EECE 574 - Adaptive Control

EECE 574 - Adaptive Control
Model-Reference Adaptive Control - An Overview
Guy Dumont
Department of Electrical and Computer Engineering

University of British Columbia
January 2013
Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 1 / 73

Model-Reference Adaptive Systems
The MRAC or MRAS is an important adaptive control methodology 1
1 see Chapter 5 of the Åström and Wittenmark textbook, or H. Butler,
”Model-Reference Adaptive Control-From Theory to Practice”, Prentice-Hall, 1992

Model-Reference Adaptive Systems
The MIT rule

Lyapunov stability theory
Design of MRAS based on Lyapunov stability theory
Hyperstability and passivity theory
The error model
Augmented error
A model-following MRAS

MIT Rule The Basics
The MIT Rule
Original approach to MRAC developed around 1960 at MIT for

aerospace applications
With e = y − ym , adjust the parameters θ to minimize
1
J(θ ) = e2
2
It is reasonable to adjust the parameters in the direction of the negative
gradient of J:
dθ ∂J ∂e
= −γ = −γe
dt ∂θ ∂θ
∂ e/∂ θ is called the sensitivity derivative of the system and is evaluated
under the assumption that θ varies slowly

MIT Rule The Basics
The MIT Rule
The derivative of J is then described by

2
dJ ∂e ∂e
=e = −γe2
dt ∂t ∂θ
Alternatively, one may consider J(e) = |e| in which case
dθ ∂J ∂e
= −γ = −γ sign(e)
dt ∂θ ∂θ
The sign-sign algorithm used in telecommunications where simple
implementation and fast computations are required, is

dθ ∂J ∂e
= −γ = −γsign sign(e)
dt ∂θ ∂θ

MIT Rule Examples
MIT Rule: Example 1
Process: y = kG(s) where G(s) is known but k is unknown

The desired response is ym = k0 G(s)uc
Controller is u = θ uc
Then e = y − ym = kG(p)θ uc − k0 G(p)uc
Sensitivity derivative
∂e k
= kG(p)uc = ym
∂θ k0
MIT rule
dθ k
= γ 0 ym e = −γym e
dt k0

MIT Rule Examples
MIT Rule: Example 1
Figure: MIT rule for adjustment of feedforward gain (from textbook).

MIT Rule Examples
MIT Rule: Example 1
Figure: MIT rule for adjustment of feedforward gain: Simulation results (from
textbook).

MIT Rule Examples
MIT Rule: Example 2
Consider the first-order system

dy
= −ay + bu
dt
The desired closed-loop system is
dym
= −am ym + bm uc
dt
Applying model-following design (see lecture notes on pole placement)
deg A0 ≥ 0 deg S = deg R = deg T = 0

MIT Rule Examples
MIT Rule: Example 2
The controller is then

u(t) = t0 uc (t) − s0 y(t)
For perfect model-following
dy
= −ay(t) + b[t0 uc (t) − s0 y(t)]
dt
= −(a + bs0 )y(t) + bt0 uc
= −am ym (t) + bm uc
This implies
am − a
s0 =
b
bm
t0 =
b

MIT Rule Examples
MIT Rule: Example 2
With the controller we can write2

bt0
y= uc
p + a + bs0
With e = y − ym , the sensitivity derivatives are
∂e b
= uc
∂ t0 p + a + bs0
∂e b 2 t0 b
= 2
uc = y
∂ s0 (p + a + bs0 ) p + a + bs0
2p is the differential operator d(.)/dt

MIT Rule Examples
MIT Rule: Example 2
However, a and b are unknown3 . For the nominal parameters, we know that
a + bs0 = am
p + a + bs0 ≈ p + am
Then

dt0 0 1 am
= −γ b uc e = −γ uc e
dt p + am p + am

ds0 0 1 am
= γb y e=γ y e
dt p + am p + am
with γ = γ 0 b/am , i.e. b is absorbed in γ and the filter is normalized with static
gain of one.
3 after all, we want to design an adaptive controller!
MIT Rule Examples
MIT Rule: Example 2
Figure: MIT rule for first-order (from textbook)

MIT Rule Examples
MIT Rule: Example 2
Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2 and

γ = 1(from textbook)

MIT Rule Examples
MIT Rule: Example 2
Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2.

