕w, Φ ≜ E2(C2G2)†C2, Av ≜ (ΦG1 − E1)SV1, Az ≜ (E1 − ΦG1)S κ(x) ≜ (ΦG1 − E1)Sh1(x) − Φf (x) M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 12 / 21
  • 21. 3-Step Recursive Observer                            γk ≜ Λ(I − NC2)xk = xk − Λ(N(z2,k − V2vk ) − ϵk ) N = G2(C2G2)† → cancel out the effect of dk L(z2,k − C2xk − ψ2(xk ) − V2vk ) = 0 ˜ f (x) ≜ Λ(I − NC2)(f (x) − G1Sh1(x)) = Ax + ρ(x) | {z } mixed-monotone decomposition ⇒ γk+1 = (A − LC2)γk + ρ(xk ) − Lψ2(xk ) − V̂ vk + Ŵ wk − Dϵk + ẑk ⇓ Step 2. Auxiliary State Propagation γk+1 = (A − LC2)⊕γk − (A − LC2)⊖γk + ρd (xk , xk ) +D⊖ϵ − D⊕ϵ + L⊖ψ2,d (xk , xk ) − L⊕ψ2,d (xk , xk ) +V̂ ⊖v − V̂ ⊕v + Ŵ ⊖w −Ŵ ⊖w + ẑk , γk+1 = (A − LC2)⊕γk − (A − LC2)⊖γk + ρd (xk , xk ) +D⊖ϵ − D⊕ϵ + L⊖ψ2,d (xk , xk ) − L⊕ψ2,d (xk , xk ) +V̂ ⊖v − V̂ ⊕v + Ŵ ⊕w − Ŵ ⊖w + ẑk , M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 13 / 21
  • 22. 3-Step Recursive Observer xk = γk + Λ(N(z2,k − V2vk ) − ϵk ) ⇓ Step 3. State Framer Computation xk = γk + ΛNz2,k + Λ⊖ϵ − Λ⊕ϵ + (ΛNV2)⊖v − (ΛNV2)⊕v, xk = γk + ΛNz2,k + Λ⊖ϵ − Λ⊕ϵ + (ΛNV2)⊖v − (ΛNV2)⊕v, by construction and properties of decomposition functions: ( xk ≤ xk ≤ xk dk ≤ dk ≤ dk , for all L. M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 14 / 21
  • 23. H∞-Optimal State and Input Observer Design to design L to satisfy stability and optimality Error Dynamics ex k+1 ≜ xk − xk = ≤ (|A − LC2| + Fρ + |L|Fψ2 )ex k + |Ŵ |δw +(|Va − LVb| − |A − LC2||ΛNV2| + |ΛNV2|)δv +(|Λ| + |Da − LDb| − |A − LC2||Λ|)δϵ, M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 15 / 21
  • 24. H∞-Optimal State and Input Observer Design Theorem (ISS H∞-Observer Design) f , h have bounded Jacobians, C2G2 has full column rank if ψ2(x) ̸= 0 then Ag is invertible (P∗ ≻ 0, η∗ 0, Γ∗ ≥ 0) is a solution to min {η,P,Γ} η s.t.     P PÃ − ΓC̃ PB̃ − ΓD̃ 0 ∗ P 0 I ∗ ∗ ηI 0 ∗ ∗ ∗ ηI    ≻0, (P, Γ) ∈ C, −P ∈ Mn then the observer is ISS and optimal with L = (P∗)−1Γ∗ M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 16 / 21
  • 25. H∞-Optimal State and Input Observer Design Theorem (ISS H∞-Observer Design) f , h have bounded Jacobians, C2G2 has full column rank if ψ2(x) ̸= 0 then Ag is invertible (P∗ ≻ 0, η∗ 0, Γ∗ ≥ 0) is a solution to min {η,P,Γ} η s.t.     P PÃ − ΓC̃ PB̃ − ΓD̃ 0 ∗ P 0 I ∗ ∗ ηI 0 ∗ ∗ ∗ ηI    ≻0, (P, Γ) ∈ C, −P ∈ Mn then the observer is ISS and optimal with L = (P∗)−1Γ∗ M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 16 / 21
  • 26. Different Choices for the Set C (i) C={(P, Γ) | P A Va Da −Γ C2 Vb Db ≥ 0}, if: à = A + Fρ, C̃ = C2 − Fψ2 , B̃ = Va + (I − A)|ΛNV2| |Ŵ | Da + (I − A)|Λ| , D̃ = Vb − C2|ΛNV2| 0 Db − C2|Λ| . (ii) C = {(P, Γ) | Γ C2 Vb Db ≥ 0}, if à = |A| + Fρ, C̃ = −C2 − Fψ2 , B̃ = |Va|+(I−|A|)|ΛNV2| |Ŵ | (I−|A|)|Λ|+|Da| , D̃ = C2|ΛNV2| − Vb 0 C2|Λ| − Db . (iii) C = {(P, Γ) | PA − ΓC2 ≥ 0}, if: à = A + Fρ, C̃ = C2 − Fψ2 , D̃ = −V2 0 0 , B̃ = |Λ(I−NC2)G1SV1|+|ΛNV2| |Ŵ | |Λ| . M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 17 / 21
