Basic Definitions and The Spectral Estimation Problem

Lecture 1

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L1–1

Informal Definition of Spectral Estimation

Given: A finite record of a signal. Determine: The distribution of signal power over frequency.
signal t ω t=1, 2, ... ω ω+∆ω π spectral density

! = (angular) frequency in radians/(sampling interval) f = !=2 = frequency in cycles/(sampling interval)

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L1–2

Applications Temporal Spectral Analysis Vibration monitoring and fault detection Hidden periodicity finding Speech processing and audio devices Medical diagnosis Seismology and ground movement study Control systems design Radar, Sonar

Spatial Spectral Analysis Source location using sensor arrays
Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997 Slide L1–3

Deterministic Signals fy(t)g1 .1 y(t)e. Moses.i!t exists and is called the Discrete-Time Fourier Transform (DTFT) Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall.1 = t= If: discrete-time deterministic data sequence X 1 t=. Stoica and R. 1997 Slide L1–4 .1 jy(t)j2 < 1 X Then: Y (! ) = 1 t=.

i!k (k) = 1 t=. S (!)d! t=. 1997 . Stoica and R.1 X Z 1 where 4 S (!) = jY (!)j2 = Energy Spectral Density We can write S (!) = where X 1 k=. Moses.1 y(t)y (t .Energy Spectral Density Parseval's Equality: jy(t)j2 = 21 . k) Slide L1–5 Lecture notes to accompany Introduction to Spectral Analysis by P.1 X (k)e. Prentice Hall.

Moses. Stoica and R.1 n jy(t)j2 = 1 o But: E jy(t)j2 < 1 E f g = Expectation over the ensemble of realizations E jy(t)j2 n o = Average power in y (t) PSD = (Average) power spectral density Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. 1997 Slide L1–6 .Random Signals Random Signal random signal probabilistic statements about future variations t current observation time Here: X 1 t=.

Moses. k)g r(k) = r (. Stoica and R.k) r(0) jr(k)j Z Note that r(k) = 21 .1 r(k)e. Prentice Hall. (!)ei!k d! n o (Inverse DTFT) Interpretation: so r(0) = E jy(t)j2 = 21 .i!k (k) is the autocovariance sequence (ACS) r(k) = E fy(t)y (t . (!)d! Z (!)d! = infinitesimal signal power in the band ! d! 2 Lecture notes to accompany Introduction to Spectral Analysis by P.First Definition of PSD (! ) = where r X 1 k=. 1997 Slide L1–7 .

i!t Lecture notes to accompany Introduction to Spectral Analysis by P.i!t (!) = Nlim E N !1 t=1 X > : 8 > < 2 9 > = > Note that 1 jY (!)j2 (!) = Nlim E N N !1 YN (!) = X where N is the finite DTFT of fy(t)g. 1997 Slide L1–8 . Prentice Hall. t=1 y(t)e. Moses.Second Definition of PSD 1 N y(t)e. Stoica and R.

Moses. Thus.! ) P3: If y Lecture notes to accompany Introduction to Spectral Analysis by P. Then: (! ) = (. 1997 Slide L1–9 . Prentice Hall.1=2 1=2] (!) 0 (t) is real. we can restrict attention to !2 . P2: ] () f 2 .! ) Otherwise: (! ) 6= (.Properties of the PSD P1: (!) = (! + 2 ) for all !. Stoica and R.

i!k H (! ) = X 1 y (!) = jH (!)j2 e(!) k=0 Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall.1 H (q) = X 1 k=0 hk q.Transfer of PSD Through Linear Systems System Function: where q . k) hk e.k = unit delay operator: q. 1997 Slide L1–10 . Stoica and R.1y(t) = y(t . 1) H (q ) y (t) 2 y (! ) = jH (! )j e( e(t) e (! ) Then !) y(t) = X 1 k=0 hk e(t . Moses.

]g Find an estimate of Two Main Approaches : Nonparametric: – Derived from the PSD definitions. Stoica and R. Prentice Hall. Moses. 1997 Slide L1–11 . Parametric: – Assumes a parameterized functional form of the PSD Lecture notes to accompany Introduction to Spectral Analysis by P.The Spectral Estimation Problem The Problem: From a sample fy(1) : : : y(N )g (!): f ^(!) ! 2 .

Periodogram and Correlogram Methods Lecture 2 Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. Prentice Hall. 1997 Slide L2–1 . Stoica and R.

1997 Slide L2–2 . Prentice Hall.i!t N t=1 Natural estimator N 2 Used by Schuster ( 1900) to determine “hidden periodicities” (hence the name).Periodogram Recall 2nd definition of (! ): 1 N y(t)e. Stoica and R. Moses. Lecture notes to accompany Introduction to Spectral Analysis by P.i!t (!) = Nlim E N !1 t=1 X > : 8 > < 2 9 > = > Given : Drop “ fy(t)gN t=1 lim ” and “E N !1 f g” to get X ^p(!) = 1 y(t)e.

Correlogram Recall 1st definition of (! ): (! ) = P X 1 Truncate the “ ” and replace “r k=. 1997 Slide L2–3 . Prentice Hall. Moses.1 N .1 X r(k)e. Stoica and R.i!k ^c(!) = k=.(N .1) Lecture notes to accompany Introduction to Spectral Analysis by P.i!k (k)” by “^(k)”: r r ^(k)e.

k t=k+1 X Standard biased estimate: 1 N y(t)y (t .Covariance Estimators (or Sample Covariances) Standard unbiased estimate: 1 N y(t)y (t . k) k 0 r(k) = N ^ t=k+1 X For both estimators: r ^(k) = ^ (. Stoica and R. Moses. 1997 Slide L2–4 .k) k < 0 r Lecture notes to accompany Introduction to Spectral Analysis by P. k) k 0 ^ r(k) = N . Prentice Hall.

1997 . Moses.(N . N ^p(!) = 1 y(t)e.i!k ^p(!) = ^c(!) Consequence: Both p ! and ^( ) ^c(!) can be analyzed simultaneously.i!t N t=1 X 2 = N . Stoica and R.Relationship Between ^p (! ) and ^c(! ) If: the biased ACS estimator r Then: ^(k) is used in ^c(!).1 X = ^c(!) k=. Prentice Hall.1) r ^(k)e. Slide L2–5 Lecture notes to accompany Introduction to Spectral Analysis by P.

r(k) may be large for large jkj ^ Even if the errors r k r k Nj. p! f^( ) . Prentice Hall. 1997 Slide L2–6 . ( )g ^ ( ) . Stoica and R. jk =0 there are “so many” that when summed in ! . Thus. ^p(!) and ^c(!) have poor performance.1 are small. Moses. even for large N . Intuitive explanation: r(k) . the PSD error is large. ( )] Lecture notes to accompany Introduction to Spectral Analysis by P.Statistical Performance of ^p(! ) and ^c(! ) Summary: Both are asymptotically (for large N ) unbiased: E ^p(!) ! (!) as N ! 1 n o Both have “large” variance.

(N .Bias Analysis of the Periodogram E ^p(!) = E ^c(!) = E f^(k)g e. Moses. or triangular. window E ^p(!) = 21 . ) d Z Ideally: WB (!) = Dirac impulse (!).i!k k=. 1 jkj N Bartlett. ( )WB (! .1 X wB (k) = 8 < : = Thus. 1997 .i!k r k=. Prentice Hall. N k=.1) N . Stoica and R.(N .i!k = 1. jNj jkj N .1 n o n o X ! X N .1) 1 = wB (k)r(k)e. Slide L2–7 Lecture notes to accompany Introduction to Spectral Analysis by P. n o 0 k 1 .1 jkj r(k)e.