Controller parameters for γ = 0.2, 1 and 5(from textbook)

MIT Rule Examples
MIT Rule: Example 2
Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2.

Relation between controller parameters θ1 and θ2 . The dashed line is θ2 = θ1 − a/b.
(from textbook)

MIT Rule Stability of the Adaptive Loop
Stability of the Adaptive Loop
Consider again the adaptation of a feedforward gain

1
G=
s2 + a1 s + a2
e = (θ − θ0 )Guc
∂e ym
= Guc =
∂θ θ0
dθ ∂e ym
= −γ 0 e = −γ 0 e = −γeym
dt ∂θ θ0

Thus the system can be described as
d 2 ym dym
+ a1 + a2 ym = θ0 uc
dt2 dt
d2 y dy
+ a1 + a2 y = θ uc
dt2 dt
dθ
= −γ(y − ym )ym
dt
This is difficult to solve analytically

d3 y d2 y dy duc
3
+ a1 2
+ a2 + γuc ym y = θ + γuc y2m
dt dt dt dt
Assume uoc and yom constant, equilibrium is
θ0 uoc
y(t) = yom =
a2
which is stable if
γ o 2
a1 a2 > γuoc yom = (u )
a2 c
Thus, if γ or uc is sufficiently large, the system will be unstable

Figure: Influence on convergence and stability of signal amplitude for MIT rule
(from textbook)

MIT Rule Modified MIT Rule
Modified MIT Rule
Normalization can be used to protect against dependence on the signal

amplitudes
With ϕ = ∂ e/∂ θ , the MIT rule can be written as
dθ
= −γϕe
dt
The normalized MIT rule is then
dθ −γϕe
= 2
dt α + ϕTϕ
For the previous example,
dθ −γeym /θ0
= 2
dt α + y2m /θ02

MIT Rule Modified MIT Rule
Modified MIT Rule
Figure: Influence on convergence and stability of signal amplitude for modified MIT
rule (from textbook)

Lyapunov Design of MRAC Lyapunov Theory
Lyapunov Stability Theory
Aleksandr M Lyapunov
(1857-1918, Russia)
Classmate of Markov, taught by
Chebyshev
Doctoral thesis The general
problem of the stability of motion
became a fundamental
contribution to the study of
dynamic systems stability
He shot himself three days after
his wife died of tuberculosis

Consider the nonlinear time-varying system
ẋ = f (x, t) with f (0, t) = 0
If at time t = t0 kx(0)k = δ , i.e. if the initial state is not the equlibrium

state x∗ = 0, what will happen?
There are four possibilities

The system is stable. For a sufficiently small δ , x stays within ε of x∗ . If

δ can be chosen independently of t0 , then the system is said to be
uniformly stable.
The system is unstable. It is possible to find ε which does not allow any
δ.
The system is asymptotically stable. For δ < R and an arbitrary ε, there
exists t∗ such that for all t > t∗ , kx − x∗ k < ε. It implies that kx − x∗ k → 0
as t → ∞.
If asymptotic stability is guaranteed for any δ , the system is globally
asymptotically stable.

A scalar time-varying function is locally positive definite if V(o, t) = 0

and there exists a time-invariant positive definite function V0 (x) such that
∀t ≥ t0 , V(x, t) ≥ V0 (x). In other words, it dominates a time-invariant
positive definite function.
V(x, t) is radially unbounded if 0 < αkxk ≤ V(x, t), α > 0.
A scalar time-varying function is said to be decrescent if V(o, t) = 0 and
there exists a time-invariant positive definite function V1 (x) such that
∀t ≥ t0 , V(x, t) ≤ V1 (x).
In other words, it is dominated by a time-invariant positive definite
function.

