Professional Documents
Culture Documents
2003book Notes
2003book Notes
(Supplemental Notes)
Gang Tao
Department of Electrical and Computer Engineering
University of Virginia
Charlottesville, VA 22903
A Wiley-Interscience Publication
Matrix Theory 1
Additional Lemmas 10
i
ii Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Parametrization of (9.102) 50
Derivation of (10.152) 61
Matrix Theory
Symmetric Matrices
For a matrix M = M T , we have M = ∑ni=1 λi ei eTi where λi and ei are the eigen-
values and eigenvectors of M such that eTi ei = 1 and eTi e j = 0 with i 6= j. With
P = [e1 , e2 , . . . , en ] and Λ = diag{λ1 , λ2 , . . . , λn }, it follows that M = PΛPT , where
PPT = PT P = I, and, in addition, for M nonsingular, that M −1 = PΛ−1 PT .
√
For M = M T ≥ 0, we define M 1/2 = ∑ni=1 λi ei eTi = PΛ1/2 PT and express
M = M 1/2 M 1/2 , where (M 1/2 )T = M 1/2 . It also follows that for M nonsingular,
(M 1/2 )−1 = ∑ni=1 √1λ ei eTi = PΛ−1/2 PT . On the other hand, for Q = PΛ1/2 , we have
i
M = QQT (as compared with M = M 1/2 M 1/2 for (M 1/2 )T = M 1/2 = PΛ1/2 PT ).
x12
V (x) = 2
+ x22 . (1)
1 + x1
For this V (x), such surface plots are closed curves for c < 1 but are open curves for
c > 1, as V (x) is not radially unbounded.
One can draw the phase-plane plot of ẋ = f (x), by obtaining its numerical solu-
tions for some typical initial conditions, for
6x1 2(x1 + x2 )
ẋ1 = − 2
+ 2x2 , ẋ2 = − . (2)
(1 + x1 )2 (1 + x12 )2
For some initial conditions, the solution trajectories do not converge to the origin.
(Are there any trajectories going to ∞?)
6x1 2(x1 +x2 )
One can draw the vector field of ẋ1 = − (1+x 2 )2 + 2x2 , ẋ2 = − (1+x2 )2 .
1 1
2√
The solution x = [x1 , x2 ]T of this system on the hyperbola x2 = x1 − 2
satisfies
ẋ2 1
g1 (x) = =− √ , (3)
ẋ1 1 + 2 2x1 + 2x12
M ẍ + Dẋ + Kx = F, (5)
where M is the mass, D is the damping constant, K is the spring constant, x is the
mass position, and F is the force acting on the mass.
The system energy is
1 1
V (x, ẋ) = M ẋ2 + Kx2 (6)
2 2
whose time derivative is
d
V (x, ẋ) = F ẋ − Dẋ2 . (7)
dt
Over any time interval [0, T ], it follows that
Z T Z T
V (x(T ), ẋ(T )) = V (x(0), ẋ(0)) + F(t)ẋ(t) dt − Dẋ2 (t) dt. (8)
0 0
Since D ≥ 0, we have
Z T
− F(t)ẋ(t) dt ≤ V (x(0), ẋ(0)), (9)
0
which means that the energy extracted from the system is less than or equal to the
initial system energy. From (7), the term F ẋ clearly represents the system absorbed
power from the input force F. For passivity analysis, the product of system input
and output is defined as such power. In this sense we consider the velocity v = ẋ as
system output. Then, the system admittance (the reciprocal of impedance) is
V (s) s
G(s) = = 2
, (10)
F(s) Ms + Ds + K
which is positively real (the mechanical-electric analogy pairs are “force vs. volt-
age” and “velocity vs. current”).
If we consider x as system output, the term Fx does not represents system power
(i.e., 0T F(t)x(t) dt does not represents system energy) so that the passivity analysis
R
is not applicable. In other words, the system transfer function for passivity analysis
in terms of positive realness is defined in terms of system impedance (or admit-
tance) relating current (velocity) to voltage (force) (or voltage (force) to current
(velocity)).
Without the passivity property, the transfer function from the input force to the
1
output position, Ms2 +Ds+K , cannot be positive real.
4 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
whenever ω2 > (−ε3 + 2ε2 − 2ε + 1)/ε, that is, h(s − ε) cannot be positive real
for any ε > 0. (P(ε) = −ε3 + 2ε2 − 2ε + 1 = (ε − 1)(−ε2 + ε − 1) > 0 for ε ∈
(0, 1) and P(ε) < 0 for ε > 1. For h(s − ε) to be stable, ε ∈ [0, 0.5) is needed as
(s − ε)2 + s − ε + 1 = s2 + (1 − 2ε)s + 1 − ε + ε2 and 1 − ε + ε2 > 0 for any ε.)
6 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
where hc (s) = c(sI − A)−1 and hb (s) = (sI − A)−1 b. Here we derive an alternative
form of this lemma, explicitly in terms of the parameters of the function h(s).
∗ ∗
Denoting Pm (s) = sn + an∗ −1 sn −1 + · · · + a1 s + a0 , for vector signals θ(t) and
ω(t), from (5.138) and with an∗ = 1, we have
1 1
θT (t) [ω](t) − [θT ω](t)
Pm (s) Pm (s)
∗ −i ∗ !
n∗ ∑nj=0 an∗ − j sn −i− j T s
i−1
= ∑ Pm (s)
θ̇
Pm (s)
[ω] (t). (17)
i=1
∗
Introducing F(s) = fn∗ −1 sn −1 + · · · + f1 s + f0 , we express
1
F(s) θ (t)
T
[ω] (t)
Pm (s)
∗ −2
1 s
f0
= fn∗ −1 s n
+ · · · + f1 θ̇ T
[ω] + θT
[ω] (t) + θT (t) [ω](t)
Pm (s) Pm (s) Pm (s)
∗ −2
1
= fn∗ −1 s n
+ · · · + f1 θ̇ T
[ω] (t)
Pm (s)
∗ −3
s s2
+ fn∗ −1 s n
+ · · · + f2 θ̇ T
[ω] + θT
[ω] (t)
Pm (s) Pm (s)
f1 s + f0
+ θT (t) [ω](t) = · · ·
Pm (s)
!
n∗ −1 n∗ −i
si−1 F(s)
= ∑ ∑ fn∗ − j s
∗
n −i− j
θ̇ T
[ω] (t) + θT (t) [ω](t) (18)
i=1 j=1 Pm (s) Pm (s)
where
n∗ −i n∗ −i
∑ n∗ −i− j
∑ an∗− j sn −i− j , i = 1, 2, . . . , n∗ − 1,
∗
αi (s) = Pm (s) fn∗ − j s − F(s) (21)
j=1 j=0
△
are polynomials of degrees n∗ − 1 or less, and αn∗ (s) = −F(s).
For a proper transfer function h(s) = fn∗ + PF(s)
m (s)
, it follows that
n∗
αi (s) T si−1
h(s)[θ ω](t) − θ (t)h(s)[ω](t) = ∑
T T
θ̇ [ω] (t). (22)
i=1 Pm (s) Pm (s)
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 9
C(zI −A)−1 B[u](t), and its modification: w(t) = C ∑ti=0 At−1 Bu(i) for w(t) = C(zI −
A)−1 Bz[u](t) = C(zI − A)−1 B[u](t + 1), we express
where, by definition,
t−1
c ∑ At−i−1 b ω(i) = h(z)[ω](t) (25)
i=0
t−1
c ∑ At−τ−1 b ωT (τ)θ(τ) = h(z)[θT ω](t). (26)
τ=0
This is the discrete-time version of the swapping lemma (5.331), whose explicit
version can also be derived, similar to the continuous-time case above.
10 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Additional Lemmas
for some β(t) such that limt→∞ β(t) = 0 and some γ(t) ∈ L∞ , then w(t) is bounded,
that is, w(t) ∈ L∞ . If, in addition, limt→∞ γ(t) = 0, then limt→∞ w(t) = 0.
Proof: Assume that w(t) is not bounded. Then there exists a subsequence {tn } such
that limtn →∞ |w(tn )| = ∞ and |w(t)| ≤ |w(tn )| for t ≤ tn , that is, kwktn ≤ |w(tn )|, and
it follows that
Since limtn →∞ β(tn ) = 0 and γ(t) ∈ L∞ , the above inequality implies that w(tn ) is
bounded, a contradiction to the assumption that limtn →∞ w(tn ) = ∞ or w(t) is un-
bounded. Hence, w(t) is bounded, and so is kwkt .
If limt→∞ γ(t) = 0 also holds, then from (27), the boundedness of w(t) and the
condition that limt→∞ β(t) = 0, it follows that limt→∞ |w(t)| = limt→∞ (β(t)kwkt +
γ(t)) = 0. ∇
Lemma 2 For kwkt = sup0≤τ≤t kw(τ)k with kw(t)k being a vector norm of w(t) ∈
Rn , if
kw(t)k ≤ β(t)kwkt + γ(t) (29)
for some β(t) such that limt→∞ β(t) = 0 and some γ(t) ∈ L∞ , then w(t) is bounded,
that is, w(t) ∈ L∞ . If, in addition, limt→∞ γ(t) = 0, then limt→∞ w(t) = 0.
