Scribd
Upload
177 views11 pages

Group # 7: Kalman Filter Based Parameter Estimation

This document contains a lab report submitted by a group of students to their professor on parameter estimation using an extended Kalman filter. The report summarizes the Kalman filter and extended Kalman filter concepts. It then describes applying an EKF to estimate the parameter ω in a second order system model. The students implemented the EKF algorithm in MATLAB code and a Simulink model. They provided plots showing the estimated parameter tracks the true parameter over time.

Uploaded by

sameerfarooq420840
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
177 views11 pages

Group # 7: Kalman Filter Based Parameter Estimation

This document contains a lab report submitted by a group of students to their professor on parameter estimation using an extended Kalman filter. The report summarizes the Kalman filter and extended Kalman filter concepts. It then describes applying an EKF to estimate the parameter ω in a second order system model. The students implemented the EKF algorithm in MATLAB code and a Simulink model. They provided plots showing the estimated parameter tracks the true parameter over time.

Uploaded by

sameerfarooq420840
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 11

PIEAS

Kalman Filter based Parameter Estimation


Lab Report

Group # 7
Submitted to: Dr AQ Khan

Submitted by: Kamran Iftikhar-51997 Moazzam Nazir-50481 Muhammad Ahmed Muhammad Aqeel Abassi Muhammad Awais Tayyab-52251

2/27/2012

Lab 7
Kalman filter based parameter estimation/optimization

Kalman Filter:
The Kalman filter is a set of mathematical equations that provides an efficient computational (recursive) means to estimate the state of a process, in a way that minimizes the mean of the squared error. The filter is very powerful in several aspects: it supports estimations of past, present, and even future states, and it can do so even when the precise nature of the modeled system is unknown. Kalman filter is used in industrial navigation, radar tracking, signal processing, sensor calibration etc

Extended Kalman Filter:


If the process to be estimated or the measurement relationship to the process is non-linear then extended Kalman filter is used. A Kalman filter that linearizes about the current mean and covariance is referred to as an extended Kalman filter or EKF.

Objective:
The objective of this lab was to write m-file , s-function and built a simulink model for parameter estimation using extended Kalman filter.

Algorithm:
Let some process

Where is a process noise

In case of non-linear system we also have to linearizes system. So given

The Kalman filter estimates a process by using a form of feedback control: the filter estimates the process state at some time and then obtains feedback in the form of (noisy) measurements. As such, the equations for the Kalman filter fall into two groups: time update equations and measurement update equations. The time update equations can also be thought of as predictor equations, while the measurement update equations can be thought of as corrector equations. Predictor-corrector algorithm can be viewed as shown in figure

Time Update Equations (predict):


Computing the predicted state estimate: Computing the predicted measurement: Computing the a priori covariance matrix: ( )

Measurement Update Equations (correct):


Computing the Kalman gain: Conditioning the predicted estimate on the measurement: given

Computing the a posteriori covariance matrix {

Task:

In this task we have to estimate the value of

. To do this we take

as the third state.

This means First we discritize the plant by using T=0.01 second. We also take w=10 and zeta=0.7, so the discritze plant is

To estimate the value of , time update equations and measurement update equations are used recursively. For Matlab implementation code is written for these equations and for sfunctions same code is used in the s-function template file to generate a s-function for the given problem.

Matlab code:
T=.01; w=10; zeta=.7; A=[1 T 0;-0.1 1 0;0 0 1]; B=[0;12*T;0]; E=[0;T;0]; C=[1 0 0]; D=0; F=1;