Prentice Hall. Moses. WB 1=N . 1997 . Slide L2–8 Lecture notes to accompany Introduction to Spectral Analysis by P.Bartlett Window WB (! ) 1 sin(!N=2) 2 WB (!) = N sin(!=2) " # 0 WB (!)=WB (0). for N = 25 −10 −20 dB −30 −40 −50 −60 −3 −2 −1 0 1 ANGULAR FREQUENCY 2 3 Main lobe 3dB width For “small” N . Stoica and R. (!) may differ quite a bit from (!).

φ(ω) smearing ω ω (!) separated in f by less than 1=N are not ^ φ(ω) <1/Ν Thus: Periodogram resolution limit = 1=N .Smearing and Leakage Main Lobe Width: smearing or smoothing Details in resolvable. ^ φ(ω)≅W (ω) B Sidelobe Level: leakage φ(ω)≅δ(ω) leakage ω ω Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L2–9 . Stoica and R. Prentice Hall. Moses.

Prentice Hall. WB ! !. severe bias As N . Moses. !1 ^( ) ( )! ( ) Lecture notes to accompany Introduction to Spectral Analysis by P.Periodogram Bias Properties Summary of Periodogram Bias Properties: For “small” N . so ! is asymptotically unbiased. 1997 Slide L2–10 . Stoica and R.

Prentice Hall. (!2) 2(!1) !1 = !2 = 0 !1 6= !2 nh ih ( io Inconsistent estimate Erratic behavior ^ φ(ω) ^ φ(ω) asymptotic mean = φ(ω) ω + 1 st. dev = φ(ω) too - Resolvability properties depend on both bias and variance. 1997 Slide L2–11 . Stoica and R. (!1) ^p(!2) .Periodogram Variance As N !1 E ^p(!1) . Lecture notes to accompany Introduction to Spectral Analysis by P. Moses.

i!t 2 k and W = e. 1 Direct computation of O N 2 flops ( ) .i 2 .1 fY (k)gN=0 from fy(t)gN : t=1 k Lecture notes to accompany Introduction to Spectral Analysis by P.Discrete Fourier Transform (DFT) Finite DTFT: YN (!) = X N t=1 y(t)e. N Let ! = N Then YN ( 2 k) is the Discrete Fourier Transform (DFT): N X Y (k) = N t=1 y(t)W tk k = 0 ::: N . Stoica and R. Prentice Hall. 1997 Slide L2–12 . Moses.

Stoica and R. 1997 . Prentice Hall. 2 = 2p: Y (2p) = For k X ~ N = 1 3 5 : : : N .1 for odd k Assume: X X X Let N For k ~ ~ = N=2 and W = W 2 = e. ~ 4 = 0 2 4 : : : N .i2 =N . Moses. 1 = 2p + 1: X t=1 ~ ~ y(t) + y(t + N )]W tp Y (2p + 1) = Each is a N ~ N ~ = N=2-point DFT computation. y(t + N )]W tgW tp Lecture notes to accompany Introduction to Spectral Analysis by P.Radix–2 Fast Fourier Transform (FFT) N = 2m N=2 N Y (k) = y(t)W tk + y(t)W tk t=1 t=N=2+1 N=2 2 = y(t) + y(t + N=2)W Nk ]W tk t=1 Nk +1 for even k with W 2 = . Slide L2–13 t=1 ~ ~ f y(t) .

Stoica and R.1 k2k ) ck = 2 ck = 1 N log2 N 2 Thus. 1997 Slide L2–14 . Prentice Hall. 2k + 2c ck = 2 k.FFT Computation Count Let ck Then = number of flops for N = 2k point FFT. Lecture notes to accompany Introduction to Spectral Analysis by P. Moses.

1997 Slide L2–15 . Prentice Hall. N=8 ω 2 k N . Stoica and R. Moses.1 N k=0 sampled.1 ! N k=0 ^ φ(ω) continuous curve Lecture notes to accompany Introduction to Spectral Analysis by P.Zero Padding Append the given data by zeros prior to computing DFT (or FFT): fy(1) : : : y(N ) 0 : : : 0g | N {z } Goals: Apply a radix-2 FFT (so N Finer sampling of = power of 2) ^(!): 2 k N .

Moses. Stoica and R. 1997 Slide L3–1 . Prentice Hall.Improved Periodogram-Based Methods Lecture 3 Lecture notes to accompany Introduction to Spectral Analysis by P.

(M . Prentice Hall. ^( ) jj ^BT (!) = M .1 X k=.1) w(k)^(k)e.i!k r fw(k)g = Lag Window w(k) 1 -M M k Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. Stoica and R.Blackman-Tukey Method Basic Idea: Weighted correlogram. with small weight applied to covariances r k with “large” k . 1997 Slide L3–2 .

Prentice Hall.Blackman-Tukey Method. 1997 . Stoica and R. )d 2 . Moses. con't ^BT (!) = 1 ^p( )W (! . Z W (!) = = Conclusion: Effect: DTFT fw(k)g Spectral Window ^BT (!) = “locally” smoothed periodogram Variance decreases substantially Bias increases slightly By proper choice of M : MSE = var + bias2 ! 0 as N ! 1 Slide L3–3 Lecture notes to accompany Introduction to Spectral Analysis by P.

W (!)d! = equiv bandwidth e= W (0) Ne e = 1 Lecture notes to accompany Introduction to Spectral Analysis by P.1 X .(M (0) w Z w (k ) = equiv time width 1 2 . Stoica and R. 0 Z | {z } If W (!) 0 (. )d 2 . 1997 Slide L3–4 .1) Ne = k=. w(k) is a psd sequence) ^BT (!) 0 (which is desirable) Then: Time-Bandwidth Product M .Window Design Considerations Nonnegativeness: ^BT (!) = 1 ^p( ) W (! . Prentice Hall. Moses.

As M increases.Window Design. Ne (and hence fixed. Lecture notes to accompany Introduction to Spectral Analysis by P. ) Choose M as a tradeoff between variance and bias. Moses. Stoica and R. The energy in the main lobe and in the sidelobes cannot be reduced simultaneously. once M is given. Once M is given. bias decreases and variance increases. Prentice Hall. e) is essentially ) Choose window shape to compromise between smearing (main lobe width) and leakage (sidelobe level). 1997 Slide L3–5 . con't e = 1=Ne = 0(1=M ) is the BT resolution threshold.

M 0 −10 = 25 −20 dB −30 −40 −50 −60 −3 −2 −1 0 1 ANGULAR FREQUENCY 2 3 Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L3–6 . Moses. M 0 = 25 −10 −20 dB −30 −40 −50 −60 −3 −2 −1 0 1 ANGULAR FREQUENCY 2 3 Rectangular Window.Window Examples Triangular Window. Prentice Hall. Stoica and R.

Bartlett Method Basic Idea:
1 Μ 2Μ ... Ν

^1(ω) ^2(ω) φ φ

... average

^L(ω) φ

^B(ω) φ
Mathematically:

yj (t) = y((j ; 1)M + t) t = 1 : : : M = the j th subsequence 4 (j = 1 : : : L = N=M ])

^j (!) = 1 yj (t)e;i!t M t=1
X

M

2

1 L ^ (!) ^B (!) = L j =1 j
X

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L3–7

Comparison of Bartlett and Blackman-Tukey Estimates

1 L ^B (!) = L j =1
X

8 > < > :

M ;1 X

=

'
Thus:

k=;(M ;1) M ;1 X k=;(M ;1)

k=;(M ;1) 8 M ;1 < 1 X L X
:

r ^j (k)e;i!k
9 =

9 > = >

L j =1

r ^j (k) e;i!k

r ^(k)e;i!k

^B (!) ' ^BT (!)
Since B ! implicitly uses method has

with a rectangular lag window wR k

()

^( )

fwR(k)g, the Bartlett

High resolution (little smearing) Large leakage and relatively large variance
Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997 Slide L3–8

Welch Method Similar to Bartlett method, but allow overlap of subsequences (gives more subsequences, and thus “better” averaging) use data window for each periodogram; gives mainlobe-sidelobe tradeoff capability

1

2 . . . subseq subseq #1 #2 subseq #S

N

Let S # of subsequences of length M . (Overlapping means S > N=M “better averaging”.)