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 11
Proof: We need to show that for any given η > 0, there exists a T > 0 such that
|ė(t)| < η for any t > T .
The proof makes use of two related fictitious filters with a virtual parameter
1 a
a > 0: H(s) = s+a and K(s) = s+a such that 1 = sH(s) + K(s). Operating this
identity on ė(t), we have
Under the condition that ë(t) ∈ L∞ , for any given η > 0, H(s)[ë](t) can be made
smaller than η2 by a choice of a sufficiently large and finite a. Under the condition
that limt→∞ e(t) = 0, for the chosen value of a, there exists a T > 0 such that for
any t > T , |sK(s)[e](t)| < η2 .
This derivation means that for any given η > 0, there exists a T > 0 such that
|ė(t)| < η for any t > T , which is equivalent to the convergence of ė(t) to 0 as
t → ∞. ∇
Remark 2 This result also follows from the Barbalat Lemma: If a scalar function
f (t) is uniformly continuous such that limt→∞ 0t f (τ) dτ exists and is finite, then e
R
limt→∞ f (t) = 0. In this case, f (t) = ė(t), limt→∞ 0t f (τ) dτ = limt→∞ 0t ė(τ) dτ =
R R
limt→∞ (e(t) − e(0)) = −e(0) exists and is finite, and f (t) is uniformly continuous,
as f˙(t) = ë(t) is bounded. 2
⋄ Convergence of e(t) (which can be a tracking error signal e(t) = y(t) − ym (t)):
With f (t) = e2 (t), we have that f˙(t) = 2e(t)ė(t) is bounded, i.e., f (t) is uniformly
12 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Rt
continuous, so that, with e(t) ∈ L2 (that is, limt→∞ 0 f (τ) dτ exists and is finite),
limt→∞ f (t) = 0, that is, limt→∞ e(t) = 0.
Proposition 1: If limt→∞ e(t) = e0 being a constant and ë(t) is bounded, then
limt→∞ ė(t) = 0.
• Barbalat lemma based proof
⋄ Convergence of ė(t) with limt→∞ e(t) = 0: For f (t) = ė(t), ë(t) being bounded
means f (t) being uniformly continuous, and limt→∞ 0t f (τ) dτ = e(∞) − e(0) =
R
−e(0) exists and is finite, so that limt→∞ f (t) = 0, that is, limt→∞ ė(t) = 0.
⋄ Convergence of ė(t) with limt→∞ e(t) = e0 ∈ R: For f (t) = ė(t), ë(t) being
bounded means f (t) being uniformly continuous, and limt→∞ 0t f (τ) dτ = e(∞) −
R
For MIMO systems, there are also different ways to express the system transfer
matrix G(s) = C(sI −A)−1 B (we first assume that G(s) has full rank m for an M ×M
G(s), that is, rank[G(s)] < M only for a finite number of values of s).
The first (simple) form of G(s) is
N(s)
G(s) = C(sI − A)−1 B = , d(s) = det[sI − A], ∂d(s) = n. (31)
d(s)
If we use a controller of order n − 1 = ∂Λ(s), we need to assume that all zeros of
det[N(s)] are stable, which may be thought as the system zeros (note that the zeros
of G(s), as defined in the literature, may be only a subset of those zeros if (A, B,C)
is not a minimal realization). In fact, different from the SISO case, the controller
order can be reduced to n − M in the MIMO case, because a reduced-order observer
has order n − M, so that the state feedback control law u = K1∗T x + K2∗ r can be
reparametrized using the input and output signals, leading to an (n − M)th order
output feedback controller.
Given that the observability index of a minimal realization of G(s) is ν ≤ n −
M +1, that is, ν−1 ≤ n−M, the order of the usual MRAC structure, which has been
chosen as ν − 1, is the minimal order to meet the desired plant-model matching. If
for a system, the order of a reduced-order observer can be chosen as ν − 1, the
parametrization of the output feedback controller with order ν − 1 can then be seen
as a reparametrization of a state feedback controller, using input and output signals.
If not, the controller parametrization may be considered as being one used for plant-
model matching. In MRAC for multivariable systems, we call ν the observability
index of the system transfer matrix G(s), in the sense that ν is the observability
index of a minimal realization of the system transfer matrix G(s).
The second form of G(s) is a left matrix-fraction description:
where P̄l (s) and Z̄l (s) are M × M polynomial matrices with P̄l (s) row reduced (a
polynomial matrix P̄l (s) is row reduced if the elements in the ith row of P̄l (s) have
a largest degree ν̄i and the matrix Γr = lims→∞ diag{s−ν̄1 , s−ν̄2 , . . . , s−ν̄M }P̄l (D) is
nonsingular), and the ith row degree of Z̄l (s) is less than the ith row degree of P̄l (s),
i = 1, 2, . . . , q (it is denoted that ∂P̄l (s) = ν̄, that is, ν̄ = max{ν̄i }). A controller can
be designed with ∂Λ(s) = ν̄ − 1. In this case, for stable MRAC, we need to as-
sume that the zeros of det[Z̄l (s)] are stable and also that (A, B,C) is stabilizable and
detectable (as some pole-zero cancellations may occur when obtaining the system
model P̄l−1 (s)Z̄l (s) from (A, B,C)).
Note that a left matrix-fraction description G(s) = P̄l−1 (s)Z̄l (s), can always be
made to have a row reduced P̄l (s), by using elementary row operations represented
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 15
G(s) = P̃l−1 (s)Z̃l (s) = (M(s)P̃l (s))−1 M(s)Z̃l (s) = P̄l−1 (s)Z̄l (s), (33)
for P̄l (s) = M(s)P̃l (s) and Z̄l (s) = M(s)Z̃l (s).
The third form of G(s) is a left co-prime matrix-fraction description:
where Pl (s) and Zl (s) are left co-prime M × M polynomial matrices (Pl (s) and Zl (s)
are left co-prime if any M × M polynomial matrix W (s) such that Zl (s) = W (s)Z(s)
and Pl (s) = W (s)P(s) for some polynomial matrices Z(s) and P(s)—such a W (s)
is called a common left divisor of Zl (s) and Pl (s), is a unimodular matrix, that is,
det[W (s)] is a non-zero constant), with Pl (s) row reduced and ∂Pl (s) = ν ≤ ν̄. A
controller can be designed with ∂Λ(s) = ν − 1. In this case, we also need to assume
that the zeros of det[Zl (s)] are stable and that (A, B,C) is stabilizable and detectable.
The situation is similar for G(s) in a right matrix-fraction description:
with Pr (s) being column reduced (that is, PrT (s) being row reduced) and Zr (s) and
Pr (s) being right co-prime (that is, ZrT (s) and PrT (s) being left co-prime). In par-
ticular, the column degrees of Pr (s) are denoted as µi (or µ̄i for P̄r (s) if G(s) =
C(sI − A)−1 B = Z̄r (s)P̄r−1 with Z̄r (s) and P̄r (s) not right co-prime), i = 1, 2, . . . , m,
and the column degrees of Zr (s) (Z̄r (s)) are less than that of Pr (s) (P̄r (s)). If (A, B,C)
is a minimal realization, then µi , i = 1, 2, . . . , m, are the controllability indexes of
(A, B). If (A, B,C) is not a minimal realization, then one can a controllable realiza-
tion (Āc , B̄c , C̄c ) whose controllability indexes are µ̄i , for Z̄r (s)P̄r−1 (s) (or (Ac , Bc ,Cc )
whose controllability indexes are µi , for Zr (s)Pr−1 ).
Note that if (A, B,C) is a minimal realization, then ν is the observability index
of (A,C) (we may also call ν the observability index of G(s)). When (A, B,C) is not
minimal, for the second case (or the third case), we can find an observable realiza-
tion (Āo , B̄o , C̄o ) whose observability indexes are ν̄i , for P̄l−1 (s)Z̄l (s) (or (Ao , Bo ,Co )
whose observability indexes are νi , for Pl−1 (s)Zl (s)). In this sense, we may also call
ν̄ (or ν) the observability of G(s) = P̄l−1 (s)Z̄l (s) (or G(s) = Pl−1 (s)Zl (s)).
In summary, for MRAC of a system (A, B,C), a basic assumption is that (A, B,C)
is stabilizable and detectable, in addition to the assumption that all zeros of the sys-
tem transfer matrix G(s) = C(sI −A)−1 B are stable; in other words, all system zeros
are required to be stable.
For more about the definitions of the zeros and poles of G(s), see Kailath (1980)
and Rugh (1996), and also see more notes.
16 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
where A ∈ Rn×n , B ∈ Rn are some unknown constant parameter matrix and param-
eter and vector, and assume that the state vector x(t) is available for measurement.
The control objective is to design a state feedback control law for u(t) such
that all signals in the closed-loop system are bounded and the plant state vector
x(t) asymptotically tracks a reference state vector xm (t) generated from a chosen
reference model system
where k1 (t) and k2 (t) are the estimates of k1∗ and k2∗ .
