Linear-Optimal State Estimation
|
|
- Anabel Wright
- 5 years ago
- Views:
Transcription
1 Linear-Optimal State Estimation Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 218 Linear-optimal Gaussian estimator for discrete-time system (Kalman filter) 2 nd -order example Alternative forms of the Kalman filter equations Ill conditioning Correlated inputs and measurement noise ime-correlated measurement noise Copyright 218 by Robert Stengel. All rights reserved. For educational use only Uncertain Linear, ime-varying (LV) Dynamic Model x = Φ 1 x 1 Γ 1 u 1 Λ 1 w 1 z = x n = s 1 = r 1 dim w dim z Initial condition and disturbance inputs are not nown precisely Measurement of state is transformed and is subject to error 2
2 Actual System State Estimation Goals Minimize effects of measurement error on nowledge of the state Reconstruct full state from reduced measurement set (r n) Average redundant measurements (r n) to estimate the full state Method Provide optimal balance between measurements and estimates based on the dynamic model alone 3 Linear-Optimal State Estimation Kalman filter is the optimal estimator for discrete-time linear systems with Gaussian uncertainty It has five equations 1) State estimate extrapolation 2) Covariance estimate extrapolation 3) Filter gain computation 4) State estimate update 5) Covariance estimate update Notation ˆx : Estimate at th instant before measurement update : Estimate at th instant after measurement update ˆx 4
3 Equations of the Kalman Filter 1) State estimate extrapolation (or propagation) ˆx = Φ 1 ˆx 1 ( ) Γ 1 u 1 2) Covariance estimate extrapolation (or propagation) P = Φ 1 P 1 ( )Φ 1 Q 1 5 Equations of the Kalman Filter K = P 3) Filter gain computation P 4) State estimate update ˆx ( ) = ˆx K z ˆx H R 5) Covariance estimate update 1 = P 1 H R 1 P 1 6
4 Diagram of State Estimate Fig , OCE [1] [4] 7 Covariance Estimate and Filter Gain Matrix Fig , OCE [3] [2] [5] 8
5 Alternative Expressions for K I r = R 1 R Matrix Identity K = P P H R 1 = P H R 1 R P H 1 R K I r P H 1 R = P H 1 R K = P H 1 R P H R 1 1 I r Multiply by inverse K = P H R 1 K P H 1 R = ( I n K )P H 1 R Subtract from both sides Collect terms 9 = P 1 H R 1 P Alternative Expressions for K and P () From matrix inversion lemma = P P = ( I n K )P 1 P H 1 R P his covariance update does not inherently preserve symmetry With P () nown, estimation gain is K = ( I n K )P H R 1 = P ( )H 1 R 1
6 Second-Order Example of Kalman Filter Rolling motion of an airplane, continuous-time Δ p Δ φ = L p Δp 1 Δφ L δ A Δδ A L p Δp w Rolling motion of an airplane, discrete-time Δp Δφ = Roll rate, rad/s Roll angle, rad Δδ A = Aileron deflection, rad Δp w = urbulence disturbance, rad/s Δp Δφ = ( e Lp 1) 1 L p e L p ϕ 11 = ϕ 21 1 Δp 1 Δφ 1 Δp 1 Δφ 1 γ 1 ~ L δa ΔδA 1 ΔδA 1 λ 1 Δp w 1 ~ L p Δp w 1 = sampling interval, s 11 Second-Order Example of Kalman Filter Rate and Angle Measurement Δp M Δφ M = Δp Δn p Δφ Δn φ = IΔx Δn 1) State Estimate Extrapolation Δˆp Δ ˆφ = ϕ 11 ϕ 21 1 Δˆp 1 Δ ˆφ 1 γ 1 Δδ A 1 12
7 Second-Order Example of Kalman Filter 2) Covariance Extrapolation p 11 p 21 p 12 p 22 = ϕ 11 ϕ 21 1 p 12 ( ) p 22 ( ) p 11 p 21 1 ϕ 11 ϕ 21 1 σ p 2 Q 1 L p Q' L C p = 2 σ p 3) Gain Computation = p 11 p 21 p 12 p 22 p 11 p 21 p 12 p 22 2 σ pm 2 σ φm 1 R δ j = 2 σ pm 2 σ φm 13 Second-Order Example of Kalman Filter 4) State Estimate Update Δˆp Δ ˆφ = Δˆp Δ ˆφ Δp M Δφ M Δˆp Δ ˆφ 5) Covariance Update p 12 ( ) p 22 ( ) p 11 p 21 = p 11 p 21 p 12 p 22 1 σ pm σ φm 1 14
8 Example: Propagating a Scalar Probability Density Function Scalar LI system with zero-mean Gaussian random input x 1 = bx 1 b 2 u 1 b 2 w, x given Propagation of the mean value x i1 = bx i 1 b 2 u i, x given Propagation of the variance P 1 = b 2 P ( 1 b 2 )Q, P given 2 Q = E( w ) 15 Example: System Response to Disturbance, Large Measurement Error Large initial uncertainty 16
9 Measurement Error and State Estimate Standard Deviations 17 Actual State, Estimated State and Standard Deviation! High measurement error! State estimate emphasizes system model 18
10 Example: System Response to Disturbance, Small Measurement Error Large initial uncertainty 19 Measurement Error and State Estimate Standard Deviations 2