Theorem (Lyapunov Theorem)

Stability: if in a ball BR around the equilibrium point 0, there exists a scalar function
V(x, t) with continuous partial derivatives such that
1 V is positive definite
2 V̇ is negative semi-definite
then the equilibrium point is stable.
Uniform stability and uniform asymptotic stability: If furthermore,
3 V is decrescent
then the origin is uniformly stable. If condition 2 is strengthened by requiring that V̇ be
negative definite, then the equilibrium is uniformly asymptotically stable.
Global uniform asymptotic stability: If BR is replaced by the whole state space, and
condition 1, the strengthened condition 2, condition 3 and the condition
4 V(x,t) is radially unbounded
are all satisfied, then the equilibrium point at 0 is globally uniformly asymptotically
stable.

The Lyapunov function has similarities with the energy content of the system and must
be decreasing with time.
This result can be used in the following way to design a stable a stable adaptive
controller
1 First, the error equation, i.e. a differential equation describing either the output
error y − ym or the state error x − xm is derived.
2 Second, a Lyapunov function, function of both the signal error e = x − xm and the
parameter error φ = θ − θm is chosen. A typical choice is
V = eT Pe + φ T Γ−1 φ
where both P and Γ−1 are positive definite matrices.

3 The time derivative of V is calculated. Typically, it will have the form
V̇ = −eT Qe + some terms including φ
Putting the extra terms to zero will ensure negative definiteness for V̇ if Q is
positive definite, and will provide the adaptive law.

Lyapunov Design of MRAC Lyapunov Design of MRAC
Stability of Adaptive Loop

Lyapunov Design of MRAC
Determine controller structure

Derive the error equation
Find a Lyapunov equation
Determine adaptation law that satisfies Lyapunov theorem

Adaptation of Feedforward Gain
Process model
dy
= −ay + ku
dt
Desired response
dym
= −aym + k0 uc
dt
Controller
u = θ uc
Introduce error e = y − ym
The error equation is
de
= −ae + (kθ − k0 )uc
dt

System model
dy
= −ay + ku
dt
dθ
= ??
dt
Desired equilibrium
e = 0
k0
θ = θ0 =
k

Consider the Lyapunov function

γ 2 k
V(e, θ ) = e + (θ − θ0 )2
2 2
dV dθ
= −γae2 + (kθ − k0 )( + γuc e)
dt dt
Choosing the adjustment rule
dθ
= −γuc e
dt
gives
dV
= −γae2
dt

dθ
Lyapunov rule: dt = −γuc e MIT rule: dθ
dt = −γym e

Lyapunov rule: MIT rule:

First-order System
Process model
dy
= −ay + bu
dt
Desired response
dym
= −am ym + bm uc
dt
Controller
u = θ1 uc − θ2 y
Introduce error e = y − ym
The error equation is
de
= −am e − (bθ2 + a − am )y + (bθ1 − bm )uc
dt

First-order System
Candidate Lyapunov function

1 2 1 2 1 2
V(t, θ1 , θ2 ) = e + (bθ2 + a − am ) + (bθ1 − bm )
2 bγ bγ
Derivative
dV de 1 dθ 1 dθ
= + (bθ2 + a − am ) 2 + (bθ1 − bm ) 1
dt dt γ dt γ dt
1 dθ 2 1 dθ
= −am e2 + (bθ2 + a − am )( − γye) + (bθ1 − bm )( 1 + γuc e)
γ dt γ dt
This suggests the adaptation law

dθ1
= −γuc e
dt
dθ2
= γye
dt

First-order System
Lyapunov Rule MIT Rule

First-order System

Lyapunov-based MRAC Design
The main advantage of Lyapunov design is that it guarantees a

closed-loop system.
For a linear, asymptotically stable governed by a matrix A, a positive
symmetric matrix Q yields a positive symmetric matrix P by the equation
AT P + PA = −Q
This equation is known as Lyapunov’s equation.

Lyapunov-based MRAC Design
The main drawback of Lyapunov design is that there is no systematic

way of finding a suitable Lyapunov function V leading to a specific
adaptive law.
For example, if one wants to add a proportional term to the adaptive law,
it is not trivial to find the corresponding Lyapunov function.
The hyperstability approach is more flexible than the Lyapunov
approach.

Lyapunov Design of MRAC State Feedback
The Lyapunov Equation
Let the linear system

ẋ = AX
be stable
Let Q be an arbitrary positive definite matrix
Then the Lyapunov equation
AT P + PA = −Q
always has a unique solution where P is positive definite.