For the tracking error x(t) − xm (t) and the parameter errors
θ̄˜ (t) = (Γ̄T1 )−1 θ̃(t), θ̃(t) = [k̃1T (t), k̃2 (t)]T ∈ Rn+1 , (45)
Since P̄ = k2∗ P = P̄T > 0, there is a nonsingular matrix P̄1 = P̄1T ∈ Rn×n such that
P̄ = P̄1 P̄1 .1 With such a nonsingular P̄1 = P̄1T , we introduce the transformed tracking
error
ē(t) = P̄1 e(t). (47)
Then, with Ām = P̄1 Am P̄1−1 and B̄ = P̄1 B, we can express the tracking error equation
(41) as
ē˙ (t) = P̄1 Am e(t) + P̄1 BωT (t)θ̃(t) = Ām ē(t) + B̄ω̄T (t)θ̄˜ (t) (48)
and, for k2∗ > 0, write the adaptive laws (42)–(43) as
˙
θ̄˜ (t) = −ω̄(t)B̄T ē(t). (49)
Here, the matrix Ām = P̄1 Am P̄1−1 with P̄ = k2∗ P = P̄1 P̄1 and P̄1 = P̄1T has the property:
To study the convergence properties of the parameter error θ̃(t) in (41), or equiv-
alently, that of θ̄˜ (t) in (50), we first introduce the following definition:
Result (Morgan and Narendra 1977; Narendra and Annaswamy 1989): For the
system
ż(t) A0 B0 ζT0 (t) z(t)
= , (53)
φ̇(t) −ζ0 (t)BT0 0 φ(t)
with z(t) ∈ Rn , φ(t) ∈ Rn+1 , ζ0 (t) ∈ Rn+1 , A0 ∈ Rn×n being stable (i.e., all its eigen-
values are in Re[s] < 0) such that A0 + AT0 < 0, and B0 ∈ Rn and (A0 , B0 ) being
controllable, if ζ0 (t) is bounded and PE, then limt→∞ φ(t) = 0 and limt→∞ z(t) = 0,
both exponentially.
Result: The MRAC system has been transformed to (50) (with Ām = P̄1 Am P̄1−1
stable and Ām + ĀTm < 0 as shown in (51), and ω̄(t) = Γ̄1 ω(t)) which has the same
form as that in (53).
Remark: ω(t) is PE iff and ω̄(t) is PE, for ω̄(t) = Γ̄1 ω(t).
Result (Narendra and Annaswamy (1989); Ioannou and Sun (1996)): For the sys-
tem
ẋm (t) = Am xm (t) + Bm r(t), xm (t) ∈ Rn , r(t) ∈ R (54)
where Am ∈ Rn×n is stable and (Am , Bm ) is controllable, if the input signal r(t) has
n + 1 or more frequencies, then ωm (t) = [xm
T (t), r(t)]T is PE.
Remark: r(t) = 1 has one frequency at 0, and r(t) = sin 2t has two frequencies at
2 and −2, etc.
Result: MRAC ensures that all closed-loop system signals are bounded and (ω(t)−
ωm (t)) ∈ L2 .
Remark: (A, B) is controllable iff (Am , Bm ) is controllable iff (Ām , B̄) is control-
lable.
Remark: limt→∞ θ̃(t) = 0 and limt→∞ e(t) = 0, both exponentially iff limt→∞ θ̄˜ (t) =
0 and limt→∞ ē(t) = 0, both exponentially, for θ̄˜ (t) = (Γ̄T1 )−1 θ̃(t) and ē(t) = P̄1 e(t).
In conclusion, with the application of the result for (53) to the MRAC system
equation (50), the state feedback state tracking MRAC system, with (A, B) con-
trollable and r(t) of n + 1 or more frequencies, ensures that limt→∞ θ̃(t) = 0 and
limt→∞ e(t) = 0, both exponentially.
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 19
Z(s)
y(t) = G(s)[u](t), G(s) = k p (55)
P(s)
with some unknown monic polynomials P(s) and Z(s) of degrees n and m respec-
tively, and some gain k p known or unknown, and assume that Z(s) is a stable poly-
nomial for MRAC.
u(t) = θT1 ω1 (t) + θT2 ω2 (t) + θ20 y(t) + θ∗3 r(t), (56)
where
a(s) a(s)
ω1 (t) = [u](t), ω2 (t) = [y](t), (57)
Λ(s) Λ(s)
a(s) = [1, s, . . . , sn−2 ]T , (58)
θ1 ∈ Rn−1 , θ2 ∈ Rn−1 and θ20 ∈ R are parameters to be adaptively updated, Λ(s) is
a monic stable polynomial of degree n − 1, and θ∗3 = k−1 p .
The control objective is to design the input signal u(t) to ensure all closed-loop
signals bounded and the output y(t) asymptotically tracking the reference output
signal ym (t) satisfying
Pm (s)[ym ](t) = r(t), (59)
where Pm (s) is a monic stable polynomial of degree n∗ = n − m, and r(t) is a
bounded and piecewise continuous reference input signal.
The parameters θ1 ∈ Rn−1 , θ2 ∈ Rn−1 and θ20 ∈ R are estimates of the nominal
parameters θ∗1 ∈ Rn−1 , θ∗2 ∈ Rn−1 and θ∗20 ∈ R satisfying the matching equation
θ∗T
1 a(s)P(s) + (θ2 a(s) + θ20 Λ(s))k p Z(s) = Λ(s)(P(s) − k p θ3 Z(s)Pm (s)). (60)
∗T ∗ ∗
1 , θ2 , θ20 ] ∈ R
where θ̃0 (t) = θ0 (t) − θ∗0 , with θ∗0 = [θ∗T ∗T ∗ T 2n−1 being the nominal
value of θ0 (t) and depends on the parameters of P(s) and Z(s) as the plant-model
matching parameter vector. The estimation error ε(t) can be expressed as
ε(t) = θ̃T0 (t)ζ0 (t), (69)
and the adaptive law with k p > 0 can be expressed as
ζ0 (t)ζ0 (t)θ̃0 (t)
T
θ̃˙ 0 (t) = −Γ . (70)
m2 (t)
Since Γ = ΓT > 0, there exists a nonsingular matrix Γ1 such that Γ = ΓT1 Γ1 . Hence,
we have
Γ1 ζ0 (t)ζT0 (t)ΓT1 T −1
(ΓT1 )−1 θ̃˙ 0 (t) = − (Γ1 ) θ̃0 (t). (71)
m2 (t)
Γ1 ζ0 (t)
Letting φ0 (t) = (ΓT1 )−1 θ̃0 (t) and ψ0 (t) = m(t) , we arrive at
aT (s) aT (s)
ζ0 (t) = Wm (s)[ [u](t), [y](t), y(t)]T
Λ(s) Λ(s)
T
a (s) −1 aT (s)
= Wm (s)[ G (s)[y](t), [y](t), y(t)]T , (74)
Λ(s) Λ(s)
1
where Wm (s) = Pm (s) , In a similar structure, we introduce the signal
aT (s) −1 aT (s)
ζ0m (t) = Wm (s)[ G (s)[ym ](t), [ym ](t), ym (t)]T
Λ(s) Λ(s)
= Wm (s)H0 (s)[r](t) ∈ R2n−1 , (75)
where
aT (s) −1 aT (s)
H0 (s) = [ G (s)Wm (s), Wm (s),Wm (s)]T . (76)
Λ(s) Λ(s)
Result (Anderson 1997): If ψ0 (t) in (72) is PE, then limt→∞ φ0 (t) = 0 exponentially,
that is, limt→∞ (θ0 (t) − θ∗0 ) = 0 exponentially.
Result: If the signal ζ0 (t) is PE, then ψ0 (t) is PE (which follows from the bound-
edness of m(t)).
Result: MRAC ensures that (ζ0 (t) − ζ0m (t)) ∈ L2 (which can be seen from ζ0 (t) −
ζ0m (t) = H(s)[e](t), where H(s) is strictly proper and stable, and e(t) ∈ L2 ).
Result: Given that (ζ0 (t) − ζ0m (t)) ∈ L2 , ζ0 (t) is PE iff ζ0m (t) is PE.
Result (Boyd and Satry 1983): If r(t) has 2n − 1 or more frequencies and P(s) and
Z(s) are coprime, then ζ0m (t) is PE.
In conclusion, for a MRAC system with P(s) and Z(s) coprime and k p known, if
the reference signal r(t) has 2n − 1 or more frequencies, then limt→∞ (θ(t) − θ∗0 ) = 0
and limt→∞ (y(t) − ym (t)) = 0, both exponentially.