=

])

Additional flexibility: The data in each subsequence are weighted by a temporal window Welch is approximately equal to non-rectangular lag window.

^BT (!) with a
Slide L3–9

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

k +J 1 ^D(!k ) = ^p (!j ) 2J + 1 j=k.1 ! = N j =0 Idea: “Local averaging” of (2J + 1) samples in the j (2 + 1) frequency domain should reduce the variance by about J .Daniell Method By a previous result. Prentice Hall. 1997 Slide L3–10 . f ^p(!j )g are (nearly) uncorrelated random variables for 2 j N . for N 1.J X Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses.

for N 1. Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. Prentice Hall.Daniell Method. 1997 Slide L3–11 . Then. Hence: ^D (! ) ' ^BT (! ) with a rectangular spectral Let Z window. Stoica and R. con't As J increases: Bias increases (more smoothing) Variance decreases (more averaging) = 2J=N . ^D(!) ' 1 ^p(!)d! 2 .

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. ^ () – Implemented by averaging subsample periodograms. ^p(!) by a suitably selected ^(k) using a lag – Implemented by truncating and weighting r window in c ! ^( ) Bartlett. Moses.Summary of Periodogram Methods Unwindowed periodogram – reasonable bias – unacceptable variance Modified periodograms – Attempt to reduce the variance at the expense of (slightly) increasing the bias. Welch periodograms – Approximate interpretation: BT ! with a suitable lag window (rectangular for Bartlett. ^BT (!) with a rectangular Slide L3–12 – Implemented by local averaging of periodogram values. Prentice Hall. Daniell Periodogram – Approximate interpretation: spectral window. 1997 . BT periodogram – Local smoothing/averaging of spectral window. more general for Welch).

Moses.Parametric Methods for Rational Spectra Lecture 4 Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L4–1 . Prentice Hall. Stoica and R.

provided m and n are chosen sufficiently large. Prentice Hall. 1997 Slide L4–2 ( ) .i! . Moses. ! can approximate arbitrarily well any continuous PSD. Stoica and R.i!k (!) is a rational function in e.θ) ^ θ Estimate PSD ^ ^ =φ(ω. however: choice of m and n is not simple some PSDs are not continuous Lecture notes to accompany Introduction to Spectral Analysis by P.i!k (!) = P jkj n k e. Note.θ) φ(ω) possibly revise assumption on φ(ω) Rational Spectra jkj m k e.Basic Idea of Parametric Spectral Estimation Observed Data Assumed functional form of φ(ω. P By Weierstrass theorem.θ) Estimate parameters in φ(ω.

Prentice Hall. Moses. e. and ARMA Models By Spectral Factorization theorem.n B (z) = 1 + b1z.1 + + anz.1 + + bmz. MA. A(! ) = A(z )jz =ei! Signal Modeling Interpretation: e(t) e (! ) = 2 - B (q) A(q) white noise B (!) 2 2 A(!) ltered white noise y(t)y (! ) = ARMA: AR: MA: A(q)y(t) = B (q)e(t) A(q)y(t) = e(t) y(t) = B (q)e(t) Slide L4–3 Lecture notes to accompany Introduction to Spectral Analysis by P..m and.AR. 1997 .g. Stoica and R. a rational factored as (!) can be B (!) 2 2 (!) = A(!) A(z) = 1 + a1z.

j )y (t .ARMA Covariance Structure ARMA signal model: y(t)+ X n i=1 n aiy(t . 1997 Slide L4–4 .k = j =0 = 0 for k > m B (q) = 1 h q. i) = m j =0 bj E fe(t . k) and take E f g to give: X r (k ) + X 2 m bj hj .k (h = 1) where H (q ) = A(q ) 0 k k=0 X X i=1 air(k . j ) (b0 = 1) Multiply by y (t . i) = X m j =0 bj e(t . k)g Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. Moses. Stoica and R.

. Lecture notes to accompany Introduction to Spectral Analysis by P. 2 1 a1 = 0 . Moses. Prentice Hall. . 1997 Slide L4–5 . r (. . r(n) : : : r(0) 1 = 2 R 0 2 6 6 6 4 7 7 7 5 " # " 32 6 6 6 4 3 7 7 7 5 # These are the Yule–Walker (YW) Equations.AR Signals: Yule-Walker Equations AR: m = 0. .1) : : : r(. . an 0 3 7 7 7 5 2 6 6 6 4 Writing covariance equation in matrix form for k : : : n: =1 r(0) r(. . r(1) r(0) . .1) . Stoica and R.n) .

1 = 0 a . . . Stoica and R.AR Spectral Estimation: YW Method Yule-Walker Method: Replace r 2 6 6 6 4 (k) by r(k) and solve for f^ig and ^2: ^ a r ^(0) ^(. r ^(n) : : : r(0) ^ ^n a 0 32 7 7 7 5 6 6 6 4 3 7 7 7 5 2 6 6 6 4 3 7 7 7 5 Then the PSD estimate is ^2 ^(!) = ^ 2 jA(!)j Lecture notes to accompany Introduction to Spectral Analysis by P. r .n) 1 r r ^2 . r ^(1) r(0) ^ ^.1) : : : ^(. . . ^(. . Prentice Hall. . 1997 Slide L4–6 .1) . Moses.

(Y Y ). y(N . Find i=1 aiy(t . 2) Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R.AR Spectral Estimation: LS Method Least Squares Method: e(t) = y(t) + X n 4 = y(t) + y(t) ^ where '(t) = y (t . . 1) y(N . ^ = . . n)]T . n) . i) = y(t) + 'T (t) = a1 : : : an]T to minimize f( ) = X N This gives 2 y=6 4 y(n + 1) y(n + 2) y(N ) . . 1) y(n + 1) y(n) . Prentice Hall. . Moses. .1(Y y) where 3 7 5 2 6 t=n+1 je(t)j2 y(1) y(2) y(N . 1997 Slide L4–7 . 3 7 5 Y =4 y(n) y(n . . 1) : : : y (t .