11 Actual State, Estimated State and Standard Deviation! Small measurement error! State estimate emphasizes measurements 21 Linear-Optimal Predictor t : Current time, sec t K : Future time, sec State estimate propagation from last estimate ˆx K = Φ( t K t ) ˆx ( ) Γ ( t K t )u Covariance estimate propagation P P K = Φ t K t Q ( t K t ) Φ t K t Predictor analogous to Kalman filter without measurement 22
12 Prediction Depends Entirely on the Model Bad model bad prediction High disturbance covariance bad prediction Prediction of mean might seem good but if covariance grows, uncertainty may render prediction meaningless 23 Discrete-ime Linear-Optimal Prediction, 1 points 24
13 Discrete-ime Linear-Optimal Prediction, u =.2 sin /2π, 1 points 25 Alternative Forms of the Kalman Filter 26
14 Simplifying the Gain Calculation Filter gain computation for r measurements K = P P r x r inversion required H R 1 K = For r = 1 P P H r scalar division 27 Consider One Measurement at Each Sampling Interval With r measurements and diagonal R, consider just one measurement at a time z = z 1 z 2 z r r 11 r ; R = 22 ; H = r rr ˆx = Φ 1 ˆx 1 Γ 1 u 1 ˆx ( ) = ˆx K i z i H i ˆx Scalar Update K i = P H i H i P H i r i Cycle through all measurements varying H, and repeat cycle Wasteful, as it may not use all available information H 1 H 2 H r z 1 = z 1 z 2 = z 2 z r = z r 28
15 Sequential Processing of Measurements at Each Sampling Interval With r measurements, form an inner loop of calculations, processing one sample at a time for i = 1,r K i = P i1 ( )H i H i P i1 ( )H i r i P i ( ) = ( I n K i H i )P i1 ( ), P end ˆx i ( ) = ˆx i1 ( ) K i z i H i ˆx i1 ( ) = ˆx r ( ) = P r ( ) ˆx P = P z 1 = z 1 z 2 = z 2 z r = z r H 1 = H 1 H 2 = H 2 H r = H r 29 ε Joseph Form of the Filter ( Stabilized Kalman Filter ) Guaranteed to retain positivedefiniteness and symmetry State update is ˆx ( ) = [ I n K ] ˆx K z Pre- and post-update measurement errors = x ˆx ; ε [ ]ε ε ( ) = I n K = x ˆx ( ) Measurement error is updated by K n 3
16 Joseph Form of the Filter ( Stabilized Kalman Filter ) ε E ε ε [ ]ε = I n K Definitions K n = P ; E( n n ) = R ; E( ε n ) = hen, covariance update is the outer product of the expected error P ( ) = [ I n K ]P [ I n K ] K R K Update equation is symmetric Equation updates covariance whether or not K is optimal Evaluate error covariance of a sub-optimal filter Design and evaluate a low-order filter for a high-order system Does not require an (n x n) inversion 31 P : Information Matrix Form of the Kalman Filter Filter based on the inverse of P State error covariance (small is good) I = P 1 : Information matrix (large is good) Fewer continuing inversions = P 1 H R 1 P I ( ) = I H R
17 Information Matrix Propagation and Gain Matrix I 1 = I n B 1 Λ 1 A 1, I = P where A 1 = Φ 1 1 I 1 ( )Φ 1 B 1 = A 1 Λ A 1 Λ 1 Q' 1 and K = I 1 ( )H 1 R 33 Ill Conditioning of the Filter Computation Calculations may be inaccurate if system contains very fast and slow modes very noisy and near-perfect measurements very large and small disturbance inputs Condition number of a matrix, P, is the ratio of singular values = λ max P P P λ min ( P P) 1/2 = σ max P P σ min = σ ( P) σ ( P) 34
18 Solutions to Ill Conditioning Double, triple,..., precision arithmetic, or Formulate the equations to solve for the square root of P Define P! SS or S S then ( P) = ( SS ) = 2 ( S), which is order 1 x ( S) = ( P), which is order 1 x/2 35 U-D Formulation of the Kalman Filter U : Factorization of P P = UDU ( UD 1/2 )( UD 1/2 ) SS where Unit upper triangular matrix: D : Diagonal matrix: " " " 1 " " 1 " 1 No square roots in the factorization, but formulation has squareroot conditioning Covariance update uses sequential processing Algorithm originally expressed in pseudo-code (Bierman and hornton, 1977) See OCE (pp ) for equations and pseudo-code 36
19 Kalman Filters for More Complex Systems* Correlated Disturbance Input and Measurement Error Measurements may be corrupted by the same processes that force the system (e.g., turbulence) Consider estimation-error dynamics, as in Joseph form Stepwise minimization of estimation cost function w.r.t. filter gain matrix See OCE ime-correlated ( Colored ) Measurement Error Augment system with measurement error dynamics Use measurement differencing wo-step state and covariance estimates See OCE * Arthur Bryson and students 37 Next ime: Kalman-Bucy Filters for Continuous-ime Systems 38