22 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
1
ξ(t) = θT (t)ζ(t) − [θT ω](t) (81)
Pm (s)
θ(t) = [θT1 (t), θT2 (t), θ20 (t), θ3 (t)]T ∈ R2n , (82)
and the adaptive laws for θ(t) and ρ(t) are chosen as
ζ(t)ε(t)
θ̇(t) = −sign[k p ]Γ (83)
m2 (t)
γε(t)ξ(t)
ρ̇(t) = − (84)
m2 (t)
where ρ∗ = k p , θ̃(t) = θ(t) − θ∗ and ρ̃(t) = ρ(t) − ρ∗ . Introducing θ̄˜ (t) = θ̄(t) − θ̄∗
with
θ̄(t) = [θT (t), ρ(t)]T ∈ R2n+1 , θ̄∗ = [θ∗T , ρ∗ ]T ∈ R2n+1 (88)
and defining Γ̄ = diag{Γk−1
p , γ} (for k p > 0), and
While this expression has the same form as that in (70), it does not render the
same procedure for the convergence analysis of θ̄˜ (t), because ζ̄(t) cannot be PE as
limt→∞ ξ(t) = 0 from MRAC.
In this case, we consider (83) and (87) with k p > 0:
ζ(t)ζ (t)
T ζ(t)ρ̃(t)ξ(t)
θ̃˙ (t) = −Γk p 2
θ̃(t) − Γ , (91)
m (t) m2 (t)
aT (s) aT (s)
ζ(t) = Wm (s)[ [u](t), [y](t), y(t), r(t)]T
Λ(s) Λ(s)
aT (s) −1 aT (s)
= Wm (s)[ G (s)[y](t), [y](t), y(t), r(t)]T , (92)
Λ(s) Λ(s)
aT (s) −1 aT (s)
ζm (t) = Wm (s)[ G (s)[ym ](t), [ym ](t), ym (t), r(t)]T
Λ(s) Λ(s)
= Wm (s)H(s)[r](t) ∈ R2n , (93)
where
aT (s) −1 aT (s)
H(s) = [ G (s)Wm (s), Wm (s),Wm (s), 1]T . (94)
Λ(s) Λ(s)
We have the following results.
Result: If r(t) has 2n or more frequencies, and P(s) and Z(s) are coprime, then
ζm (t) is PE.
Result: MRAC ensures that (ζ(t) − ζm (t)) ∈ L2 , and ξ(t) ∈ L2 and limt→∞ ξ(t) = 0.
24 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Result: If ζ(t) is PE, then the solution θ̃(t) of the homogeneous part of (91): θ̃˙ (t) =
−Γk p ζ(t)ζ
T (t)
2
m (t)
θ̃(t), has the property: limt→∞ θ̃(t) = 0 exponentially.
Result: If ζ(t) is PE, then the solution θ̃(t) of (91) has the property:
This follows from the system signal boundedness and the ζ(t) PE condition,
which imply that
ζ(t)ζT (t)
θ̃˙ (t) = −Γk p θ̃(t) (96)
m2 (t)
is an exponentially stable system, and from the properties: limt→∞ ξ(t) = 0 and
ξ(t) ∈ L2 .
In conclusion, for a MRAC system with P(s) and Z(s) coprime and k p > 0 un-
known, if the reference signal r(t) has 2n or more frequencies, then limt→∞ (θ(t) −
θ∗ ) = 0 and (θ(t) − θ∗ ) ∈ L2 , in addition to the usual MRAC tracking error proper-
ties: limt→∞ (y(t) − ym (t)) = 0 and (y(t) − ym (t)) ∈ L2 .
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 25
Theorem 1 For a model reference adaptive control system with the plant relative
degree n∗ > 0, the tracking error e(t) = y(t) − ym (t) has the convergence property:
i
limt→∞ d dte(t) ∗
i = 0, for i = 0, 1, . . . , n − 1.
Proof: The property: limt→∞ e(t) = 0 follows directly from the established prop-
erties: 0∞ e2 (t)dt < ∞ and de(t)
R
dt is bounded. However, it is not clear whether
d i e(t) i
dt i
is square-integrable or not for i > 0. To prove the convergence of d dte(t) i , for
i = 1, . . . , n −1, we recall that for a function f (t) defined on [t0 , ∞), limt→∞ f (t) = 0
∗
if for every η > 0, there exists a T = T (η) > 0 such that | f (t)| < η, ∀t > T . Hence,
our goal now is to show that in each case of i = 1, . . . , n∗ − 1, for any given η,
i
i | < η, for all t > T . To reach this goal, we use
there exists a T > 0 such that | d dte(t)
i
a method to decompose the signal d dte(t)i into two fictitious parts: one being small
enough and one converging to zero asymptotically with time going to infinity. To
this end, we first introduce two fictitious filters K(s) and H(s) from
∗
an
K(s) = ∗ , sH(s) = 1 − K(s) (97)
(s + a)n
where a > 0 is a generic constant to be specified. The filter H(s) is given as
∗ ∗
1 1 (s + a)n − an
H(s) = (1 − K(s)) = ∗ (98)
s s (s + a)n
which is strictly proper (with relative degree 1) and stable and whose impulse re-
sponse function is
n∗ ∗
an −i n∗ −i
h(t) = L −1
[H(s)] = e −at
∑ ∗ t . (99)
i=1 (n − i)!
26 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
where Pm (s) is a stable polynomial of degree n∗ , and θ̃(t) is the parameter error
vector and ω(t) is the controller regressor vector, and both are bounded. Using
(97): 1 = sH(s) + K(s), we express ė(t) = de(t)
dt as
k ps T
ė(t) = [θ̃ ω](t)
Pm (s)
k p s2 T kp
= H(s) [θ̃ ω](t) + sK(s) [θ̃T ω](t)
Pm (s) Pm (s)
k ps 2
= H(s) [θ̃T ω](t) + sK(s)[e](t). (102)
Pm (s)
To demonstrate the new technique for proving the new result in Theorem 1:
i
limt→∞ d dte(t) ∗
i = 0, for i = 1, . . . , n − 1, under the previously proved properties of
MRAC: all closed-loop signals are bounded and the tracking error e(t) = y(t) −
ym (t) satisfies limt→∞ e(t) = 0, we first consider the case of n∗ = 2, where
a2 s + 2a
K(s) = 2
, H(s) = . (103)
(s + a) (s + a)2
k p s2 k p s2
Since θ̃T (t)ω(t) is bounded and Pm (s) is stable and proper, Pm (s) [θ̃ ω](t) is bounded.
T
k p s2 T c0
|H(s) [θ̃ ω](t)| ≤ (104)
Pm (s) a
for any t ≥ 0 and some constant c0 > 0 independent of a > 0. Since limt→∞ e(t) = 0
as established and sK(s) is stable and strictly proper (with relative degree n∗ − 1 = 1
in this case of n∗ = 2), it follows, for any a > 0 in K(s), that
To show that, for every η > 0, there exists a T > 0 such that |ė(t)| < η, ∀t > T ,
we set a > a(η) = 2cη0 for the fictitious filter H(s) in (98) so that ca0 < η2 in (104),
and let T = Ta (a(η), η) = T (η) > 0 such that |sK(s)[e](t)| < η2 for all t ≥ T (which
△
is ensured by the property that limt→∞ sK(s)[e](t) = 0 for any finite a > 0 in the
other related fictitious filter K(s) in (97)) (the peak value of |sK(s)[e](t)| depends
on the parameter a, so that the above time instant T = Ta (a(η), η) also depends on
a too). Then, it follows from (102) and (104) that
η η
|ė(t)| < + = η, ∀t > T, (106)
2 2
which implies that limt→∞ ė(t) = 0. In other words, if limt→∞ ė(t) 6= 0, then there
exist an η0 > 0 and a sequence of time instants ti with limi→∞ ti = ∞ such that
|ė(ti )| > η0 for all i = 1, 2, . . ., which, from the above analysis, is impossible.
a3 s2 + 3as + 3a2
K(s) = , H(s) = , (107)
(s + a)3 (s + a)3
k p s2
we see that in (102), Pm (s) is stable and strictly proper, and so is sK(s), so that
2 d 2 e(t)
limt→∞ ė(t) = 0. To show limt→∞ d dte(t)
2 = 0, from (102), we express dt 2
as
d 2 e(t) k p s3 T
= H(s) [θ̃ ω](t) + s2 K(s)[e](t), (108)
dt 2 Pm (s)
k p s3 s3
Pm (s) [θ̃ ω](t) is stable and proper and θ̃T (t)ω(t) is
in which T is bounded (as Pm (s)
bounded) and H(s) satisfies (100), and limt→∞ s2 K(s)[e](t) = 0. Hence, similar to
2 d 2 e(t)
(106), now with d dte(t)
2 replacing ė(t) in (106), it also follows that limt→∞ dt 2 = 0.
In general, for a MRAC system with an arbitrary relative degree n∗ > 0, from
i
(102), we express the ith-order time derivative d dte(t)
i of e(t) as
d i e(t) k p si+1 T
= H(s) [θ̃ ω](t) + si K(s)[e](t). (109)
dt i Pm (s)
k p si+1
Since θ̃T (t)ω(t) is bounded, and for each i = 1, 2, . . . , n∗ − 1, Pm (s) is stable and
strictly proper (proper for i = n∗ − 1), we have
k p si+1 T ci
|H(s) [θ̃ ω](t)| ≤ (110)
Pm (s) a
28 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Hence, with the use of the fictitious parameter a > 0 in H(s) and K(s), similar to
(106), it can be shown that for every η > 0, there exists a T = T (η) > 0 such that
i d i e(t)
i | < η for all t > T , so that limt→∞ dt i = 0, for i = 1, 2, . . . , n − 1.
| d dte(t) ∗ ∇
For a general MRAC system, the reference input signal r(t) is only required to
be bounded (which is sufficient for the result of Theorem 1). If the time-derivative
of r(t) is also bounded, we have the following additional property.