  • 27. Simulation Results: A Three-Area Power Station M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 18 / 21
  • 28. Takeaway leveraged mixed-monotone decompositions and affine outer-approximations for resilient estimation designed state and input interval observers the design is correct-by-construction derived sufficient conditions for stability and optimality future work: ▶ alternative designs for minimizing L1 gain ▶ continuous-time and hybrid systems M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 19 / 21
  • 29. References L. Yang, O. Mickelin, and N. Ozay. “On sufficient conditions for mixed monotonicity”. IEEE Transactions on Automatic Control, 64(12):5080–5085, 2019 M. Khajenejad, Z. Jin, and S.Z. Yong. “Interval observers for simultaneous state and model estimation of partially known nonlinear systems”. In American Control Conference, pages 2848–2854. IEEE, 2021. M. Khajenejad and S. Z. Yong. “Simultaneous mode, input and state set-valued observers with applications to resilient estimation against sparse attacks”. In IEEE Conference on Decision and Control (CDC), pages 1544–1550, 2019. S.Z. Yong, M. Zhu, and E. Frazzoli. “Switching and data injection attacks on stochastic cyber-physical systems: Modeling, resilient esti- mation, and attack mitigation”. ACM Transactions on Cyber-Physical Systems, 2(2):9, 2018. Y. Bar-Shalom, T. Kirubarajan, and X.R. Li. “Estimation with Applications to Tracking and Navigation”. John Wiley Sons, Inc., New York, NY, USA, 2002. Y. Nakahira and Y. Mo. “Attack-resilient H2 , H∞ , and ℓ11 state estimator”. IEEE Transactions on Automatic Control, 63(12):4353–4360, 2018. J. Kim, C. Lee, H. Shim, Y. Eun, and J.H. Seo. “Detection of sensor attack and resilient state estimation for uniformly observable nonlinear systems having redundant sensors”. IEEE Transactions on Automatic Control, 64(3):1162–1169, 2018. M.S. Chong, H. Sandberg, and J.P. Hespanha. “A secure state estimation algorithm for nonlinear systems under sensor attacks”. In IEEE Conference on Decision and Control, pages 5743–5748, 2020. M. Milanese and C. Novara. “Set membership identification of nonlinear systems”. Automatica, 40:957–975, 2004. Z.B. Zabinsky, R.L. Smith, and B.P. Kristinsdottir. “Optimal estimation of univariate black-box Lipschitz functions with upper and lower error bounds”. Computers Operations Res., 30(10):1539–1553, 2003. Z. Jin, M. Khajenejad, and S.Z. Yong. “Data-driven model invalidation for unknown Lipschitz continuous systems via abstraction”. In American Control Conference (ACC), pages 2975–2980. IEEE, 2020. K.R. Singh, Q. Shen, and S.Z. Yong. “Mesh-based affine abstraction of nonlinear systems with tighter bounds”. In IEEE Conference on Decision and Control (CDC), pages 3056–3061, 2018. M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 20 / 21
  • 30. Thank you! Questions? M. Khajenejad, Z. Jin, T.N. Dinh, S. Z. Yong Resilient State Estimation Dec., 13, 2023 21 / 21