. order-recursive solution to YW equations 2 6 6 6 4 | 0 ..1 1 . 32 7 7 7 5 } 6 6 6 4 1 n 3 7 7 7 5 2 6 6 6 4 ^ k = either r(k) or r(k). 0 2 n 3 7 7 7 5 O(n3) flops = 1 : : : n: O(n4) flops Levinson–Durbin Algorithm: Exploits the Toeplitz form of Rn+1 to obtain the solutions for k : : : n in O n2 flops! =1 ( ) Lecture notes to accompany Introduction to Spectral Analysis by P.0 . 1997 Slide L4–8 . n .. {z . .1 ..n .. . . Moses. Prentice Hall. Stoica and R. . Direct Solution: For one given value of n: For k Rn+1 1 0 = 0 .. .Levinson–Durbin Algorithm Fast.

n 3 7 5 1 Rn+2 4 1 n = 0 3 5 2 4 Rn+1 n+1 r ~n n+1 r ~n 0 32 54 1 n n = 0 0 n 3 5 2 4 2 3 5 where r n = n+1 + ~n n Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. Moses. 1997 Slide L4–9 . Stoica and R.Levinson-Durbin Alg. . where x = xn : : : x1]T ~ ~ ~ Nested structure 2 Rn+2 = Thus. 2 4 Rn+1 n+1 r ~n n+1 r ~n 0 3 5 2 r ~n = 6 4 . con't Relevant Properties of R: Rx = y $ Rx = y.

n = n ) n+1 = 0 + kn 2+1 = n + kn n = n(1 . Prentice Hall. jknj2) 2 2 n " n # " ~n 1 # Computation count: k flops for the step k 2 ) n2 flops !k+1 2 to determine f k k gn=1 k Slide L4–10 This is O (n2) times faster than the direct solution. 1997 . Lecture notes to accompany Introduction to Spectral Analysis by P.Levinson-Durbin Alg. 2 kn = . Stoica and R. con't 2 Rn+2 4 82 < 4 +2 : 1 n n = 0 0 n 3 5 2 4 2 3 5 2 Rn+2 4 39 = 5 2 0 ~n = 0n 2 1 n 3 5 2 4 3 5 2 3 5 Combining these gives: Rn 0 1 n + kn ~n 0 1 3 5 2 4 = 4 n + kn n 2 0 n + kn 2 n = 4 n 2 +1 3 5 0 0 Thus. Moses.

m) Thus. Moses. Prentice Hall. 1) + + bme(t . r(k) = 0 for jkj > m and 2 2 = m r(k)e.i!k (!) = jB(!)j k=.MA Signals MA: n =0 y(t) = B (q)e(t) = e(t) + b1e(t . Stoica and R.m X Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L4–11 .

Lecture notes to accompany Introduction to Spectral Analysis by P. Estimate fbkg and 2 and insert them in (!) = jB(!)j2 2 nonlinear estimation problem ^(!) is guaranteed to be 0 2. Prentice Hall.i!k k=. 1997 Slide L4–12 . Insert sample covariances (! ) = This is length X m f^(k)g in: r r(k)e. Moses.m ^BT (!) with a rectangular lag window of 2m + 1. with AR model order n . ^(!) is not guaranteed to be 0 =0 Both methods are special cases of ARMA methods described below.MA Spectrum Estimation Two main ways to Estimate (! ): 1. Stoica and R.

j ) Lecture notes to accompany Introduction to Spectral Analysis by P. k)] g = E f A(q)y(t)] A(q)y(t . Prentice Hall.ARMA Signals ARMA models can represent spectra with both peaks (AR part) and valleys (MA part). Moses.m k e. 1997 Slide L4–13 . k)] g = X n X n j =0 p=0 aj ap r(k + p . A(q)y(t) = B (q)e(t) B (!) 2 = m=. Stoica and R.i!k k (!) = 2 A(!) jA(!)j2 P where k = E f B(q)e(t)] B(q)e(t .

Prentice Hall. Estimate fai bj A(!) nonlinear estimation problem.m k e jA(!)j2 linear estimation problem (the Modified Yule-Walker method). Stoica and R. Moses. 1997 Slide L4–14 . ^(!) is not guaranteed to be 0 Lecture notes to accompany Introduction to Spectral Analysis by P.i!k k=.ARMA Spectrum Estimation Two Methods: 2g in (!) = 2 B(!) 2 1. Estimate fai r(k)g in (!) = P m . can use an approximate linear two-stage least squares method ^(!) is guaranteed to be 0 2.

Moses. Prentice Hall. 1) + + K y(t . K ) 1a) Estimate the coefficients modelling techniques. for some large K e(t) ' y(t) + 1y(t . 1997 Slide L4–15 . Stoica and R. f kgK=1 by using AR k 1b) Estimate the noise sequence ^(t) = y(t) + ^1y(t .Two-Stage Least-Squares Method Assumption: The ARMA model is invertible: A e(t) = B(q) y(t) (q ) = y(t) + 1y(t . K ) ^2 = N 1 j^(t)j2 e N . 1) + e and its variance + ^K y(t . K t=K +1 X Lecture notes to accompany Introduction to Spectral Analysis by P. 2) + = AR(1) with j kj ! 0 as k ! 1 Step 1: Approximate. 1) + 2y(t .

Stoica and R. m)] e e = a1 : : : an b1 : : : bm]T Lecture notes to accompany Introduction to Spectral Analysis by P. n) e ^(t . 1997 Slide L4–16 . e A(q)y(t) ' B (q)^(t) e and obtain estimates of fai bj g by applying least squares Step 2: Replace techniques.y(t . Moses. Note that the ai and bj coefficients enter linearly in the above equation: y(t) . 1) : : : . 1) : : : ^(t . Prentice Hall. ^(t) ' .Two-Stage Least-Squares Method. y(t . con't fe(t)g by ^(t) in the ARMA equation.

= m + 1 ::: m + M a1 an . . Prentice Hall. Moses. least f^ g Note: For narrowband ARMA signals. 1) : : : r(m . . the accuracy of ai is often better for M > n Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. . i) = 0 k > m 3 7 5 " r(m) r(m + 1) . .6 4 Replace fr(k)g by f^(k)g and solve for faig. n + M ) .. . r(m + M ) r(m + 1) r(m + 2) 3 7 If M n. r f^ g . = f^ g Y W system of equations. . : : : r(m .. fast Levinson-type algorithms exist for obtaining ai . n + 1) r(m . If M > n overdetermined squares solution for ai . n + 2) . 1997 Slide L4–17 . .Modified Yule-Walker Method ARMA Covariance Equation: r(k) + 2 6 4 X n In matrix form for k i=1 air(k . 5 . . # 2 = r(m + M .

Moses.Summary of Parametric Methods for Rational Spectra Method AR: YW or LS low low-medium medium-high medium-high medium low-medium No No Yes Computational Burden low Accuracy medium Guarantee ^(! ) 0 ? Yes MA: BT ARMA: MYW ARMA: 2-Stage LS Use for Spectra with (narrow) peaks but no valley Broadband spectra possibly with valleys but no peaks Spectra with both peaks and (not too deep) valleys As above Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. 1997 Slide L4–18 . Prentice Hall.

1997 Slide L5–1 . Prentice Hall. Moses.Parametric Methods for Line Spectra — Part 1 Lecture 5 Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R.

Examples: communications radar. Prentice Hall. 1997 . Stoica and R. Moses. sonar geophysical seismology ARMA model is a poor approximation Better approximation by Discrete/Line Spectrum Models 2 6(2 1) 2 6 (!) 6 (2 2) 2 6 (2 2) 3 ! Slide L5–2 .Line Spectra Many applications have signals with (near) sinusoidal components. !1 !2 !3 - An “Ideal” line spectrum Lecture notes to accompany Introduction to Spectral Analysis by P.

] (realistic and mathematically convenient) A3: e(t) = circular white noise with variance 2 E fe(t)e (s)g = 2 t s E fe(t)e(s)g = 0 (can be achieved by “slow” sampling) Slide L5–3 Lecture notes to accompany Introduction to Spectral Analysis by P. superimposed in white noise of k 2. 1997 .Line Spectral Signal Model Signal Model: Sinusoidal components of frequencies !k and powers 2 . Moses. power f g f g y(t) = x(t) + e(t) t = 1 2 : : : n x(t) = k ei(!k t+ k ) k=1 xk (t) X | {z } Assumptions: A1: A2: k>0 !k 2 . Prentice Hall. uniformly distributed on . ] (prevents model ambiguities) f'k g = independent rv's. Stoica and R.

for p 6= j =2 . Stoica and R. for p = j e i'p e. r(k) = E fy(t)y (t .i'j o = E nei'p o E ne. k)g = n=1 2ei!pk + 2 k 0 p p P and (! ) = 2 X n p=1 2 (! . k) = 2 ei!pk p j p n o Hence.i'j o e Z n 1 i' d' 2 = 0.Covariance Function and PSD Note that: E E n i'p e. Moses.i'j o = 1. e E xp(t)xj (t . !p) + 2 p Slide L5–4 Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 . Prentice Hall.