Corollary 1 For a MRAC system with relative degree n∗ > 0, if both r(t) and ṙ(t)
are bounded, then the tracking error e(t) = y(t) − ym (t) has the convergence prop-
i
erty: limt→∞ d dte(t) ∗
i = 0, for i = 0, 1, . . . , n .
n∗
Proof: We just need to show the additional convergence property: limt→∞ d dte(t)
n∗ =
0. To this end, we consider (109) with i = n∗ :
∗ ∗
d n e(t) k p sn ∗
n∗ = H(s) s[θ̃T ω](t) + sn K(s)[e](t). (112)
dt Pm (s)
T
In this case, s[θ̃T ω](t) = dt
d
(θ̃T (t)ω(t)) = θ̃˙ (t)ω(t) + θ̃T ω̇(t) is bounded, because
θ̃˙ (t) and ω̇(t) (whose last component is ṙ(t) which is bounded by assumption) are
∗
k sn ∗
bounded. Hence Pmp (s) s[θ̃T ω](t) is bounded. Since sn K(s) in (112) is stable and
proper and H(s) satisfies (100), we have, with the property: limt→∞ e(t) = 0 and an
∗ n∗
d n e(t)
inequality similar to (106) with dt n
∗ replacing ė(t), that limt→∞ d dte(t)
n∗ = 0. ∇
for some exponentially decaying and initial condition related term ε1 (t), where
Pm (s) is a stable polynomial of degree n∗ . It follows from this equation that the
i
tracking error e(t) = y(t) − ym (t) has the convergence property: limt→∞ d dte(t)
i = 0,
∗
for i = 0, 1, . . . , n (under the condition that r(t) is bounded). The difference be-
tween the nominal control and adaptive control cases is that in the adaptive control
n∗
case the additional condition that ṙ(t) is bounded was used to show limt→∞ d dte(t)
n∗ =
0, which is not needed for the nominal control case. 2
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 29
i
Remark 4 For the case when ym (t) = 0, it follows that limt→∞ d dty(t) i = 0, for i =
∗
0, 1, . . . , n − 1, an output regulation result of MRAC.
For a state-space system model: ẋ = Ax + Bu, y = Cx with x(t) ∈ Rn , if n∗ = n,
i
then limt→∞ d dty(t)i = 0, for i = 0, 1, . . . , n−1, as established above, and limt→∞ x(t) =
0, as (A,C) is observable implied by the condition that n∗ = n (the system transfer
function does not have any finite zeros as its numerator is just a constant k p ). This
is the adaptive asymptotic state regulation result, guaranteed by a MRAC design.
In particular, for an nth order system with (A, B) in the controllable canonical
form and y(t) = x1 (t), the system transfer function explicitly has no finite zeros
as its numerator is a constant, leading to n∗ = n and an observable (A,C), so that
limt→∞ x(t) = 0, an inherent property of a MRAC system. 2
i
This new higher-order tracking error convergence property: limt→∞ d dte(t)
i = 0,
∗
for i = 1, . . . , n , of MRAC, is extendable to adaptive nonlinear tracking control
systems in which the controlled nonlinear plant has a defined relative degree n∗ .
30 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
θ∗ = (θ∗T
1 , θ2 , θ20 , θ3 ) ∈ R
∗T ∗ ∗ T 2n̄
(114)
θ(t) = (θT1 (t), θT2 (t), θ20 (t), θ3 (t))T ∈ R2n̄ (115)
ω(t) = (ωT1 (t), ωT2 (t), y(t), r(t))T ∈ R2n̄ (116)
θ̃(t) = θ(t) − θ∗ , e(t) = y(t) − ym (t), (117)
and derived the tracking error equation
kp
e(t) = [θ̃T ω](t)
Pm (s)
1 1
= −k p (θ∗T [ω](t) − [θT ω](t)), (118)
Pm (s) Pm (s)
where Pm (s) is a monic and stable polynomial of degree n∗ .
We also defined the estimation error
that is, ε(t) = e(t)− the estimate of −k p (θ∗T Pm1(s) [ω](t) − Pm1(s) [θT ω](t)).
The estimation error ε(t) can be expressed as its theoretical form
In the MRAC literature, it was shown that limt→∞ ε(t) = 0, and limt→∞ e(t) = 0,
kp
that is, limt→∞ k p θ̃T (t)ζ(t) + ρ̃(t)ξ(t) = 0, and limt→∞ Pm (s) [θ̃T ω](t) = 0.
How about limt→∞ k p θ̃T (t)ω(t) = 0?
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 31
Γφ(t)ε(t)
θ̇(t) = − , θ(0) = θ0 (125)
m2 (t)
we can show that if, in addition, some order derivatives of y(t) and φ(t) are bounded,
then certain order derivatives of ε(t) also converge to 0, using the following lemma.
Lemma: If limt→∞ f (t) = f0 has a constant limit f0 and f¨(t) is bounded, then
limt→∞ f˙(t) = 0.
For example, with limt→∞ ε(t) = 0, if ε̈(t) is bounded, then limt→∞ ε̇(t) = 0.
To specify the conditions for ε̈(t) to be bounded, from ε(t) = θT (t)φ(t) − y(t),
we have
ε̇(t) = θ̇T (t)φ(t) + θT (t)φ̇(t) − ẏ(t) (129)
ε̈(t) = θ̈T (t)φ(t) + θ̇T (t)φ̇(t) + θ̇T (t)φ̇(t) + θT (t)φ̈(t) − ÿ(t). (130)
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 33
kp
e(t) = [θ̃T ω](t), (132)
Pm (s)
where k p is the plant high frequency gain, Pm (s) is a monic and stable polynomial
of degree n∗ (the plant relative degree), and
θ∗ = (θ∗T
1 , θ2 , θ20 , θ3 ) ∈ R
∗T ∗ ∗ T 2n
(134)
θ(t) = (θT1 (t), θT2 (t), θ20 (t), θ3 (t))T ∈ R2n (135)
where y(t) ∈ R and u(t) ∈ R are the measured plant input and output,
are polynomials in s with s being the time differentiation operator s[x](t) = ẋ(t), and
pi , i = 0, 1, . . . , n − 1, k p , and zi , i = 0, 1, . . . , m − 1, with n > m, are the unknown
constant parameters.
Choosing a stable polynomial Λe (s) = sn + λen−1 sn−1 + · · · + λe1 s + λe0 and defin-
ing the parameter and regressor vectors
θ∗p = [k p z0 , k p z1 , . . . , k p zm−1 , k p ,
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 35
Λn−1 (s)
ε(t) = θTp (t)φ(t) − y(t) + [y](t), (143)
Λe (s)
which can be expressed as
It follows that
△
ẏ˜ (t) = ẏˆ (t) − ẏ(t) = θ̃Tp (t)φ̇(t). (151)
From ε(t) = θ̃Tp (t)φ(t), we have
T
ε̇(t) = θ̃˙ p (t)φ(t) + θ̃Tp (t)φ̇(t) (152)
T T T
ε̈(t) = θ̃¨ p (t)φ(t) + θ̃˙ p (t)φ̇(t) + θ̃˙ p (t)φ̇(t) + θ̃Tp (t)φ̈(t). (153)
If an adaptive law without parameter projection for θ p (t):
Γφ(t)ε(t)
θ̇ p (t) = − , Γ = ΓT > 0, α > 0 (154)
1 + αφ (t)φ(t)
T
can achieve the objectives of the asymptotic tracking and signal boundedness of the
MRAC system, then it can be verified that ε̈(t) is bounded and θ̃¨ p (t) is bounded, so
that, with limt→∞ ε(t) = 0, we have limt→∞ ε̇(t) = 0. Then, since limt→∞ θ̃˙ p (t) = 0,
it follows that limt→∞ θ̃Tp (t)φ̇(t) = 0, that is, limt→∞ (ẏˆ (t) − ẏ(t)) = 0.