20 Supplemental Material 39 E x Descriptions of Random Variables = ˆx o ; E ( x ˆx o )( x ˆx o ) E u E( w ) = ; E( w j w ) = Q' δ j [ ] = u ; E [ u u ][ u u ] = P { } = E( n ) = ; E n j n = E w j n = R δ j 4
21 Program for Covariance Propagation and Kalman Filter Estimatation % Scalar Propagation of Mean and Variance % Kalman Filter % Copyright by Robert Stengel. All rights reserved. % 4/12/211 clear '==========================' 'Propagation of Uncertainty' date b =.99 b2 = b*b bb = 1 - b^2 sqrb = sqrt(bb) w = []; x = []; xbar = []; x(1) = 1 xbar(1) = x(1) P(1) = 3 Q =.1 u = for = 1:999 w() = sqrt(q)*randn(1); x(1) = b*x() sqrb*u sqrb*w(); xbar(1) = b*xbar() sqrb*u; P(1) = b2*p() bb*q; end = [1:1]; w(1) = sqrt(q)*randn(1); figure plot(,w,'c',,x,'b',,(xbar sqrt(p)),'',,(xbar - sqrt(p)),'',,xbar,'r'),grid title(['propagation of Uncertainty, b =:', num2str(b), ', Q =:',num2str(q)]), xlabel('sampling Index, '), ylabel('x, xbar, xbar ± sqrt(p)') legend('disturbance','state','mean Sigma','Mean - Sigma','Mean') % Kalman Filter '=============' 'Kalman Filter' xm(1) = xp(1) = Pm(1) = 1 Pp(1) = 1 z(1) = xm(1) Q = Q R =.1 u = for = 1:999 n() = sqrt(r)*randn(1); xm(1) = b*xp() sqrb*u; Pm(1) = b2*pp() Q; K = Pm(1) / (Pm(1) R); z(1) = x(1) n(); xp(1) = xm(1) K*(z(1) - xm(1)); Pp(1) = 1 / ((1/Pm(1)) 1/R); end = [1:1]; n(1) = sqrt(r)*randn(1); figure plot(,z,'c',,(xp sqrt(pp)),'g',,(xp - sqrt(pp)),'g',,xp,'r',,x,'b'),grid title(['kalman Filter, Q =:',num2str(q),', R =:',num2str(r)]), xlabel('sampling Index, '), ylabel('x, xp, xp ± sqrt(pp)') legend('measurement','estimate Sigma','Estimate - Sigma','Estimate','Actual') % Covariance Comparison figure SP = sqrt(p); SPp = sqrt(pp); plot(,n,'c',,w,'g',,sp,'b',,spp,'r'), grid title(['covariance Comparison, Q =:',num2str(q),', R =:' num2str(r)]), xlabel('sampling Index, '),ylabel('w, n, Sqrt(P), Sqrt(Pp)') legend('measurement Error','Disturbance','Estimate of Actual State Std Dev','Estimate Error Std Dev') 41
OPTIMAL CONTROL AND ESTIMATION
OPTIMAL CONTROL AND ESTIMATION Robert F. Stengel Department of Mechanical and Aerospace Engineering Princeton University, Princeton, New Jersey DOVER PUBLICATIONS, INC. New York CONTENTS 1. INTRODUCTION
More informationOptimal Control and Estimation MAE 546, Princeton University Robert Stengel, Preliminaries!
Optimal Control and Estimation MAE 546, Princeton University Robert Stengel, 2018 Copyright 2018 by Robert Stengel. All rights reserved. For educational use only. http://www.princeton.edu/~stengel/mae546.html
More informationReturn Difference Function and Closed-Loop Roots Single-Input/Single-Output Control Systems
Spectral Properties of Linear- Quadratic Regulators Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 2018! Stability margins of single-input/singleoutput (SISO) systems! Characterizations
More informationHere represents the impulse (or delta) function. is an diagonal matrix of intensities, and is an diagonal matrix of intensities.
19 KALMAN FILTER 19.1 Introduction In the previous section, we derived the linear quadratic regulator as an optimal solution for the fullstate feedback control problem. The inherent assumption was that
More informationThe Kalman Filter. Data Assimilation & Inverse Problems from Weather Forecasting to Neuroscience. Sarah Dance
The Kalman Filter Data Assimilation & Inverse Problems from Weather Forecasting to Neuroscience Sarah Dance School of Mathematical and Physical Sciences, University of Reading s.l.dance@reading.ac.uk July
More informationNonlinear State Estimation! Particle, Sigma-Points Filters!
Nonlinear State Estimation! Particle, Sigma-Points Filters! Robert Stengel! Optimal Control and Estimation, MAE 546! Princeton University, 2017!! Particle filter!! Sigma-Points Unscented Kalman ) filter!!
More informationStochastic Optimal Control!
Stochastic Control! Robert Stengel! Robotics and Intelligent Systems, MAE 345, Princeton University, 2015 Learning Objectives Overview of the Linear-Quadratic-Gaussian (LQG) Regulator Introduction to Stochastic
More informationMinimization of Static! Cost Functions!
Minimization of Static Cost Functions Robert Stengel Optimal Control and Estimation, MAE 546, Princeton University, 2017 J = Static cost function with constant control parameter vector, u Conditions for
More informationThe Kalman Filter (part 1) Definition. Rudolf Emil Kalman. Why do we need a filter? Definition. HCI/ComS 575X: Computational Perception.