^ ei!k t + residuals k Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. Stoica and R. 1997 Slide L5–5 . Moses.Parameter Estimation Estimate either: f!k k 'k gn=1 k 2 (Signal Model) f!k 2gn=1 2 k k Once (PSD Model) ^ Major Estimation Problem: f!k g f!kg are determined: ^ X f ^2g can be obtained by a least squares method from k n 2ei!pk + residuals ^ ^ r(k) = p p=1 OR: Both k and 'k can be derived by a least squares method from f^ g f^ g X y(t) = with n k= k=1 k ei'k .

B = eiN!1 eiN!n k 2 6 4 3 7 5 Lecture notes to accompany Introduction to Spectral Analysis by P. .Nonlinear Least Squares (NLS) Method min' g f! Let: X N k k k t=1 | y(t) . X n k=1 {z F (! ') k ei(!k t+'k ) 2 } = kei'k = 1 : : : n]T Y = y(1) : : : y(N )]T ei!1 ei!n . . . Moses. 1997 Slide L5–6 . Stoica and R. Prentice Hall.

(B B).1B Y This gives: ^ = (B B). Moses.1B Y ] B B] . Y B(B B).1B Y ^ ! Lecture notes to accompany Introduction to Spectral Analysis by P.1B Y !=^ ! and ! = arg max Y B (B B ). B k2 = . B ) = kY . B ) (Y .1B Y ] +Y Y .Nonlinear Least Squares (NLS) Method. (B B). con't Then: F = (Y . Prentice Hall. Stoica and R. 1997 Slide L5–7 .

! Difficult Implementation: The NLS cost function F is multimodal. Stoica and R. Moses. 1997 Slide L5–8 .NLS Properties Excellent Accuracy: 6 2 (for N var (^k ) = 3 2 ! N k Example: N = 300 SNRk = 2= 2 = 30 dB k q 1) Then var (^k ) 10. Prentice Hall. Lecture notes to accompany Introduction to Spectral Analysis by P. it is difficult to avoid convergence to local minima.5.

Unwindowed Periodogram as an Approximate NLS Method

For a single (complex) sinusoid, the maximum of the unwindowed periodogram is the NLS frequency estimate: Assume: Then:

n=1
X

B B=N
N t=1

BY=

y(t)e;i!t = Y (!)

(finite DTFT)

1 Y B (B B );1B Y = N jY (!)j2 = ^p(!) = (Unwindowed Periodogram)
So, with no approximation,

! = arg max ^p(!) ^ !

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L5–9

Unwindowed Periodogram as an Approximate NLS Method, con't

Assume: Then:

n>1

f!kgn=1 ' the ^ k
provided that

locations of the peaks of p !

^( )

n

largest

inf j!k ; !pj > 2 =N
which is the periodogram resolution limit. If better resolution desired then use a High/Super Resolution method.

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L5–10

High-Order Yule-Walker Method Recall:

y(t) = x(t) + e(t) =

k ei(!k t+'k ) +e(t) | {z } k=1 xk (t)
X

n

“Degenerate” ARMA equation for y (t):
n

(1 ; ei!k q;1)xk (t) = k ei(!kt+'k) ; ei!k ei !k(t;1)+'k] = 0
o

Let

B (q) = 1 +

4 bk q;k = A(q)A(q) k=1 A(q) = (1 ; ei!1 q;1) (1 ; ei!n q;1) A(q) = arbitrary
X

L

Then B

(q)x(t) 0 ) B (q)y(t) = B (q)e(t)
Slide L5–11

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

High-Order Yule-Walker Method. Stoica and R. f!kgn=1. n ^ k Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L5–12 . along with L . con't Estimation Procedure: Estimate f^igL=1 using an ARMA MYW technique b i ^( ) Roots of B q give “spurious” roots. Moses. Prentice Hall.

High-Order and Overdetermined YW Equations ARMA covariance: r (k ) + X L i=1 bir(k . i) = 0 k > L In matrix form for k 2 6 6 6 4 | = L + 1 ::: L + M 3 7 7 7 5 2 6 6 6 4 3 7 7 7 5 } : : : r(1) r(L + 1) : : : r(2) b = . Lecture notes to accompany Introduction to Spectral Analysis by P. . Moses. . Prentice Hall. . Stoica and R. 1997 Slide L5–13 . 1) : : : r(M ) r(L + M ) 4 4 = = {z } | {z r(L) r(L + 1) This is a high-order (if L > n) and overdetermined (if M > L) system of YW equations. . r(L + M . . r(L + 2) .

if the Minimum-Norm solution b is used. diagonal and nonsingular Thus.1U Important property: The additional (L . The Minimum-Norm solution is b = .High-Order and Overdetermined YW Equations. Prentice Hall. con't Fact: SVD of rank : ( )=n =U V U = (M n) with U U = In V = (n L) with V V = In = (n n).V . 1997 Slide L5–14 . y = . Moses. Stoica and R. Lecture notes to accompany Introduction to Spectral Analysis by P. n) spurious zeros of B (q ) are located strictly inside the unit circle. (U V )b = .

SVD of .V ^ . this is the price paid for increased accuracy when L > n.1U ^ ^ ^ b ^(q) that Then !k n=1 are found from the n zeroes of B k are closest to the unit circle. Lecture notes to accompany Introduction to Spectral Analysis by P. Compute ^ = . Prentice Hall. this approach may give spurious frequency estimates when L > n.HOYW Equations. f^ g When the SNR is low. Moses. V be defined similarly to U . r . Practical Solution Let ^= ^ ^ ^ ^ but made from f^(k)g instead of fr(k)g. V from the Let U . Stoica and R. 1997 Slide L5–15 . .

Stoica and R.Parametric Methods for Line Spectra — Part 2 Lecture 6 Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. 1997 Slide L6–1 . Moses.

i! : : : e. m + 1)]T e 4 R = E fy(t)~ (t)g = APA + 2I ~ y Then with 2 P = E fx(t)~ (t)g = ~ x 6 4 2 1 0 . m + 1) .. Stoica and R. 1997 Slide L6–2 . Moses. 1) (for m 3 7 7 7 5 n) 2 46 y(t) = 6 ~ 6 4 y(t . = Ax(t) + ~(t) ~ e where x(t) = x1(t) : : : xn(t)]T ~ ~(t) = e(t) : : : e(t .1)! ]T A = a(!1) : : : a(!n)] (m n) rank Note: Define (A) = n y(t) y(t . 0 2 n 3 7 5 Lecture notes to accompany Introduction to Spectral Analysis by P.. Prentice Hall. .The Covariance Matrix Equation Let: a(!) = 1 e.i(m.

n)) Thus. Stoica and R. 2 R = S G] 6 4 1 3 " . Moses. Prentice Hall..ng: orthonormal eigenvectors associated with f n+1 : : : m g S = s1 : : : sn] (m n) G = g1 : : : gm. m 7 5 S G # Lecture notes to accompany Introduction to Spectral Analysis by P.n] (m (m . 1997 Slide L6–3 .Eigendecomposition of R and Its Properties R = APA + 2I (m > n) Let: 1 2 ::: m: eigenvalues of R fs1 : : : sng: orthonormal eigenvectors associated with f 1 : : : ng fg1 : : : gm..