However, if the parameter projection applied to the adaptive law (154) (for the
estimate of k p , the (m + 1)th element of θ p (t)) is in action, then the (m + 1)th
element of θ̇ p (t) may be discontinuous at the parameter boundary of k p (its lower
bound), so that the (m + 1)th element of θ̈ p (t) may not be bounded. In this case,
ε̈(t) may not be bounded, and limt→∞ ε̇(t) = 0 may not hold. Hence, from ε̇(t) =
T
θ̃˙ p (t)φ(t) + θ̃Tp (t)φ̇(t), we may not conclude limt→∞ θ̃Tp (t)φ̇(t) = 0 for ẏˆ (t) − ẏ(t) =
θ̃T (t)φ̇(t), even if lim
p t→∞ p θ̃˙ (t) = 0 (as lim ε(t) = 0).
t→∞
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 37
(that is, with a fixed σ-modification). For V given in (5.43): V = e2 + |k p |θ̃T Γ−1 θ̃,
we have from (5.204) that
d¯2 (t)
V̇ ≤ −am e2 (t) + − 2|k p |σ0 θ̃T (t)θ(t)
am
d¯2 (t)
= −am e2 (t) + − 2|k p |σ0 θ̃T (t)θ̃(t) + 2|k p |σ0 θ̃T (t)θ∗
am
d¯2 (t)
2
≤ −am e (t) + − |k p |σ0 θ̃T (t)θ̃(t) + |k p |σ0 θ∗T θ∗ . (155)
am
For Γ = γI and am > γ σ0 (that is, with a small σ0 ), we have
d¯2 (t)
V̇ ≤ −γ σ0V + + |k p |σ0 θ∗T θ∗ (156)
am
¯
which, for |d(t)| ≤ d0 , leads to
d02 |k p | ∗T ∗
lim V (t) ≤ + θ θ . (157)
t→∞ am γ σ0 γ
This implies that the upper bound for the tracking error e(t) = y(t) − ym (t), for
|k |
d(t) = d0 = 0 (in the absence of disturbances d(t)), may be as large as γp θ∗T θ∗ ,
independent of σ0 . We already knew that for d(t) = 0 and σ1 (t) = σ0 = 0, the
adaptive control system ensures that limt→∞ e(t) = 0. Now from the above analysis
|k |
we know that for a small σ0 6= 0, the error bound on |e(t)| can be up to γp θ∗T θ∗ .
On the other hand, from
d¯2 (t)
V̇ ≤ −am e2 (t) + − 2|k p |σ0 θ̃T (t)θ(t) (158)
am
and the adaptive control system signal boundedness, similar to (5.204), we have
Z t2
e2 (t) dt ≤ γ0 + k0 (t2 − t1 )d¯02 + c0 (t2 − t1 )σ0 (159)
t1
38 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
for some constant c0 > 0, where the σ0 related term is due to using a fixed σ-
modification σ1 (t) = σ0 , instead of the switching σ-modification which leads to
(5.205) and in turn to (5.204). The inequality (159) implies that, when d0 = 0 (in
the absence of disturbances), we have the mean error
γ0
Z t2
1
e2 (t) dt ≤ + c 0 σ0 (160)
t2 − t1 t1 t2 − t1
which is of the magnitude of c0 σ0 , but the absolute error |e(t)|, as from (157),
|k |
could be as large as γp θ∗T θ∗ , independent of σ0 . This is the so-called “bursting
phenomenon” of robust adaptive control with a fixed σ-modification: the tracking
error e(t) may go to a large value independent of σ0 for a small interval of time
but in the mean sense the error e(t) is of the order of σ0 . This analytically explains
what was observed in the simulation results [134].
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 39
we have
ε2 (t) ε(t)η(t)
V̇ = −2 2
+2 2 − 2σ1 (t)|ρ∗ |θ̃T (t)θ(t) − 2σ2 (t)ρ̃(t)ρ(t)
m (t) m (t)
ε (t)
2 ε(t) η(t) 2 η2 (t)
=− 2 − −µ + µ2 2
m (t) m(t) m(t) m (t)
−2σ1 (t)|ρ∗ |θ̃T (t)θ(t) − 2σ2 (t)ρ̃(t)ρ(t)
ε2 (t) η2 (t)
≤ − 2 + µ2 2 − 2σ1 (t)|ρ∗ |θ̃T (t)θ(t) − 2σ2 (t)ρ̃(t)ρ(t) (162)
m (t) m (t)
|η(t)|
which is (5.224). Since m(t) ≤ b0 for some constant b0 > 0, we have V̇ < 0 if
This implies that there exist constants θ0 > 0 and ρ0 > 0 such that kθ(t)k2 ≥ θ0
or/and |ρ(t)| ≥ ρ0 implies that V̇ < 0. Hence, the boundedness of θ(t) and ρ(t) is
ensured. One choice of such (θ0 , ρ0 ) is
r
1 1
θ = max{2M1 ,
0
(2µ2 b20 + kθ∗ k22 + ρ∗2 ) + kθ∗ k}
4σ 2
r 10
1 1
ρ0 = max{2M2 , (2µ2 b20 + kθ∗ k22 + ρ∗2 ) + |ρ∗ |}. (166)
4σ20 2
40 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
with two unknown parameters a p and b p and choose the reference model
with |am | < 1 for stability. We use the adaptive controller structure
−a p − am ∗ bm
k1∗ = , k2 = . (170)
bp bp
where k̃1 (t) = k1 (t) − k1∗ and k̃2 (t) = k2 (t) − k2∗ .
Defining ρ∗ = b p and
ρ∗
e(t) = [θT ω − θ∗T ω](t)
z + am
1
=ρ∗
[θ ω](t) − θ ζ(t) .
T ∗T
(175)
z + am
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 41
The design task is to find adaptive laws to update the parameter estimates θ(t) and
ρ(t) (which is an estimate of ρ∗ ) such that the estimation error
1
ε(t) = e(t) − ρ(t) [θ ω](t) − θ (t)ζ(t)
T T
(176)
z + am
where
1
ξ(t) = θT (t)ζ(t) − [θT ω](t) (178)
z + am
θ̃(t) = θ(t) − θ∗ , ρ̃(t) = ρ(t) − ρ∗ . (179)
We choose the gradient adaptive laws for θ(t) and ρ(t):
sign[b p ]Γε(t)ζ(t) 2
θ(t + 1) = θ(t) − 2
, 0 < Γ = ΓT < 0 I2 , (180)
m (t) bp
γ ε(t)ξ(t)
ρ(t + 1) = ρ(t) − , 0 < γ < 2, (181)
m2 (t)
where G(s) is an M × M strictly proper rational transfer matrix which can be ex-
pressed as
G(s) = D−1 (s)N(s) = Cg (sI − Ag )−1 Bg , (184)
where L ∈ RMν×M is such that A − LCg is a desired stable matrix. For the state
estimation error x̃(t) = x̂(t) − x(t), it follows that
0
x̃˙ (t) = (A − LCg )x̃(t) + Ã p (t)y(t) + u(t), ỹ(t) = Cg x̃(t) (194)
B̃ p (t)
Adaptive control law. The adaptive control law is an indirect-design (that is,
its parameters are calculated from plant parameter estimates) and explicit-observer
(that is, it uses the state estimate x̂(t) for feedback control) based control law:
where K1 (t) ∈ RMν×M and K2 (t) ∈ RM× are parameter matrices which satisfy
point-wise in the time variable t, where Âg (t) and B̂g (t) are the on-line estimates
of Ag and Bg . It can be shown that such solutions K1 and K2 exist if the estimated
plant (Âg , B̂g ,Cg ) has ξm (s) as its modified left interactor matrix point-wise, that is,
limλ→∞ ξm (λ)Cg (λI − Âg )−1 B̂g = K̂ p is finite and non-singular, for each time t as
Âg = Â p (t) and B̂g = B̂ p (t) (note that K̂ p may be a function of time t).
44 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
are known with Am being stable, for constructing a good reference model system
Then, for K1 , K2 and K3 being the estimates of K1∗ , K2∗ and K3∗ , we use the adaptive
controller
u(t) = K1T (t)x(t) + K2 (t)r(t) + K3 (t)φ(x). (201)
From the definitions of K1∗ , K2∗ and K3∗ , we have
u(t) = K̃1T (t)x(t) + K̃2 (t)r(t) + K̃3 (t)φ(x) + K1∗T x(t) + K2∗ r(t) + K3∗ φ(x) (204)
Then, in view of this equation and (199), the tracking error e(t) = x(t) − xm (t)
satisfies
ė(t) = Am e(t) + BΛ(K̃1T (t)x(t) + K̃2 (t)r(t) + K̃3 (t)φ(x)). (206)
Introducing θ∗i such that θ∗T ∗T ∗ ∗
i is the ith row of [K1 , K2 , K3 ], i = 1, 2, . . . , M, letting
θi be the estimate of θ∗i and θ̃i (t) = θi (t) − θ∗i , i = 1, 2, . . . , M, and defining
we express (206) as
θ̃T1 (t)ω(t)
θ̃T2 (t)ω(t)
..
ė(t) = Am e(t) + BΛ . . (208)
T
θ̃M−1 (t)ω(t)
θ̃TM (t)ω(t)
for a chosen constant matrix Q ∈ Rn×n such that Q = QT > 0, and ēi (t) be the ith
component of eT (t)PB, i = 1, 2, . . . , M, we design the adaptive law for θi (t) as
where Γi = ΓTi > 0 is a chosen constant adaptation gain matrix, and sign[λi ] is the
sign of λi , i = 1, 2, . . . , M.