The Kalman Filter (part 1) HCI/ComS 575X: Computational Perception Instructor: Alexander Stoytchev http://www.cs.iastate.edu/~alex/classes/2007_spring_575x/ March 5, 2007 HCI/ComS 575X: Computational Perception
More informationOnline monitoring of MPC disturbance models using closed-loop data
Online monitoring of MPC disturbance models using closed-loop data Brian J. Odelson and James B. Rawlings Department of Chemical Engineering University of Wisconsin-Madison Online Optimization Based Identification
More informationAdaptive State Estimation Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 2018
Adaptive State Estimation Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 218! Nonlinearity of adaptation! Parameter-adaptive filtering! Test for whiteness of the residual!
More informationNonlinear State Estimation! Extended Kalman Filters!
Nonlinear State Estimation! Extended Kalman Filters! Robert Stengel! Optimal Control and Estimation, MAE 546! Princeton University, 2017!! Deformation of the probability distribution!! Neighboring-optimal
More informationLinear-Quadratic-Gaussian Controllers!
Linear-Quadratic-Gaussian Controllers! Robert Stengel! Optimal Control and Estimation MAE 546! Princeton University, 2017!! LTI dynamic system!! Certainty Equivalence and the Separation Theorem!! Asymptotic
More informationOptimal Control and Estimation MAE 546, Princeton University Robert Stengel, Preliminaries!
Optimal Control and Estimation MAE 546, Princeton University Robert Stengel, 2017 Copyright 2017 by Robert Stengel. All rights reserved. For educational use only. http://www.princeton.edu/~stengel/mae546.html
More informationKalman Filter Computer Vision (Kris Kitani) Carnegie Mellon University
Kalman Filter 16-385 Computer Vision (Kris Kitani) Carnegie Mellon University Examples up to now have been discrete (binary) random variables Kalman filtering can be seen as a special case of a temporal
More informationState Observers and the Kalman filter
Modelling and Control of Dynamic Systems State Observers and the Kalman filter Prof. Oreste S. Bursi University of Trento Page 1 Feedback System State variable feedback system: Control feedback law:u =
More informationECONOMETRIC METHODS II: TIME SERIES LECTURE NOTES ON THE KALMAN FILTER. The Kalman Filter. We will be concerned with state space systems of the form
ECONOMETRIC METHODS II: TIME SERIES LECTURE NOTES ON THE KALMAN FILTER KRISTOFFER P. NIMARK The Kalman Filter We will be concerned with state space systems of the form X t = A t X t 1 + C t u t 0.1 Z t
More informationLinear Prediction Theory
Linear Prediction Theory Joseph A. O Sullivan ESE 524 Spring 29 March 3, 29 Overview The problem of estimating a value of a random process given other values of the random process is pervasive. Many problems
More informationInformation Formulation of the UDU Kalman Filter
Information Formulation of the UDU Kalman Filter Christopher D Souza and Renato Zanetti 1 Abstract A new information formulation of the Kalman filter is presented where the information matrix is parameterized
More informationTime-Invariant Linear Quadratic Regulators Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 2015
Time-Invariant Linear Quadratic Regulators Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 15 Asymptotic approach from time-varying to constant gains Elimination of cross weighting
More informationControl Systems. State Estimation.
State Estimation chibum@seoultech.ac.kr Outline Dominant pole design Symmetric root locus State estimation We are able to place the CLPs arbitrarily by feeding back all the states: u = Kx. But these may
More informationSince the determinant of a diagonal matrix is the product of its diagonal elements it is trivial to see that det(a) = α 2. = max. A 1 x.
APPM 4720/5720 Problem Set 2 Solutions This assignment is due at the start of class on Wednesday, February 9th. Minimal credit will be given for incomplete solutions or solutions that do not provide details
More informationHomework: Ball Predictor for Robot Soccer
Homework: Ball Predictor for Robot Soccer ECEn 483 November 17, 2004 1 Introduction The objective of this homework is to prepare you to implement a ball predictor in the lab An extended Kalman filter will
More informationENGR352 Problem Set 02
engr352/engr352p02 September 13, 2018) ENGR352 Problem Set 02 Transfer function of an estimator 1. Using Eq. (1.1.4-27) from the text, find the correct value of r ss (the result given in the text is incorrect).
More informationEvery real system has uncertainties, which include system parametric uncertainties, unmodeled dynamics
Sensitivity Analysis of Disturbance Accommodating Control with Kalman Filter Estimation Jemin George and John L. Crassidis University at Buffalo, State University of New York, Amherst, NY, 14-44 The design
More informationFor final project discussion every afternoon Mark and I will be available
Worshop report 1. Daniels report is on website 2. Don t expect to write it based on listening to one project (we had 6 only 2 was sufficient quality) 3. I suggest writing it on one presentation. 4. Include
More informationCramér-Rao Bounds for Estimation of Linear System Noise Covariances
Journal of Mechanical Engineering and Automation (): 6- DOI: 593/jjmea Cramér-Rao Bounds for Estimation of Linear System oise Covariances Peter Matiso * Vladimír Havlena Czech echnical University in Prague
More informationStochastic and Adaptive Optimal Control
Stochastic and Adaptive Optimal Control Robert Stengel Optimal Control and Estimation, MAE 546 Princeton University, 2018! Nonlinear systems with random inputs and perfect measurements! Stochastic neighboring-optimal
More information2 Introduction of Discrete-Time Systems
2 Introduction of Discrete-Time Systems This chapter concerns an important subclass of discrete-time systems, which are the linear and time-invariant systems excited by Gaussian distributed stochastic
More informationA Review of Matrix Analysis
Matrix Notation Part Matrix Operations Matrices are simply rectangular arrays of quantities Each quantity in the array is called an element of the matrix and an element can be either a numerical value
More informationTime-Invariant Linear Quadratic Regulators!