2 2 6 4 3 7 5 2 6 4 Note: RS = APA S + 2S = S 1 0 . con't As rank (APA ) = n: k > 2 k = 1 ::: n k = 2 k = n + 1 ::: m 0 1. 0 n 3 7 5 with C (since rank S . . =0 A G=0 Slide L6–4 Lecture notes to accompany Introduction to Spectral Analysis by P.1) = AC S = A(PA S ( ) = rank(A) = n).. Prentice Hall.Eigendecomposition of R and Its Properties. Stoica and R.. 1997 . Therefore. 2 .. since S G j j 6= 0 4 . Moses. = = nonsingular 0 n.

. Stoica and R.MUSIC Method a (!1) . f!kgn=1 are the unique solutions of k a (!)GG a(!) = 0. Moses. Prentice Hall. 1997 Slide L6–5 . Let: 1 N y(t)~ (t) ^ ~ y R = N t=m ^^ S G = S G made from the ^ eigenvectors of R X Lecture notes to accompany Introduction to Spectral Analysis by P. A G= G=0 a (!n) ) fa(!k )gn=1 ? R(G) k 2 6 4 3 7 5 Thus.

Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. a(z) = 1 z.(m.1)]T Note: Both variants of MUSIC may produce spurious frequency estimates.1 : : : z. 1997 Slide L6–6 .Spectral and Root MUSIC Methods Spectral MUSIC Method: f!kgn=1 = the locations of the n highest peaks of the ^ k “pseudo-spectrum” function: 1 ^G a(!) ! 2 .1)GG a(z) = 0 that are closest to the unit circle. Moses. a (!)G ^ Root MUSIC Method: ] f!kgn=1 = the angular positions of the n roots of: ^ k ^^ aT (z. Here. Stoica and R.

1)^1 = 0 g no problem with spurious frequency estimates computationally simple (much) less accurate than MUSIC with m n+1 Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. ) f!kgn=1 can be found from the roots of ^ k aT (z. Prentice Hall. If: =n+1 m=n+1 Then: ^ g G = ^1. 1997 Slide L6–7 .Pisarenko Method Pisarenko is a special case of MUSIC with m (the minimum possible value). Stoica and R.

that has minimum Euclidean norm. Stoica and R. with first element equal ^ ^ to one. Prentice Hall. Moses. g # Lecture notes to accompany Introduction to Spectral Analysis by P.Min-Norm Method Goals: Reduce computational burden. but only one vector in G (as in Pisarenko). R( ) Let " 1 = the vector in R(G). 1997 Slide L6–8 . and reduce risk of false frequency estimates. Uses m n (as in MUSIC).

Prentice Hall. con't Spectral Min-Norm f!gn=1 = ^k the locations of the n highest peaks in the “pseudo-spectrum” 2 1 = a (!) 1 ^ g " # Root Min-Norm f!gn=1 = ^k the angular positions of the n roots of the polynomial aT (z.Min-Norm Method. Lecture notes to accompany Introduction to Spectral Analysis by P.1) 1 ^ g " # that are closest to the unit circle. Stoica and R. 1997 Slide L6–9 . Moses.

(S S ).Let S ^= S g1 . k k2) = (S S ) ) (S S ). Prentice Hall.S(S S ).S =(1 . k k2) ) ^ = . Stoica and R.) ^^ I=S S= Multiplying the above equation by gives: (1 . for N 1. g " # Min-Norm solution: As: ^ = . 1997 Slide L6–10 . k k2) g Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. at least.1 exists iff = k k2 6= 1 (This holds.1 gm " # " # Min-Norm Method: Determining g ^ Then: 1 2 R(G) ) S 1 = 0 ^ ^ ^ ^ g g ) S ^= .1 = =(1 .1 g + S S .

determined as is uniquely = (S1S1).1]A Then A2 Also.1]S 2 6 4 | 3 7 5 Recall S = AC with jC j 6= 0. Then S2 = A2C = A1DC = S1 C . Stoica and R.i!1 0 . where e.1DC {z } So has the same eigenvalues as D ..1S1S2 Lecture notes to accompany Introduction to Spectral Analysis by P.1 0]S S2 = 0 Im.ESPRIT Method Let A1 = Im. Moses. let = A1D. Prentice Hall. D= 0 e.. 1997 Slide L6–11 .1 0]A A2 = 0 Im.i!n S1 = Im.

find S .1S1S2 ^^ ^^ f!kgn=1 = . ^ ^ ^ ^ The frequency estimates are found by: where f^kgn=1 are the eigenvalues of k ^ = (S1S1). 1997 Slide L6–12 .ESPRIT Implementation From the eigendecomposition of R. then S1 and S2 . Stoica and R. Moses. Prentice Hall. arg(^k ) ^ k ESPRIT Advantages: computationally simple no extraneous frequency estimates (unlike in MUSIC or Min–Norm) accurate frequency estimates Lecture notes to accompany Introduction to Spectral Analysis by P.

Prentice Hall.Summary of Frequency Estimation Methods Method Periodogram Nonlinear LS Yule-Walker Pisarenko MUSIC Min-Norm ESPRIT Computational Accuracy / Risk for False Burden Resolution Freq Estimates small medium-high medium very high very high very high medium high medium small low none high high medium medium high small medium very high none Recommendation: Use Periodogram for medium-resolution applications Use ESPRIT for high-resolution applications Slide L6–13 Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. 1997 . Moses.

1997 Slide L7–1 . Moses. Prentice Hall.Filter Bank Methods Lecture 7 Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R.

by assuming that ! is nearly constant over the bands f ( )g ( ) !. ( =2 2 =1 N > 1 leads to the variability / resolution compromise associated with all nonparametric methods. Nonparametric Approach: Implicitly smooth ! !=. !+ ] N >1 1 2 is more general than 1. Stoica and R. Lecture notes to accompany Introduction to Spectral Analysis by P. Parametric Approach: Parameterize finite-dimensional model. 1997 Slide L7–2 . Prentice Hall. Moses. but 2 requires to ensure that the number of estimated values = = ) is < N .Basic Ideas Two main PSD estimation approaches: 1. (!) by a 2.

Filter Bank Approach to Spectral Estimation H 6 (!) 1 !c -! Filtered Signal y (t) . Slide L7–3 Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 . 2 ! + 2 ] Note that assumptions (a) and (b). Z 6 ~ (! ) @ -! (b) Z !c ^FB (!) ' 1 2 (a) . jH ( )j2 ( )d = ' 21 !+ !. as well as (b) and (c). Moses. !c @Q Q -! .with varying ! Power Calculation Bandpass Filter c and xed bandwidth - Power in the band - Division by lter bandwidth ^(!c ) - 6 y y (! ) . Prentice Hall. Stoica and R. are conflicting. c ( )d = (') (!) (a) consistent power calculation (b) Ideal passband filter with bandwidth (c) ( ) constant on 2 ! .

i!k = 1 eiN (~. k) 2 ! .t) t=1 X X = N where ( X 1 k=0 hk y(N . 1 H (!) = hk e ! N ei(~. 1 ~ hk = N otherwise 0 1 center frequency of H 3dB bandwidth of H (!) = ! ~ (!) ' 1=N Slide L7–4 Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R.!) . 1997 .i!t 2 ~ ^p(~) = ! 4 N t=1 2 N 1 ~ = N y(t)ei!(N .!) . Prentice Hall. Moses.Filter Bank Interpretation of the Periodogram 1 N y(t)e. 1 k=0 X 1 ei!k k = 0 : : : N .

con't jH (!)j as a function of (~ .Filter Bank Interpretation of the Periodogram. ! 0 −5 −10 −15 dB −20 −25 −30 −35 −40 −3 −2 −1 0 1 ANGULAR FREQUENCY 2 3 Conclusion: The periodogram p ! is a filter bank PSD estimator with bandpass filter as given above. !). and: narrow filter passband. Prentice Hall. Stoica and R. power calculation from only 1 sample of filter output. Moses. for N = 50. 1997 Slide L7–5 . ^( ) Lecture notes to accompany Introduction to Spectral Analysis by P.