The function
V(x) = xT Px
is then a Lyapunov function

State Feedback
Process: ẋ = Ax + Bu
Desired response: ẋm = Am xm + Bm uc
Control law: u = Muc − Lx
Closed-loop system
ẋ = Ax + Bu = (A − BL)x + BMuc = Ac (θ )x + Bc (θ )uc
The parametrization is Ac (θ ) = Am and Bc (θ ) = Bm

For compatibility we need A − Am = BL and Bm = BM

Error Equation
The error e = x − xm satisfies
ė = ẋ − ẋm = Ax + Bu − Axm − Bm uc
Hence
ė = Am e + (A − Am − BL)x + (BM − Bm )uc

= Am e + (Ac (θ ) − Am )x + (Bc (θ ) − Bm )uc
= Am e + Ψ(θ − θ0 )

The Lyapunov Function
Try
1
γeT Pe + (θ − θ0 )T (θ − θ0 )

V(e, θ ) =
2
Differentiating V
γ
V̇ = (ėT Pe + eT Pė) + (θ − θ0 )T θ̇
2
Then with ė = Am e + Ψ(θ − θ0 ) we can write
γ T T
e Am + Ψ(θ − θ0 ) Pe + eT (Am e + Ψ(θ − θ0 )

V̇ =
2
+(θ − θ0 )T θ̇

The Lyapunov Function
Now, consider the Lyapunov equation ATm P + PAm = −Q (possible

because reference model always stable)
γ
V̇ = − eT Qe + γ(θ − θ0 )T ΨT Pe + (θ − θ0 )T θ̇
2
γ
= − eT Qe + (θ − θ0 )T (θ̇ + γΨT Pe)
2
If the parameter adjustment law is chosen as
dθ
= −γΨT Pe
dt
We obtain
dV γ
= − eT Qe
dt 2
which is negative semi-definite
Note that this assumes knowledge of the state.
Lyapunov Design of MRAC Adaptation of Feedforward Gain
Consider the error
e = (kG(p)θ − k0 G(p))uc = kG(p)(θ − θ0 )uc
With
ẋ = Ax + B(θ − θ0 )uc
e = Cx
A candidate Lyapunov function is

1
V = (γxT Px + (θ − θ0 )2 )
2
Let
AT P + PA = −Q

Differentiating V
γ T
ẋ Px + xT Pẋ + (θ − θ0 )θ̇

V̇ =
2
Which after using the model for ẋ and the Lyapunov equation gives
γ
V̇ = − xT Qx + (θ − θ0 ) θ̇ + γuc BT Px

2
Using the adaptation law
dθ
= −γuc BT Px
dt
gives
dV γ
= − xT Qx
dt 2

Output Feedback
The adaptation law

dθ
= −γuc BT Px
dt
assumes knowledge of the state x
If we can find P such that
BT P = C
then the adaptation law becomes
dθ
= −γuc e
dt
i.e. this is now output feedback and the state x is not required
When is it possible to do so?

Lyapunov Design of MRAC Kalman-Yakubovich Lemma
Kalman-Yakubovich Lemma
Definition: Positive real transfer function

A rational transfer function G with real coefficients is positive real (PR) if
Re G(s) ≥ 0 for Re s ≥ 0
A rational transfer function G with real coefficients is strictly positive real

(SPR) if G(s − ε) is PR for some ε > 0
G(s) = 1/(s + 1) is SPR
G(s) = 1/s is PR bur not SPR

Lyapunov Design of MRAC Kalman-Yakubovich Lemma
Kalman-Yakubovich Lemma
Theorem (Kalman-Yakubovich Lemma)

Let the linear time-invariant system
ẋ = Ax + Bu
y = Cx
be completely controllable and completely observable. The transfer function
G(s) = C(sI − A)−1 B
is strictly positive real if and only if there exist positive definite matrices P and
Q such that
AT P + PA = −Q
BT P = C

The Small Gain Theorem Definitions
Operator View of Dynamical Systems
L2 norm
Z ∞
21
2
kuk = u (t)dt
−∞
L∞ norm
kuk = sup |u(t)|

0≤t<∞
Truncation operator:

x(t) 0 ≤ t ≤ T
xT (t) =
0 t>T
L2e = x | kxT k2 = hxT |xT i < ∞


The Notion of Gain
Gain of a nonlinear system

Let the signal space be Xe . The gain γ(S) of a system S is defined as
kSuk
γ(S) = sup
u∈Xe kuk
where u is the input signal to the system.