Time-Invariant Linear Quadratic Regulators Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 17 Asymptotic approach from time-varying to constant gains Elimination of cross weighting
More informationCONTROL SYSTEMS, ROBOTICS AND AUTOMATION Vol. VIII Kalman Filters - Mohinder Singh Grewal
KALMAN FILERS Mohinder Singh Grewal California State University, Fullerton, USA Keywords: Bierman, hornton, Cholesky decomposition, continuous time optimal estimator, covariance, estimators, filters, GPS,
More informationLeast Squares. Ken Kreutz-Delgado (Nuno Vasconcelos) ECE 175A Winter UCSD
Least Squares Ken Kreutz-Delgado (Nuno Vasconcelos) ECE 75A Winter 0 - UCSD (Unweighted) Least Squares Assume linearity in the unnown, deterministic model parameters Scalar, additive noise model: y f (
More informationInformation, Covariance and Square-Root Filtering in the Presence of Unknown Inputs 1
Katholiee Universiteit Leuven Departement Eletrotechnie ESAT-SISTA/TR 06-156 Information, Covariance and Square-Root Filtering in the Presence of Unnown Inputs 1 Steven Gillijns and Bart De Moor 2 October
More informationIntroduction to Optimization!
Introduction to Optimization! Robert Stengel! Robotics and Intelligent Systems, MAE 345, Princeton University, 2017 Optimization problems and criteria Cost functions Static optimality conditions Examples
More informationExtension of Farrenkopf Steady-State Solutions with Estimated Angular Rate
Extension of Farrenopf Steady-State Solutions with Estimated Angular Rate Andrew D. Dianetti and John L. Crassidis University at Buffalo, State University of New Yor, Amherst, NY 46-44 Steady-state solutions
More information6.4 Kalman Filter Equations
6.4 Kalman Filter Equations 6.4.1 Recap: Auxiliary variables Recall the definition of the auxiliary random variables x p k) and x m k): Init: x m 0) := x0) S1: x p k) := Ak 1)x m k 1) +uk 1) +vk 1) S2:
More informationKalman Filters with Uncompensated Biases
Kalman Filters with Uncompensated Biases Renato Zanetti he Charles Stark Draper Laboratory, Houston, exas, 77058 Robert H. Bishop Marquette University, Milwaukee, WI 53201 I. INRODUCION An underlying assumption
More informationMinimization with Equality Constraints
Path Constraints and Numerical Optimization Robert Stengel Optimal Control and Estimation, MAE 546, Princeton University, 2015 State and Control Equality Constraints Pseudoinverse State and Control Inequality
More informationControl Systems! Copyright 2017 by Robert Stengel. All rights reserved. For educational use only.
Control Systems Robert Stengel Robotics and Intelligent Systems MAE 345, Princeton University, 2017 Analog vs. digital systems Continuous- and Discretetime Dynamic Models Frequency Response Transfer Functions
More informationTime Response of Linear, Time-Invariant (LTI) Systems Robert Stengel, Aircraft Flight Dynamics MAE 331, 2018
Time Response of Linear, Time-Invariant LTI Systems Robert Stengel, Aircraft Flight Dynamics MAE 331, 2018 Learning Objectives Methods of time-domain analysis Continuous- and discrete-time models Transient
More informationKalman Filter. Predict: Update: x k k 1 = F k x k 1 k 1 + B k u k P k k 1 = F k P k 1 k 1 F T k + Q
Kalman Filter Kalman Filter Predict: x k k 1 = F k x k 1 k 1 + B k u k P k k 1 = F k P k 1 k 1 F T k + Q Update: K = P k k 1 Hk T (H k P k k 1 Hk T + R) 1 x k k = x k k 1 + K(z k H k x k k 1 ) P k k =(I
More informationCPE 310: Numerical Analysis for Engineers
CPE 310: Numerical Analysis for Engineers Chapter 2: Solving Sets of Equations Ahmed Tamrawi Copyright notice: care has been taken to use only those web images deemed by the instructor to be in the public
More informationCross Directional Control
Cross Directional Control Graham C. Goodwin Day 4: Lecture 4 16th September 2004 International Summer School Grenoble, France 1. Introduction In this lecture we describe a practical application of receding
More informationKalman Filtering in a Mass-Spring System
Department of Mathematics Center for Quantitative Sciences in Biomedicine North Carolina State University July 30, 2016 Practical Overview Goal: Implement Kalman filter for linear system 1. Solve a linear
More informationEstimation and Prediction Scenarios
Recursive BLUE BLUP and the Kalman filter: Estimation and Prediction Scenarios Amir Khodabandeh GNSS Research Centre, Curtin University of Technology, Perth, Australia IUGG 2011, Recursive 28 June BLUE-BLUP
More informationData Fusion of Dual Foot-Mounted Zero Velocity Update (ZUPT) Aided Inertial Navigation Systems (INSs) using Centroid Method
February 02, 2013 Data Fusion of Dual Foot-Mounted Zero Velocity Update (ZUPT) Aided Inertial Navigation Systems (INSs) using Centroid Method Girisha Under the guidance of Prof. K.V.S. Hari Notations Define