Each filter covers a small band centered on ! . Use several bandpass filters on the whole sample. Prentice Hall. Both approaches compromise bias for variance. Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L7–6 . and in fact are quite related to each other: splitting the data sample can be interpreted as a special form of windowing or filtering.Possible Improvements to the Filter Bank Approach 1. This approach leads to Bartlett and Welch methods. This “multiwindow approach” is similar to the Daniell method. Stoica and R. Split the available sample. ~ provides several samples for power calculation. Moses. more data points for the power calculation stage. and bandpass filter each subsample. 2.

Stoica and R.Capon Method Idea: Data-dependent bandpass filter design. = h0 h1 : : : hm] . Moses. 1997 Slide L7–7 . y(t . Prentice Hall. m) h y(t) ~ E jyF (t)j2 = h Rh R = E fy(t)~ (t)g ~ y {z } 7 5 {z } n o y(t) 3 H (! ) = where a X m (! ) = 1 k=0 e.im! ]T hk e. k) 2 6 4 | . yF (t) = X m k=0 | hk y(t .i!k = h a(!) Lecture notes to accompany Introduction to Spectral Analysis by P.i! : : : e.

Moses. n The Bandwidth 1 ' m+1 = (filter 1 length) Conclusion Estimate PSD as: +1 ^(!) = m^.1a=a R.Capon Method. Stoica and R.1a o The power at the filter output is: E jyF (t)j2 = h0Rh0 = 1=a (!)R. con't Capon Filter Design Problem: min(h Rh) h Solution: subject to h a(!) = 1 h0 = R. Prentice Hall.1 a (!)R a(!) with N 1 ^ y(t)~ (t) ~ y R = N . 1997 Slide L7–8 .1a(!) which should be the power of y (t) in a passband centered on ! .m t=m+1 X Lecture notes to accompany Introduction to Spectral Analysis by P.

Moses. Stoica and R. Prentice Hall. . but it splits the N -data point sample into (N m) subsequences of length m with maximum overlap.Capon Properties m is the user parameter that controls the compromise between bias and variance: – as m increases. bias decreases and variance increases. Lecture notes to accompany Introduction to Spectral Analysis by P. Capon uses one bandpass filter only. 1997 Slide L7–9 .

Moses.i!k ^ k=. jkj ^BT (!) = r(k)e. Prentice Hall. s)e.1a(!) Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L7–10 .s) = m+1 r t=0 s=0 ^ a (!)Ra(!) R = ^(i .m m + 1 1 m m ^(t .Relation between Capon and Blackman-Tukey Methods Consider ^BT (!) with Bartlett window: m m + 1 .i!(t. Stoica and R. j )] ^ r = m+1 X X X Then we have ^ a (!)Ra(!) ^BT (!) = m+1 m+1 ^C (!) = ^ a (!)R.

C ! generally has worse resolution and bias properties than the AR method. C ! generally has less statistical variability than the AR PSD estimator. Prentice Hall.Relation between Capon and AR Methods Let 2 ^k ^k (!) = ^ 2 jAk(!)j be the kth order AR PSD estimate of y (t). ^( ) ^( ) Lecture notes to accompany Introduction to Spectral Analysis by P. AR Then ^C (!) = 1 1 m 1= ^AR(!) m + 1 k=0 k X Consequences: Due to the average over k. Stoica and R. Moses. 1997 Slide L7–11 . Due to the low-order AR terms in the average.

1997 Slide L8–1 . Prentice Hall. Stoica and R. Moses.Spatial Methods — Part 1 Lecture 8 Lecture notes to accompany Introduction to Spectral Analysis by P.

. mechanical Receiving sensors: Hydrophones.@ . 1997 Slide L8–2 .. @.R . antennas. @.The Spatial Spectral Estimation Problem Source 1 vSource 2 B B B B N B cccc ccc c c c c c v @ . . communications. underwater surveillance Basic Approach: Determine energy distribution over space (thus the name “spatial spectral analysis”) Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. seismology.@ . v Source n @. electromagnetic. @. @. Moses. Stoica and R.. seismometers Applications: Radar. @. . Emitted energy: Acoustic. . Sensor 1 @.. Sensor 2 Sensor m Problem: Detect and locate n radiating sources by using an array of m passive sensors. sonar.

the array is calibrated. Prentice Hall. 1997 Slide L8–3 .) Lecture notes to accompany Introduction to Spectral Analysis by P. The number of sources n is known. Stoica and R. (We do not treat the detection problem) The sensors are linear dynamic elements with known transfer characteristics and known locations (That is. Moses.Simplifying Assumptions Far-field sources in the same plane as the array of sensors Non-dispersive wave propagation Hence: The waves are planar and the only location parameter is direction of arrival (DOA) (or angle of arrival. AOA).

.g. () f kg with hk(t) This is a time-delay estimation problem in the unknown input case. k) + ek (t) ( = convolution operator). thermal noise in sensor electronics. Then the output of sensor k is yk (t) = hk (t) x(t . etc..) hk (t) = ek (t) = Note: t 2 R (continuous-time signals).Array Model — Single Emitter Case x(t) = k= the signal waveform as measured at a reference point (e. Stoica and R. 1997 Slide L8–4 . Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. at the “first” sensor) the delay between the reference point and the kth sensor the impulse response (weighting function) of sensor k “noise” at the kth sensor (e.g. Prentice Hall. Basic Problem: Estimate the time delays known but x t unknown. background noise.

k ) ' (t) cos !ct + '(t) . k ) ' '(t) Time delay is now ' to a phase shift !c k : x(t . !c k + argfHk (!c)g] where Hk function (!) = Ffhk(t)g is the kth sensor's transfer Hence. Prentice Hall. Then: x(t) = (t) cos !ct + '(t)] where (t) '(t) vary “slowly enough” so that (t . !c k + arg Hk (!c)] + ek(t) Slide L8–5 Lecture notes to accompany Introduction to Spectral Analysis by P. k ) ' jHk (!c)j (t)cos !ct + '(t) . Stoica and R. 1997 . k) ' (t) '(t . Moses. the kth sensor output is yk (t) = jHk (!c)j (t) cos !ct + '(t) . !c k ] hk (t) x(t .Narrowband Assumption Assume: The emitted signals are narrowband with known carrier frequency !c .

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. 1997 Slide L8–6 or . iyk(t) sin(!ct) we get the complex representation: (for t 2 Z ) yk (t) = (t) ei'(t) jHk (!c)j ei arg H (! )] e.i!ct to obtain (t)ei (t) lowpass highpass Hence.i!c k + ek (t) where s(t) is the complex envelope of x(t). by low-pass filtering and sampling the signal yk (t)=2 = yk(t)e.i 2!ct+ (t)] g n o | {z } | {z } Lowpass filter 2z(t)e. Moses.Complex Signal Representation The noise-free output has the form: z(t) = (t) cos !ct + (t)] = = (t) ei !ct+ (t)] + e. Prentice Hall.i! + ek (t) | s(t) {z } | Hk (!c) {z k c } c k yk (t) = s(t)Hk (!c) e.i !ct+ (t)] 2 Demodulate z (t) (translate to baseband): 2z(t)e.!ct = (t)f ei (t) + e.i!ct ~ = yk(t) cos(!ct) .