The gain is thus the smallest value γ such that
kSuk ≤ γ(S)kuk for all u ∈ Xe

The Notion of Gain

For a linear system G with signals in L2e ,
γ(G) = max |G(iω)|
ω
Static nonlinear system y(t) = f (u(t)), the gain of the system is

|f (u)|
γ = max
u |u|

BIBO Stability
Definition
A system is bounded-input, bounded-output (BIBO) stable if it has bounded
gain
A BIBO stable system is a system for which the outputs will remain
bounded for all time, for any finite initial condition and input.
A continuous-time linear time-invariant system is BIBO stable if and
only if all the poles of the system have real parts less than 0.

The Small Gain Theorem
Theorem (The Small Gain Theorem)

Consider the system
Let γ1 and γ2 be the gains of the systems H1 and H2 . The closed-loop system is
then BIBO stable if
γ1
γ1 γ2 < 1 and its gain is less than γ=
1 − γ1 γ2

Positivity and Passivity Background
Positivity and Passivity
Positivity and passivity are nearly equivalent concepts.

Positivity is a property of linear systems, while passivity is a more
general concept, applicable to both linear and nonlinear systems.
The scalar, controllable system
ẋ = Ax + bu
y = cT x
with a transfer function
H(s) = cT (sI − A)−1 b
is said to be positive real (PR) if Re[H(s)] ≥ 0 for all Re(s) ≥ 0.

If s = jω, it means that the Nyquist diagram of H(jω) must lies in the
right half plane, including the imaginary axis.
Thus, the phase shift must be between −90 deg and +90 deg.
The Kalman-Yakubovich lemma states that the system H(s) is strictly
positive real if there exist positive definite matrices P and Q such that:
AT P + PA = −Q
Pb = c

Passivity is a general form of positivity

A passive linear system must be positive real and vice-versa.
Consider a system with input u and output y. The energy of the system is
such that
d
[Stored energy]=[external power input]+[internal power generation]
dt
If the internal power generation is negative, then the system is
dissipative or strictly passive. If the internal power generation is less
than or equal to zero the system is passive.

If the input u and output y are such that the external power input is
expressed as uT y, then
The system is passive if hy|ui ≥ 0
The system is input strictly passive (ISP) if there exists ε > 0 such that
hy|ui ≥ kuk2
The system is output strictly passive (OSP) if there exists ε > 0 such that
hy|ui ≥ kyk2
Examples include electrical networks (power proportional to product of
current and voltage), mass-spring systems (mechanical power is force ×
velocity).

The Passivity Theorem
Theorem (Passivity Theorem)

Consider the system
Let H1 be strictly output passive and H2 be passive. The closed-loop system is

then BIBO stable

Positive Real Transfer Functions
A rational transfer function G(s) with real coefficients is PR if and only if the
following conditions hold
1 The function has no poles in the right half-plane
2 If the function has poles on the imaginary axis or at infinity, they are
simple poles with positive residues
3 The real part of G is nonnegative along the iω axis
Re(G(iω)) ≥ 0
G is SPR if conditions 1 and 3 hold and if G(s) has no poles or zeros on the
imaginary axis

Positive real (PR): Re G(jω) ≥ 0

Input strictly passive (ISP): Re G(jω) ≥ ε > 0
Input strictly passive (ISP): Re G(jω) ≥ ε|G(jω)|2
Examples
G(s) = s + 1 is SPR and ISP but not OSP
G(s) = 1/(s + 1) is SPR and OSP but not ISP
G(s) = (s2 + 1)/(s + 1)2 is OSP and ISP but not SPR
G(s) = 1/s is PR, but not SPR, OSP or ISP

Positive Real Transfer Functions
Theorem (Lemma)
Let r be a bounded square integrable function, and let G(s) be a positive real
transfer function. The system whose input-output relation is given by
y = r(G(p)ru)
is then passive.