To analyze the adaptive control system performance, we consider the positive
definite function
M
V (e, θ̃i , i = 1, 2, . . . , M) = eT Pe + ∑ |λi |θ̃Ti Γ−1
i θ̃i (211)
i=1
and derive its time derivative along the trajectory of (208) and (210) as
M
V̇ = 2e (t)Pė(t) + 2 ∑ |λi |θ̃Ti (t)Γ−1
T
i θ̇i (t)
i=1
M M
= 2eT (t)PAm e(t) + 2 ∑ ēi (t)λi θ̃Ti (t)ω(t) + 2 ∑ |λi |θ̃Ti (t)Γ−1
i θ̇i (t)
i=1 i=1
T
= −e (t)Qe(t). (212)
From this result, it follows that x(t) and θi (t) are bounded and e(t) ∈ L2 , and in
turn, that u(t) is bounded, and so is ė(t), so that limt→∞ e(t) = 0.
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 47
Θ∗3ξ(z)y(t)
For multivariable MRAC, the parameters Θ∗1 , Θ∗2 , Θ∗20 and Θ∗3 of the matching equa-
tion
u(t) = Θ∗T
1 ω1 (t) + Θ2 ω2 (t) + Θ20 y(t) + Θ3 ξ(z)y(t)
∗T ∗ ∗
(213)
are the nominal controller parameters. For discrete-time MRAC where Λ(z) = znc
for generating ω1 (t) and ω2 (t), such parameters can be calculated from the follow-
ing procedure.
For a given system transfer matrix G(z) = A−1 (z−1 )B(z−1 ) and an interactor
matrix ξ(z) such that rlimz→∞ ξ(z)G(z) = K p finite and nonsingular, where
ξ0 zd + ξ1 zd−1 + · · · + ξd−1 z
= (F0 zd + F1 zd−1 + · · · + Fd−1 z)(I + A1 z−1 + · · · + Ana z−na )
+H0 + H1 z−1 + · · · + Hnh z−nh . (219)
F0 = ξ0 , F1 + F0 A1 = ξ1 , . . . (220)
with αi = Hi , and
β(z−1 ) = F(z)B(z−1 ). (222)
From ξ(z) = F(z)A(z−1 ) + H(z−1 ), we have
ξ(z) − H(z−1 ) = F(z)B(z−1 )B−1 (z−1 )A(z−1 ) = β(z−1 )G−1 (z) (223)
that is,
G(z) = (ξ(z) − α(z−1 ))−1 β(z−1 ). (224)
Since limz→∞ ξ(z)G(z) = K p finite and nonsingular, we have
ξ(z)y(t) = Θ∗−1
3 u(t) − Θ3 Θ1 ω1 (t) − Θ3 Θ2 ω2 (t) − Θ3 Θ20 y(t)
∗−1 ∗T ∗−1 ∗T ∗−1 ∗
(228)
with Λ(z) = znc , we can obtain Θ∗1 , Θ∗2 , Θ∗20 and Θ∗3 = K p−1 , from α(z−1 ) and β(z−1 ).
For G(z) in this example, it can be found the system order is n = 5 and its
observability index ν ≤ n−1 = 4, so that the order of Λ(z) in (227) is nc = ν−1 = 3,
which means ω1 (t) depends on u(t − 1), u(t − 2) and u(t − 3), as contained in the
term β(z−1 )u(t).
Finally, with α(z−1 ) and β(z−1 ) specified, from the matching equations (227)
and (228):
ξ(z)y(t) = α(z−1 )y(t) + β(z−1 )u(t) (240)
Since the term α(z−1 )y(t) does not contain y(−2) and y(−3), α3 = α2 = 0 (the
2 × 2 zero matrix), the first 4 × 2 part of Θ∗2 ∈ R6×2 is the 4 × 2 zero matrix. 2
50 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
Parametrization of (9.102)
To define the parameters Θ∗i , i = 1, 2, 20, 3, to satisfy (9.105):
Θ∗T
1 A(D)P0 (D) + (Θ2 A(D) + Θ20 Λ(D))Z0 (D)
∗T ∗
for the nominal version of the multivariable MRAC controller (9.102), we consider
an equivalent version (see (244) below) of (9.105) by dividing it from the right by
P0−1 (D) and use the left matrix-fraction description of G0 (D) = Pl−1 (D)Zl (D) such
that G0 (D) = Pl−1 (D)Zl (D) = Z0 (D)P0−1 (D) (as used in the proof of Lemma 9.3),
to obtain
Θ∗T
1 A(D) + (Θ2 A(D) + Θ20 Λ(D))Pl (D)Zl (D)
∗T ∗ −1
for some M × M polynomial matrices Ql (D) and Rl (D) such that ∂ci [Rl (D)] <
∂ci [Pl (D)] ≤ ν. Then, we define Θ∗1 , Θ∗2 and Θ∗20 from
Θ∗T
2 A(D) + Θ20 Λ(D) = −Rl (D),
∗
(246)
Θ∗T
1 A(D) = Λ(D)IM − Ql (D)Zl (D) (247)
(recall that ∂[Λ(D)] = ν−1 and ∂[A(D)] = ν−2 for the controller structure (9.102)).
With this definition of Θ∗2 , Θ∗20 and Θ∗1 , (244) is satisfied, and so is (9.105).
From (9.105) (that is, (244)), we have the plant–model matching equation
IM − Θ∗T
1 F(D) − (Θ2 F(D) + Θ20 )G0 (D) = Θ3Wm (D)G0 (D)
∗T ∗ ∗ −1
(248)
where Wm (D) = ξ−1 m (D). From this equation with limD→∞ Θ3Wm (D)G0 (D) = IM
∗ −1
lim Θ∗T
1 F(D) = 0, (249)
D→∞
Θ∗T
1 A(D)P0 (D) + Θ2 A(D)Z0 (D) = Λ(D)(P0 (D) − Θ3 ξm (D)Z0 (D)),
∗T ∗
(250)
IM − Θ∗T
1 F(D) − Θ2 F(D)G0 (D) = Θ3Wm (D)G0 (D).
∗T ∗ −1
(251)
Operating this identity on u(t), we may get the plant signal identity:
(This identity was used in deriving (9.119), and its similar version with Θ∗20 was
used in deriving (9.365).)
As an alternative procedure to obtain the plant parametrized signal identity (sim-
ilar to that in (5.30) for the SISO case with M = 1), we start by considering (9.84)
and (9.85):
Θ∗T
2 A(D) = Ql (D)Pl (D) − Λ(D)K p ξm (D)
−1
(253)
Θ∗T
1 A(D) = Λ(D)IM − Ql (D)Zl (D). (254)
u(t) = Θ∗T
1 ω1 (t) + Θ2 ω2 (t) + Θ20 y(t) + Θ3 r(t),
∗T ∗ ∗
(258)
Θ∗T
1 A(D)P0 (D) + (Θ2 A(D) + Θ20 Λ(D))Z0 (D)
∗T ∗
IM − Θ∗T
1 F(D) − (Θ2 F(D) + Θ20 )G0 (D) = Θ3Wm (D)G0 (D).
∗T ∗ ∗ −1
(260)
The nominal parameters Θ∗1 , Θ∗2 , Θ∗20 and Θ∗3 are defined from
2 A(D) + Θ20 Λ(D)) = −Rl (D) = Ql (D)Pl (D) − Λ(D)K p ξm (D) (261)
(Θ∗T ∗ −1
Θ∗T
1 A(D) = Λ(D)IM − Ql (D)Zl (D), (262)
that is, dividing Λ(D)K p−1 ξm (D) on the right by Pl (D) to get Rl (D) and Ql (D).
From these equations, we obtain the signal identities:
A(D) 1
(Θ∗T + Θ∗20 )[y](t) = Ql (D)Pl (D)[y](t) − K p−1 ξm (D)[y](t) (263)
2
Λ(D) Λ(D)
A(D) 1
Θ∗T [u](t) = u(t) − Ql (D)Zl (D)[u](t). (264)
1
Λ(D) Λ(D)
Using the open-loop plant signal identity: Pl (D)[y](t) = Zl (D)[u](t), we have the
plant signal identity in a feedback form:
A(D) A(D)
u(t) = Θ∗T [u](t) + Θ∗T [y](t) + Θ∗20 y(t) + K p−1 ξm (D)[y](t) (265)
1
Λ(D) 2
Λ(D)
which also holds for any input signal u(t), similar to that in (257).