More informationRelative Merits of 4D-Var and Ensemble Kalman Filter
Relative Merits of 4D-Var and Ensemble Kalman Filter Andrew Lorenc Met Office, Exeter International summer school on Atmospheric and Oceanic Sciences (ISSAOS) "Atmospheric Data Assimilation". August 29
More informationSingular value decomposition. If only the first p singular values are nonzero we write. U T o U p =0
Singular value decomposition If only the first p singular values are nonzero we write G =[U p U o ] " Sp 0 0 0 # [V p V o ] T U p represents the first p columns of U U o represents the last N-p columns
More informationLeast Squares and Kalman Filtering Questions: me,
Least Squares and Kalman Filtering Questions: Email me, namrata@ece.gatech.edu Least Squares and Kalman Filtering 1 Recall: Weighted Least Squares y = Hx + e Minimize Solution: J(x) = (y Hx) T W (y Hx)
More informationAvailable online at ScienceDirect. Energy Procedia 94 (2016 )
Available online at www.sciencedirect.com ScienceDirect Energy Procedia 94 (216 ) 191 198 13th Deep Sea Offshore Wind R&D Conference, EERA DeepWind 216, 2-22 January 216, Trondheim, Norway Vibration-based
More informationState Estimation using Gaussian Process Regression for Colored Noise Systems
State Estimation using Gaussian Process Regression for Colored Noise Systems Kyuman Lee School of Aerospace Engineering Georgia Institute of Technology Atlanta, GA 30332 404-422-3697 lee400@gatech.edu
More informationFrequentist-Bayesian Model Comparisons: A Simple Example
Frequentist-Bayesian Model Comparisons: A Simple Example Consider data that consist of a signal y with additive noise: Data vector (N elements): D = y + n The additive noise n has zero mean and diagonal
More informationMinimization with Equality Constraints!
Path Constraints and Numerical Optimization Robert Stengel Optimal Control and Estimation, MAE 546, Princeton University, 2017 State and Control Equality Constraints Pseudoinverse State and Control Inequality
More informationDATA ASSIMILATION FOR FLOOD FORECASTING
DATA ASSIMILATION FOR FLOOD FORECASTING Arnold Heemin Delft University of Technology 09/16/14 1 Data assimilation is the incorporation of measurement into a numerical model to improve the model results
More informationMathematical Concepts of Data Assimilation
Mathematical Concepts of Data Assimilation N.K. Nichols 1 Introduction Environmental systems can be realistically described by mathematical and numerical models of the system dynamics. These models can
More informationInitial estimates for R 0 and Γ0 R Γ. Γ k. State estimate R Error covariance. k+1 k+1
HERA-B -32 VDET -1 CATS track fitting algorithm based on the discrete Kalman filter D. Emeliyanov 12 I. Kisel 34 1 Institute of Control Sciences, Profsoyuznaya ul., 65, Moscow, 11786 Russia 2 Deutsches
More informationLinear Dynamical Systems
Linear Dynamical Systems Sargur N. srihari@cedar.buffalo.edu Machine Learning Course: http://www.cedar.buffalo.edu/~srihari/cse574/index.html Two Models Described by Same Graph Latent variables Observations
More informationEnsemble square-root filters
Ensemble square-root filters MICHAEL K. TIPPETT International Research Institute for climate prediction, Palisades, New Yor JEFFREY L. ANDERSON GFDL, Princeton, New Jersy CRAIG H. BISHOP Naval Research
More informationData Fusion Kalman Filtering Self Localization
Data Fusion Kalman Filtering Self Localization Armando Jorge Sousa http://www.fe.up.pt/asousa asousa@fe.up.pt Faculty of Engineering, University of Porto, Portugal Department of Electrical and Computer
More information1 Introduction. Systems 2: Simulating Errors. Mobile Robot Systems. System Under. Environment
Systems 2: Simulating Errors Introduction Simulating errors is a great way to test you calibration algorithms, your real-time identification algorithms, and your estimation algorithms. Conceptually, the
More informationLinear regression methods
Linear regression methods Most of our intuition about statistical methods stem from linear regression. For observations i = 1,..., n, the model is Y i = p X ij β j + ε i, j=1 where Y i is the response
More informationRECURSIVE IMPLEMENTATIONS OF THE SCHMIDT-KALMAN CONSIDER FILTER
RECURSIVE IMPLEMENTATIONS OF THE SCHMIDT-KALMAN CONSIDER FILTER Renato Zanetti and Christopher D Souza One method to account for parameters errors in the Kalman filter is to consider their effect in the
More informationFirst-Order Low-Pass Filter
Filters, Cost Functions, and Controller Structures Robert Stengel Optimal Control and Estimation MAE 546 Princeton University, 218! Dynamic systems as low-pass filters! Frequency response of dynamic systems!