Positivity and Passivity Application to Adaptive Control

MIT rule:
Lyapunov rule:

Analysis
Redraw (b) as:
Because γ/s is PR, H is passive

Thus, if G is SPR, the passivity theorem implies that the system is L2 (BIBO) stable
If non-zero initial conditions, e(t) is still in L2 and thus goes to zero as t → ∞
This is stable for all γ > 0, so the adaptation can be made arbitrarily fast as long as G is
SPR
If G is not SPR, then may be possible to introduce a compensator Gc such that GGc is
SPR
Analysis
The Kalman-Yakubovich lemma can be used to find Gc such that GGc is

PR
However, when relative degree of G is greater than 1, Gc will contain
derivatives
Augmented error

Augmented Error
Factor G as
G = G1 G2
where G1 is SPR
e = G(θ − θ0 )uc = (G1 G2 )(θ − θ0 )uc

= G1 (G2 (θ − θ0 )uc (θ − θ0 )G2 uc − (θ − θ0 )G2 uc )
Introduce the augmented error
ε = e+η
where η is the error augmentation defined by
η = G1 (θ − θ0 )G2uc − G(θ − θ0 )uc = G1 (θ G2 uc) − Gθ uc
The correction term η vanishes when θ is constant

Augmented Error
The adaptation law

θ̇ = −γεG2 uc
gives a closed-loop system where e(t) goes to zero as t goes to infinity

Augmented Error
Error augmentation is a fundamental idea in MRAC

Details can sometimes be messy...
See example in Section 5.8 for an output feedback design
Error signal
b0
e= (Ru + Sy − Tuc )
A0 Am
Key is introduction of a filter Q such that b0 Q/A0 Am is SPR and
ef = Qe/P
Then the closed-loop system is BIBO stable

Positivity and Passivity Example
Active Suspension

Active Suspension

Active Suspension

EECE 574 - Adaptive Control

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EECE 574 - Adaptive Control

Uploaded by

Copyright:

Available Formats

EECE 574 - Adaptive Control

Model-Reference Adaptive Control - An Overview

Department of Electrical and Computer Engineering

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 1 / 73

The MRAC or MRAS is an important adaptive control methodology 1

1 see Chapter 5 of the Åström and Wittenmark textbook, or H. Butler,

”Model-Reference Adaptive Control-From Theory to Practice”, Prentice-Hall, 1992

The MIT rule

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 3 / 73

The MIT Rule

Original approach to MRAC developed around 1960 at MIT for

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 4 / 73

The MIT Rule

The derivative of J is then described by

Alternatively, one may consider J(e) = |e| in which case

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 5 / 73

MIT Rule: Example 1

Process: y = kG(s) where G(s) is known but k is unknown

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 6 / 73

MIT Rule: Example 1

Figure: MIT rule for adjustment of feedforward gain (from textbook).

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 7 / 73

MIT Rule: Example 1

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 8 / 73

MIT Rule: Example 2

Consider the first-order system

deg A0 ≥ 0 deg S = deg R = deg T = 0

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 9 / 73

MIT Rule: Example 2

The controller is then

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 10 / 73

MIT Rule: Example 2

With the controller we can write2

2p is the differential operator d(.)/dt

MIT Rule: Example 2

MIT Rule: Example 2

Figure: MIT rule for first-order (from textbook)

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 13 / 73

MIT Rule: Example 2

Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2 and

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 14 / 73

MIT Rule: Example 2

Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2.

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 15 / 73

MIT Rule: Example 2

Figure: Simulation of MIT rule for first-order with a = 1, b = 0.5, am = bm = 2.

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 16 / 73

Stability of the Adaptive Loop

Consider again the adaptation of a feedforward gain

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 17 / 73

Stability of the Adaptive Loop

Thus the system can be described as

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 18 / 73

Stability of the Adaptive Loop

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 19 / 73

Stability of the Adaptive Loop

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 20 / 73

Modified MIT Rule

Normalization can be used to protect against dependence on the signal

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 21 / 73

Modified MIT Rule

Guy Dumont (UBC EECE) EECE 574: MRAC - 1 January 2013 22 / 73

Lyapunov Stability Theory