Both parametrized plant signal identities (257) and (265) are useful for adaptive
control: either verify the nominal controller structure
A(D) A(D)
u(t) = Θ∗T [u](t) + Θ∗T [y](t) + Θ∗3 r(t) (266)
1
Λ(D) 2
Λ(D)
for the matching equation (250), or
A(D) A(D)
u(t) = Θ∗T [u](t) + Θ∗T [y](t) + Θ∗20 y(t) + Θ∗3 r(t) (267)
1
Λ(D) 2
Λ(D)
for the matching equation (259), where Θ∗3 r(t) = K p−1 ξm (D)[ym ](t) for the reference
output ym (t) = ξ−1
m (D)[r](t), leading to ξm (D)[y − ym ](t) = 0 exponentially, or they
motivate the adaptive controller structure
A(D) A(D)
u(t) = ΘT1 [u](t) + ΘT2 [y](t) + Θ3 r(t) (268)
Λ(D) Λ(D)
for the matching equation (250), or
A(D) A(D)
u(t) = ΘT1 [u](t) + ΘT2 [y](t) + Θ20 y(t) + Θ3 r(t) (269)
Λ(D) Λ(D)
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 53
for the matching equation (259), leading to the desired tracking error equation:
where θ∗0 is a constant vector and δθ∗ (t) is the variation of θ∗ (t) with respect to
θ∗0 (note that both θ∗0 and δθ∗ (t) are unknown). We will develop and analyze some
alternative adaptive control schemes to that presented in (9.607)–(9.608).
u(t) = Y (q, qd , q̇, q̇d , q̈d ,t)θ0 (t) − m(t)ψ(t) − m1 (t)ψ1 (t) − KD s(t), (278)
m1 (t) = k1 kY (q, qd , q̇, q̇d , q̈d ,t)k, k1 > 0, ψ1 (t) = m1 (t)s(t), (280)
where σ(t) is a switching signal similar to that in (9.610), using a design parameter
σ0 > 0 and the knowledge of the upper bound M0 on kθ∗0 k:
0
if kθ0 (t)k < M0 ,
σ(t) = σ0 ( kθM
0 (t)k
− 1) if M0 ≤ kθ0 (t)k < 2M0 , (282)
0
σ0 if kθ0 (t)k ≥ 2M0 .
56 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
This adaptive control scheme has the properties: all signals in the closed-loop
system are bounded, and the tracking error e(t) = q(t) − qd (t) satisfies
2
γ γ21
Z t2
ke(t)k dt ≤ α0 2 + 2 (t2 − t1 ) + β0
2
(283)
t1 k0 k1
for some constants α0 > 0, β0 > 0 and any t2 > t1 ≥ 0, where γ1 > 0 is the upper
bound on supt≥0 kδθ∗ (t)k. Moreover, e(t) ∈ L2 and limt→∞ e(t) = 0 in the absence
of parameter time variations, that is, when δθ∗ (t) = 0 and ∂D(q,t)
∂t = 0.
The proof of these properties is based on the positive definite function
1
V0 (s, θ̃0 ) = (sT Ds + θ̃T0 Γ−1 θ̃0 ), θ̃0 (t) = θ0 (t) − θ∗0 , D = D(q(t),t), (284)
2
which has the following time derivation:
V̇0 = −sT (t)KD s(t) − m2 (t)sT (t)s(t) − m21 (t)sT (t)s(t) − σ(t)θ̃T0 (t)θ0 (t)
1 ∂D(q,t)
− sT (t) (q̇(t) + v(t)) − sT (t)Y (q, qd , q̇, q̇d , q̈d ,t)δθ∗ (t)
2 ∂t
T γ 2 γ2
≤ −s (t)KD s(t) − m(t)ks(t)k − +
4k0 16k02
γ1 2 γ21
− m1 (t)ks(t)k − + 2 − σ(t)θ̃T0 (t)θ0 (t). (285)
2k1 4k1
With this adaptive control scheme, as indicated by (283), the tracking perfor-
mance can be influenced by the design parameters k0 and k1 in the feedback control
law (278)–(280) (one may increase k0 and k1 to reduce the tracking error e(t)).
|gi (q, q̇, qd , q̇d ,t)| ≤ a∗i αi (q, q̇, qd , q̇d ,t), i = 1, 2, . . . , n (288)
|hi (q, q̇, qd , q̇d , q̈d ,t)| ≤ b∗i βi (q, q̇, qd , q̇d , q̈d ,t), i = 1, 2, . . . , n (289)
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 57
for some unknown constants a∗i and b∗i , and known functions αi (q, q̇, qd , q̇d ,t) and
βi (q, q̇, qd , q̇d , q̈d ,t), i = 1, 2, . . . , n.
If the parameters a∗i and b∗i were known, one could use the control law
u(t) = Y (q, qd , q̇, q̇d , q̈d ,t)θ0 (t) − φ∗ (t) − φ∗1 (t) − KD s(t), (290)
φ(t) = [sgn[s1 (t)]a1 (t)α1 , sgn[s2 (t)]a2 (t)α2 , . . . , sgn[sn (t)]an (t)αn ]T , (296)
φ1 (t) = [sgn[s1 (t)]b1 (t)β1 , sgn[s2 (t)]b2 (t)β2 , . . . , sgn[sn (t)]bn (t)βn ]T , (297)
where θ0 (t) is updated from (281), and the parameters ai (t) and bi (t) are estimates
of a∗i and b∗i and updated from the adaptive laws:
ȧi (t) = κai |si (t)|αi (q, q̇, qd , q̇d ,t), κai > 0, i = 1, 2, . . . , n, (298)
ḃi (t) = κbi |si (t)|βi (q, q̇, qd , q̇d , q̈d ,t), κbi > 0, i = 1, 2, . . . , n. (299)
58 Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023)
where θ̃0 (t) = θ0 (t) − θ∗0 , D = D(q(t),t), ãi (t) = ai (t) − a∗i , b̃i (t) = bi (t) − b∗i , i =
1, 2, . . . , n. Using (276), (295), (281), (298) and (299), we have
This result also implies that all signals in the closed-loop system are bounded, and
the tracking error e(t) = q(t) − qd (t) converges to zero as t goes to ∞.
However, since the adaptive control scheme (295) uses the sgn functions in φ(t)
and φ1 (t) and such switching signals are discontinuous when si (t) passes through
zero, it may lead to chattering of system response.
Remark 5 The adaptive control scheme (295) may have certain advantage for per-
formance even if when the parameters a∗i and b∗i are known. This is because the
parameters a∗i and b∗i are only the upper bounds for the parameter variation uncer-
tainties gi and hi in (286) and (287), and some smaller (and unknown) bounds may
exist and can be estimated by the adaptive laws (298) and (299). The use of smaller
bounds is desirable because it leads to smaller control signals. In this case, the
adaptive laws (298) and (299) can be modified by setting
u(t) = Y (q, qd , q̇, q̇d , q̈d ,t)θ0 (t) − φ̂(t) − φ̂1 (t) − KD s(t), (304)
φ̂(t) = [sat[s1 (t); ε1 ]a1 (t)α1 , . . . , sat[sn (t); εn ]an (t)αn ]T , (305)
φ̂1 (t) = [sat[s1 (t); η1 ]b1 (t)β1 , . . . , sat[sn (t); ηn ]bn (t)βn ]T , (306)
Such functions have the property: χ[si ; xi ](1 − χ[si ; xi ]) = 0 (that is, χ[si ; xi ] = 0
whenever 1 − χ[si ; xi ] = 1, and χ[si ; xi ] = 1 whenever 1 − χ[si ; xi ] = 0).
The adaptive laws are also modified as
ȧi (t) = χ[si ; εi ]κai |si (t)|αi (q, q̇, qd , q̇d ,t), κai > 0, i = 1, 2, . . . , n, (309)
ḃi (t) = χ[si ; ηi ]κbi |si (t)|βi (q, q̇, qd , q̇d , q̈d ,t), κbi > 0, i = 1, 2, . . . , n. (310)
n n
ai ȧi + ∑ b̃i κbi ḃi − σ(t)θ̃0 (t)θ0 (t)
− ∑ si (t)gi − ∑ si (t)hi + ∑ ãi κ−1 −1 T
i=1 i=1 i=1 i=1
n n
≤ −sT (t)KD s(t) − ∑ si (t)sat[si ; εi ]ai (t)αi − ∑ si (t)sat[si ; ηi ]bi (t)βi
i=1 i=1
n n
ai ȧi + ∑ b̃i κbi ḃi − σ(t)θ̃0 (t)θ0 (t)
+ ∑ |si (t)|a∗i αi + ∑ |si (t)|b∗i βi + ∑ ãi κ−1 −1 T
i=1 i=1 i=1 i=1
n
= −sT (t)KD s(t) − ∑ (1 − χ[si ; εi ])si (t)sat[si ; εi ]ai (t)αi
i=1
n n
− ∑ (1 − χ[si ; ηi ])si (t)sat[si ; ηi ]bi (t)βi + ∑ (1 − χ[si ; εi ])|si (t)|a∗i αi
i=1 i=1
n
+ ∑ (1 − χ[si ; ηi ])|si (t)|b∗i βi − σ(t)θ̃T0 (t)θ0 (t). (311)
i=1
This modified scheme would ensure the closed-loop signal boundedness but not
the asymptotic convergence of the tracking error e(t) = q(t) − qd (t) to zero (only a
bounded tracking error e(t) = q(t) − qd (t) of the order εi and ηi ).
Notes on Adaptive Systems and Control, Gang Tao (Copyright 2023) 61
Derivation of (10.152)
In this case, the expression (5.30) also holds
a(s) a(s)
u(t) = φ∗T [u](t) + φ∗T [y](t) + φ∗20 y(t) + φ∗3 Pm (s)[y](t) (312)
1
Λ(s) 2
Λ(s)
(with θ∗i replaced by φ∗i as the new notation and the exponentially decaying term
ε1 (t) ignored). Recall (10.39):
This calculation can be verified by using the symbolic algebra operations in Matlab.
For example, for n = 5, we can use