More informationApplications Linear Control Design Techniques in Aircraft Control I
Lecture 29 Applications Linear Control Design Techniques in Aircraft Control I Dr. Radhakant Padhi Asst. Professor Dept. of Aerospace Engineering Indian Institute of Science - Bangalore Topics Brief Review
More informationCS 532: 3D Computer Vision 6 th Set of Notes
1 CS 532: 3D Computer Vision 6 th Set of Notes Instructor: Philippos Mordohai Webpage: www.cs.stevens.edu/~mordohai E-mail: Philippos.Mordohai@stevens.edu Office: Lieb 215 Lecture Outline Intro to Covariance
More information2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008
MIT OpenCourseWare http://ocw.mit.edu 2.830J / 6.780J / ESD.63J Control of Processes (SMA 6303) Spring 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
More informationSystematic strategies for real time filtering of turbulent signals in complex systems
Systematic strategies for real time filtering of turbulent signals in complex systems Statistical inversion theory for Gaussian random variables The Kalman Filter for Vector Systems: Reduced Filters and
More informationAircraft Flight Dynamics!
Aircraft Flight Dynamics Robert Stengel MAE 331, Princeton University, 2016 Course Overview Introduction to Flight Dynamics Math Preliminaries Copyright 2016 by Robert Stengel. All rights reserved. For
More informationThe Ensemble Kalman Filter:
p.1 The Ensemble Kalman Filter: Theoretical formulation and practical implementation Geir Evensen Norsk Hydro Research Centre, Bergen, Norway Based on Evensen, Ocean Dynamics, Vol 5, No p. The Ensemble
More informationHomework #1 Solution
February 7, 4 Department of Electrical and Computer Engineering University of Wisconsin Madison ECE 734 VLSI Array Structures for Digital Signal Processing Homework # Solution Due: February 6, 4 in class.
More informationLecture 7. Gaussian Elimination with Pivoting. David Semeraro. University of Illinois at Urbana-Champaign. February 11, 2014
Lecture 7 Gaussian Elimination with Pivoting David Semeraro University of Illinois at Urbana-Champaign February 11, 2014 David Semeraro (NCSA) CS 357 February 11, 2014 1 / 41 Naive Gaussian Elimination
More informationDecember 20, MAA704, Multivariate analysis. Christopher Engström. Multivariate. analysis. Principal component analysis
.. December 20, 2013 Todays lecture. (PCA) (PLS-R) (LDA) . (PCA) is a method often used to reduce the dimension of a large dataset to one of a more manageble size. The new dataset can then be used to make
More informationKalman filter using the orthogonality principle
Appendix G Kalman filter using the orthogonality principle This appendix presents derivation steps for obtaining the discrete Kalman filter equations using a method based on the orthogonality principle.
More informationTSRT14: Sensor Fusion Lecture 8
TSRT14: Sensor Fusion Lecture 8 Particle filter theory Marginalized particle filter Gustaf Hendeby gustaf.hendeby@liu.se TSRT14 Lecture 8 Gustaf Hendeby Spring 2018 1 / 25 Le 8: particle filter theory,
More informationLeast Squares Estimation Namrata Vaswani,
Least Squares Estimation Namrata Vaswani, namrata@iastate.edu Least Squares Estimation 1 Recall: Geometric Intuition for Least Squares Minimize J(x) = y Hx 2 Solution satisfies: H T H ˆx = H T y, i.e.
More information1 Kalman Filter Introduction
1 Kalman Filter Introduction You should first read Chapter 1 of Stochastic models, estimation, and control: Volume 1 by Peter S. Maybec (available here). 1.1 Explanation of Equations (1-3) and (1-4) Equation
More informationSIMULTANEOUS STATE AND PARAMETER ESTIMATION USING KALMAN FILTERS
ECE5550: Applied Kalman Filtering 9 1 SIMULTANEOUS STATE AND PARAMETER ESTIMATION USING KALMAN FILTERS 9.1: Parameters versus states Until now, we have assumed that the state-space model of the system
More informationECE295, Data Assimila0on and Inverse Problems, Spring 2015
ECE295, Data Assimila0on and Inverse Problems, Spring 2015 1 April, Intro; Linear discrete Inverse problems (Aster Ch 1 and 2) Slides 8 April, SVD (Aster ch 2 and 3) Slides 15 April, RegularizaFon (ch
More informationAdvanced Introduction to Machine Learning CMU-10715
Advanced Introduction to Machine Learning CMU-10715 Gaussian Processes Barnabás Póczos http://www.gaussianprocess.org/ 2 Some of these slides in the intro are taken from D. Lizotte, R. Parr, C. Guesterin
More informationChapter 4 Systems of Linear Equations; Matrices
Chapter 4 Systems of Linear Equations; Matrices Section 5 Inverse of a Square Matrix Learning Objectives for Section 4.5 Inverse of a Square Matrix The student will be able to identify identity matrices
More informationFlying Qualities Criteria Robert Stengel, Aircraft Flight Dynamics MAE 331, 2018
Flying Qualities Criteria Robert Stengel, Aircraft Flight Dynamics MAE 331, 2018 Learning Objectives MIL-F-8785C criteria CAP, C*, and other longitudinal criteria ϕ/β, ω ϕ /ω d, and other lateral-directional
More informationTranslational and Rotational Dynamics!
Translational and Rotational Dynamics Robert Stengel Robotics and Intelligent Systems MAE 345, Princeton University, 217 Copyright 217 by Robert Stengel. All rights reserved. For educational use only.