xo=[0;0;0]; u=1; Po=[1 0 0;0 1 0;.1 0 1]; Q=[0 0 0;0 4.47 0;0 0 0]; R=.01; Pn=Po; Ew=[0;T;0]; Kk=[0;0;0]; %//////////////////////// x110=0;x220=0; x11=x110;x22=x220; x10=0;x20=0;x30=0; x1=x10;x2=x20;x3=x30; yh=0; xo0=[0;0;0]; Pn=Po; A=[1 T 0;-T*w*w 1-(2*T*w*x3) -2*T*w*x2;0 0 1]; y=[]; yhat=[]; origx1=[];origx2=[]; estx1=[];estx2=[];estx3=[]; for i=1:500 x11d=x11+T*x22; x22d=x22-T*w*w*x11-2*T*w*x22*zeta+12*T*u+T*(randn*(4.47)); y=[y;x11d+(randn*(.01))]; x11=x11d;x22=x22d; %//////// origx1=[origx1;x11d]; origx2=[origx2;x22d]; x1d=x1+T*x2; x2d=x2-T*w*w*x1-2*T*w*x2*x3-2*T*w*x3*x2+12*T*u; x3d=x3; A=[1 T 0;-T*w*w 1-(2*T*w*x3) -2*T*w*x2;0 0 1]; x1d=x1d+Kk(1)*(y(i)-x1d); x2d=x2d+Kk(2)*(y(i)-x1d); x3d=x3d+Kk(3)*(y(i)-x1d); Pn=A*Pn*A'+Ew*4.47*Ew'; Kk=Pn*C'*((C*Pn*C'+R)^-1); Pp=((eye(3)-Kk*C)*Pn);

Pn=Pp; yhat=[yhat;x1d]; x1=x1d;x2=x2d;x3=x3d; estx1=[estx1;x1]; estx2=[estx2;x2]; estx3=[estx3;x3]; end Kk x3 subplot(311) plot(origx1,'r') hold on plot(estx1) subplot(312) plot(origx2,'r') hold on plot(estx2) subplot(313) plot(estx3)

Results:

Implementation in Simulink:
Overall:

Plant:

Kalman Filter:

Results:
Plant State:

Kalman Filter State Estimate:

S-Function:
function msfuntmpl_basic(block) setup(block); function setup(block) % Register number of ports block.NumInputPorts = 4; block.NumOutputPorts = 1; % Setup port properties to be inherited or dynamic block.SetPreCompInpPortInfoToDynamic; block.SetPreCompOutPortInfoToDynamic; % Override input port properties %u(k-1) block.InputPort(1).Dimensions block.InputPort(1).DatatypeID = 0;

= 1; % double

block.InputPort(1).Complexity = 'Real'; block.InputPort(1).DirectFeedthrough = true; %z block.InputPort(2).Dimensions = 1; block.InputPort(2).DatatypeID = 0; % double block.InputPort(2).Complexity = 'Real'; block.InputPort(2).DirectFeedthrough = true; %Q(k-1) block.InputPort(3).Dimensions = [2,2]; block.InputPort(3).DatatypeID = 0; % double block.InputPort(3).Complexity = 'Real'; block.InputPort(3).DirectFeedthrough = true; %R(k) block.InputPort(4).Dimensions = 1; block.InputPort(4).DatatypeID = 0; % double block.InputPort(4).Complexity = 'Real'; block.InputPort(4).DirectFeedthrough = true; % Override output port properties block.OutputPort(1).Dimensions = 1; block.OutputPort(1).DatatypeID = 0; % double block.OutputPort(1).Complexity = 'Real'; % Register parameters block.NumDialogPrms % Register sample times block.SampleTimes = [0 0]; block.RegBlockMethod('PostPropagationSetup', @DoPostPropSetup); block.RegBlockMethod('InitializeConditions', @InitializeConditions); block.RegBlockMethod('Start', @Start); block.RegBlockMethod('Outputs', @Outputs); % Required block.RegBlockMethod('Update', @Update); block.RegBlockMethod('Derivatives', @Derivatives); block.RegBlockMethod('Terminate', @Terminate); % Required function DoPostPropSetup(block) block.NumDworks = 1; block.Dwork(1).Name block.Dwork(1).Dimensions block.Dwork(1).DatatypeID block.Dwork(1).Complexity block.Dwork(1).UsedAsDiscState = 0;

= = = = =

'x1'; 1; 0; % double 'Real'; % real true;

function InitializeConditions(block) function Start(block) block.Dwork(1).Data = 0;

10

function Outputs(block) block.OutputPort(1).Data = block.Dwork(1).Data + block.InputPort(1).Data; function Update(block) block.Dwork(1).Data = block.InputPort(1).Data; function Derivatives(block) function Terminate(block)

11

You might also like