Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 105 views

    Lecture 4

    Uploaded by

    raahan
    This document summarizes model reference adaptive control (MRAC). The controller has two loops: 1) an inner feedback loop consisting of the process and controller, and 2) an outer loop that adjusts the controller parameters. The goal is to minimize the error between the actual output and the model output. Lyapunov stability analysis is used to develop adaptive laws that ensure the error goes to zero over time while keeping the parameters bounded. The parameters may not necessarily converge to their ideal values, but the output will track the model output.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    Lecture 4

    Uploaded by

    raahan
    105 views9 pages
    This document summarizes model reference adaptive control (MRAC). The controller has two loops: 1) an inner feedback loop consisting of the process and controller, and 2) an outer loop that adjusts the controller parameters. The goal is to minimize the error between the actual output and the model output. Lyapunov stability analysis is used to develop adaptive laws that ensure the error goes to zero over time while keeping the parameters bounded. The parameters may not necessarily converge to their ideal values, but the output will track the model output.

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document summarizes model reference adaptive control (MRAC). The controller has two loops: 1) an inner feedback loop consisting of the process and controller, and 2) an outer loop that adjusts the controller parameters. The goal is to minimize the error between the actual output and the model output. Lyapunov stability analysis is used to develop adaptive laws that ensure the error goes to zero over time while keeping the parameters bounded. The parameters may not necessarily converge to their ideal values, but the output will track the model output.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    105 views9 pages

    Lecture 4

    Uploaded by

    raahan
    This document summarizes model reference adaptive control (MRAC). The controller has two loops: 1) an inner feedback loop consisting of the process and controller, and 2) an outer loop that adjusts the controller parameters. The goal is to minimize the error between the actual output and the model output. Lyapunov stability analysis is used to develop adaptive laws that ensure the error goes to zero over time while keeping the parameters bounded. The parameters may not necessarily converge to their ideal values, but the output will track the model output.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 9

    Lecture 4

    CY4A2: Advanced Nonlinear Control

    Model Reference Adaptive Control

    y:z!{|zj}|z p
    ~ z
    h S-i?STO8SRL3WjS nKlq z ulN€
    |‚€
    ƒv„…
    _1J Z SRQ
    Y[Z\^] X6MP_1SRL`M
    _aS/WcbedTL-fgX-_
    k:ljp
     } leuu zN}
    ‚o } ojn z  zD} r v
    c† } l~ z r!r
    lj€"‚D€"?v&‡
    kmlDn[nKoDpq IKJ:LNMPO8JRQQSTO UVO?J/WNS-XNX
    rs t po`ugv"wx :k lDp" } l`u
    rs t po`ugvw

    This is an adaptive control technique where


    the performance specifications are given in
    terms of a model. The model represents the
    ideal response of the process to a command
    signal.
    54
    The controller has two loops:

    1. The inner loop, which is an ordinary feed-


    back loop consisting of the process and
    the controller.

    2. The outer loop, which adjust the con-


    troller parameters in such a way that the
    error e = y - y m is small (not trivial)

    Based on ideas by Whitaker (1958), from


    MIT.

    Approaches: Gradient approach (MIT Rule)


    and Lyapunov direct approach.

    55
    Lyapunov approach to stable adaptation

    In MRAC we want the error e = y − ym to go


    to zero, we will now attempt to find a Lya-
    punov function and adaptation mechanism to
    achieve it.
    Example 13. A first order system

    Process:
    dy
    = −ay + bu
    dt
    Model:
    dym
    = −amym + bmuc
    dt
    where am > 0 and the reference uc is bounded.

    Controller:

    u = θ 1 uc − θ 2 y

    56
    Error:

    e = y − ym

    de = −ay + b(θ u − θ y) + a y − b u
    dt 1 c 2 m m m c

    de = a y − (bθ + a)y + (bθ − b )u


    dt m m 2 1 m c

    de = −a e − (bθ + a − a )y + (bθ − b )u
    dt m 2 m 1 m c

    Note: perfect model following occurs when


    the parameters have ideal values: θ 1 = bm/b
    and θ 2 = (am − a)/b

    Define φ2 = bθ2 + a − am, φ1 = bθ1 − bm so


    that when the parameters have ideal values,
    φ1 = φ2 = 0

    With that definition, we have


    de
    = −ame − φ2y + φ1uc
    dt
    57
    Candidate for Lyapunov function:
    1 2 1 1 2
    V (e, φ1, φ2) = (e + φ21 + φ2 )
    2 bγ bγ
    Note: V→0 when e→0 and the parameters
    θ 1,θ 2 are ideal.

    Derivative of the candidate Lyapunov func-


    tion:
     
    de
    h i dt 
    V̇ = ∂t + ∂V
    ∂V ∂V ∂V  dφ1 
    ∂e ∂φ1 ∂φ2  dt 
    dφ2
    dt

    = e de + 1 φ dφ1 + 1 φ dφ2
    dt bγ 1 dt bγ 2 dt

    1 φ dφ1 + 1 φ dφ2
    = −ame2 − eφ2y + eφ1uc + bγ 1 dt bγ 2 dt
    h i h i
    2 1 dφ1 1 dφ2
    = −ame + φ1 bγ dt + euc + φ2 bγ dt − ey

    58
    If we choose the following adaptation laws
    dφ1
    dt = −bγuce

    dφ2
    dt = γbye
    Then
    V̇ = −ame2
    Notice that am > 0. Note also that the sys-
    tem described by the state equations:
    de
    = −ame − φ1y + φ2uc
    dt
    dφ1
    dt = −bγuce

    dφ2
    dt = γbye
    has an equilibrium point at x = (e, φ1, φ2)T =
    (0, 0, 0)T

    59
    Since V (e, φ1, φ2) is positive definite and V̇ =
    −ame2 6 0

    then, according to Theorem 2 the equilibrium


    point x = 0 is uniformly stable, but will e →0
    and will the parameters θ1 and θ2 converge
    to their ideal values when t → ∞ ?

    The answer to this question is as follows:

    Since dV 2
    dt = −am e is negative semi-definite
    then V (t) 6 V (0).

    Then, e, θ1 and θ2 must be bounded.

    60
    Since am > 0, and uc is bounded then ym is
    bounded, and therefore y = e+ym is bounded
    as well.

    From the fact that all these variables are


    bounded, it is possible to conclude by using
    some extra mathematical arguments (see As-
    trom and Wittenmark, 1995) that V̇ → 0 as
    t → ∞, which implies that the error e(t) →0
    as t → ∞.

    The parameters θ1 and θ2 will not necessarily


    converge to the ideal values.

    61
    Notice that the adaptation laws found can
    be written in terms of the original controller
    parameters as follows:
     
    dθ1 −γ
    dt = −γuc e ⇒ θ1 = p uc e
     
    dθ2 γ
    dt = γye ⇒ θ2 = p ye

    ym

    2 To Workspace2
    s+2
    Model sub

    0.5
    Command y
    Product s+1
    Signal
    sub1 To Workspace3
    theta1 Process

    To Workspace
    -g
    Integrator
    s theta2
    Product1
    g To Workspace1
    Integrator_
    s

    Product2

    Product3 Plot
    results

    62

    You might also like