More informationAircraft Flight Dynamics Robert Stengel MAE 331, Princeton University, 2018
Aircraft Flight Dynamics Robert Stengel MAE 331, Princeton University, 2018 Course Overview Introduction to Flight Dynamics Math Preliminaries Copyright 2018 by Robert Stengel. All rights reserved. For
More informationPrinciples of the Global Positioning System Lecture 11
12.540 Principles of the Global Positioning System Lecture 11 Prof. Thomas Herring http://geoweb.mit.edu/~tah/12.540 Statistical approach to estimation Summary Look at estimation from statistical point
More information4. DATA ASSIMILATION FUNDAMENTALS
4. DATA ASSIMILATION FUNDAMENTALS... [the atmosphere] "is a chaotic system in which errors introduced into the system can grow with time... As a consequence, data assimilation is a struggle between chaotic
More informationPrincipal Component Analysis
B: Chapter 1 HTF: Chapter 1.5 Principal Component Analysis Barnabás Póczos University of Alberta Nov, 009 Contents Motivation PCA algorithms Applications Face recognition Facial expression recognition
More informationA Comparison of Multiple-Model Target Tracking Algorithms
University of New Orleans ScholarWors@UNO University of New Orleans heses and Dissertations Dissertations and heses 1-17-4 A Comparison of Multiple-Model arget racing Algorithms Ryan Pitre University of
More informationDerivation of the Kalman Filter
Derivation of the Kalman Filter Kai Borre Danish GPS Center, Denmark Block Matrix Identities The key formulas give the inverse of a 2 by 2 block matrix, assuming T is invertible: T U 1 L M. (1) V W N P
More informationMatrix notation. A nm : n m : size of the matrix. m : no of columns, n: no of rows. Row matrix n=1 [b 1, b 2, b 3,. b m ] Column matrix m=1
Matrix notation A nm : n m : size of the matrix m : no of columns, n: no of rows Row matrix n=1 [b 1, b 2, b 3,. b m ] Column matrix m=1 n = m square matrix Symmetric matrix Upper triangular matrix: matrix
More informationLecture 9. Errors in solving Linear Systems. J. Chaudhry (Zeb) Department of Mathematics and Statistics University of New Mexico
Lecture 9 Errors in solving Linear Systems J. Chaudhry (Zeb) Department of Mathematics and Statistics University of New Mexico J. Chaudhry (Zeb) (UNM) Math/CS 375 1 / 23 What we ll do: Norms and condition
More informationFinal Review Written by Victoria Kala SH 6432u Office Hours R 12:30 1:30pm Last Updated 11/30/2015
Final Review Written by Victoria Kala vtkala@mathucsbedu SH 6432u Office Hours R 12:30 1:30pm Last Updated 11/30/2015 Summary This review contains notes on sections 44 47, 51 53, 61, 62, 65 For your final,
More informationSquare-Root Algorithms of Recursive Least-Squares Wiener Estimators in Linear Discrete-Time Stochastic Systems
Proceedings of the 17th World Congress The International Federation of Automatic Control Square-Root Algorithms of Recursive Least-Squares Wiener Estimators in Linear Discrete-Time Stochastic Systems Seiichi
More informationInterpretation of two error statistics estimation methods: 1 - the Derozier s method 2 the NMC method (lagged forecast)
Interpretation of two error statistics estimation methods: 1 - the Derozier s method 2 the NMC method (lagged forecast) Richard Ménard, Yan Yang and Yves Rochon Air Quality Research Division Environment
More informationTSRT14: Sensor Fusion Lecture 9
TSRT14: Sensor Fusion Lecture 9 Simultaneous localization and mapping (SLAM) Gustaf Hendeby gustaf.hendeby@liu.se TSRT14 Lecture 9 Gustaf Hendeby Spring 2018 1 / 28 Le 9: simultaneous localization and
More informationA KALMAN FILTERING TUTORIAL FOR UNDERGRADUATE STUDENTS
A KALMAN FILTERING TUTORIAL FOR UNDERGRADUATE STUDENTS Matthew B. Rhudy 1, Roger A. Salguero 1 and Keaton Holappa 2 1 Division of Engineering, Pennsylvania State University, Reading, PA, 19610, USA 2 Bosch
More informationReview: Linear and Vector Algebra
Review: Linear and Vector Algebra Points in Euclidean Space Location in space Tuple of n coordinates x, y, z, etc Cannot be added or multiplied together Vectors: Arrows in Space Vectors are point changes
More informationAlgorithm for Multiple Model Adaptive Control Based on Input-Output Plant Model
BULGARIAN ACADEMY OF SCIENCES CYBERNEICS AND INFORMAION ECHNOLOGIES Volume No Sofia Algorithm for Multiple Model Adaptive Control Based on Input-Output Plant Model sonyo Slavov Department of Automatics
More informationStochastic Models, Estimation and Control Peter S. Maybeck Volumes 1, 2 & 3 Tables of Contents
Navtech Part #s Volume 1 #1277 Volume 2 #1278 Volume 3 #1279 3 Volume Set #1280 Stochastic Models, Estimation and Control Peter S. Maybeck Volumes 1, 2 & 3 Tables of Contents Volume 1 Preface Contents
More informationA technique for simultaneous parameter identification and measurement calibration for overhead transmission lines
A technique for simultaneous parameter identification and measurement calibration for overhead transmission lines PAVEL HERING University of West Bohemia Department of cybernetics Univerzitni 8, 36 4 Pilsen
More information