Vector Representation for a Narrowband Source

Let

= m =
a( ) =
2 6 4

the emitter DOA the number of sensors
2 6 4

H1(!c) e;i!c
. .

1

3 7 5

Hm(!c) e;i!c m y1(t) e1(t) . . y(t) = e(t) = . . ym(t) em(t)
3 7 5 2 6 4

3 7 5

Then

y(t) = a( )s(t) + e(t)
NOTE: enters a via both For omnidirectional sensors the on .

()

f kg and fHk(!c)g. fHk(!c)g do not depend

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L8–7

Analog Processing Block Diagram

Analog processing for each receiving array element
SENSOR DEMODULATION

A/D CONVERSION

-

-

Internal Noise External Noise XXXXX ? XX ll z l - Cabling and , Filtering : ,, Incoming Transducer Signal x(t ; k )

Re yk (t)] Lowpass Re yk (t)] - !! ~ Filter A/D cos(!ct) 6 y k (t) Oscillator -sin(!ct) ? ~ - Im yk (t)] Lowpass Im yk (t)] - !! Filter A/D

-

Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997

Slide L8–8

Multiple Emitter Case Given n emitters with received signals: DOAs:

fsk(t)gn=1 k

k

Linear sensors

) y(t) = a( 1)s1(t) +

+ a( n)sn(t) + e(t)

Let

A = a( 1) : : : a( n)] (m n) s(t) = s1(t) : : : sn(t)]T (n 1)
Then, the array equation is:

y(t) = As(t) + e(t)
Use the planar wave assumption to find the dependence of k on .
Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses, Prentice Hall, 1997 Slide L8–9

Uniform Linear Arrays Source h JJ J J J J JJ J ^ J J JJ JJ+ JJ JJ 2 ] J d sin J ^ JJ J JJ 3 J J v 4 1 v ]? v d - v ppp m vSensor ULA Geometry Sensor #1 = time delay reference Time Delay for sensor k: where c = wave propagation speed d sin k = (k . 1997 Slide L8–10 . 1) c Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Moses. Prentice Hall.

!s = spatial frequency The function !s 7! a( ) is one-to-one for j!sj $ dj sin j 1 d =2 As =2 d =2 is a spatial Shannon sampling theorem.1)!s ]T By direct analogy with the vector a(! ) made from uniform samples of a sinusoidal time series. Prentice Hall. Stoica and R. Moses. Slide L8–11 d = spatial sampling period Lecture notes to accompany Introduction to Spectral Analysis by P.Spatial Frequency Let: sin 4 !s = !c d sin = 2 dc=f = 2 d sin c c = c=fc = signal wavelength a( ) = 1 e. 1997 .i(m.i!s : : : e.

Spatial Methods — Part 2 Lecture 9 Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Prentice Hall. 1997 Slide L9–1 . Moses.

Stoica and R. Moses. Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. 1997 Slide L9–2 .Spatial Filtering Spatial filtering useful for DOA discrimination (similar to frequency discrimination of time-series filtering) Nonparametric DOA estimation There is a strong analogy between temporal filtering and spatial filtering.

1 X (t) = ei!t then yF (t) = h a(!)] u(t) | {z } filter transfer function a(!) = 1 e. Moses.Analogy between Temporal and Spatial Filtering Temporal FIR Filter: yF (t) = hk u(t . m + 1)]T If u m. Lecture notes to accompany Introduction to Spectral Analysis by P. k) = h y(t) k=0 h = ho : : : hm.i! : : : e. 1997 Slide L9–3 .i(m.1)! ]T We can select h to enhance or attenuate signals with different frequencies ! . Stoica and R. Prentice Hall.1] y(t) = u(t) : : : u(t .

i!c m ]T yF (t) = h a( )] s(t) | {z } The corresponding filter output is filter transfer function We can select h to enhance or attenuate signals coming from different DOAs. Spatial FIR Filter output: yF (t) = X m k=1 hk yk (t) = h y(t) Narrowband Wavefront: The array's (noise-free) response to a narrowband ( sinusoidal) wavefront with complex envelope s t is: () y(t) = a( )s(t) a( ) = 1 e.Analogy between Temporal and Spatial Filtering Spatial Filter: fyk(t)gm=1 = k the “spatial samples” obtained with a sensor array. 1997 Slide L9–4 .i!c 2 : : : e. Lecture notes to accompany Introduction to Spectral Analysis by P. Moses. Stoica and R. Prentice Hall.

? B -P C B C B B B . C C B C B C B C B C B A @ . s qq h - ? 1 ? - m.1 ? - 1 .i! C @ h a(!)]u(t) R @ C . Stoica and R.i! m | e {z a( ) aa m! point ( ) qq ( ) 1 C C C C C Cs(t) C C C C C A } - ? @ ? @ @ h - 1 @ R @ - P h a( )]s(t) - hm.. Cu(t) C B . ! hm. ? qq 1 ! (Spatial sampling) (b) Spatial filter Lecture notes to accompany Introduction to Spectral Analysis by P. Moses.B .i! Be 2 B !! B B B B . B .i m. B B @ ..Analogy between Temporal and Spatial Filtering u(t) = ei!t z 1 s q. | e {z } ? a(!) 1 ? 0 qq ( 1) 1 (Temporal sampling) (a) Temporal filter narrowband source with DOA= Z z Z Z Z ~ aa 1reference !! h 0 0 1 B aa 2.@ 0 B 1 C h @@ C B B e.1 q . 1997 Slide L9–5 . Prentice Hall.

Prentice Hall. h a 0 . Moses. Stoica and R. con't Example: The response magnitude h a of a spatial filter (or beamformer) for a 10-element ULA. Here. where 0 j ( )j = ( ) = 25 0 30 −30 Th et a (d eg ) 60 −60 90 0 2 4 6 Magnitude 8 10 −90 Lecture notes to accompany Introduction to Spectral Analysis by P.Spatial Filtering. 1997 Slide L9–6 .

Prentice Hall. Moses.Spatial Filtering Uses Spatial Filters can be used To pass the signal of interest only. hence filtering out interferences located outside the filter's beam (but possibly having the same temporal characteristics as the signal). 1997 Slide L9–7 . Stoica and R. To locate an emitter in the field of view. Lecture notes to accompany Introduction to Spectral Analysis by P. by sweeping the filter through the DOA range of interest (“goniometer”).

and evaluate the powers of the filtered signals: E jyF (t)j2 = E jh ( )y(t)j2 = h ( )Rh( ) with R = E fy (t)y (t)g. 1997 . Moses.Nonparametric Spatial Methods A Filter Bank Approach to DOA estimation. Stoica and R. Basic Ideas Design a filter h ( ) such that for each – It passes undistorted the signal with DOA = – It attenuates all DOAs 6= Sweep the filter through the DOA range of interest. n o n o The (dominant) peaks of h DOAs of the sources. ( )Rh( ) give the Slide L9–8 Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall.

Stoica and R. y t can be considered as impinging on the array from all DOAs.Beamforming Method Assume the array output is spatially white: Then: R = E fy(t)y (t)g = I E jyF (t)j2 = h h n o Hence: In direct analogy with the temporally white assumption for filter bank methods. Moses. Prentice Hall. () Filter Design: min (h h) h Solution: subject to h a( ) = 1 h = a( )=a ( )a( ) = a( )=m E jyF (t)j2 = a ( )Ra( )=m2 n o Lecture notes to accompany Introduction to Spectral Analysis by P. 1997 Slide L9–9 .

Implementation of Beamforming 1 N y(t)y (t) ^ R=N t=1 X The beamforming DOA estimates are: f^kg = the locations of the n largest peaks of a Ra . Prentice Hall. Moses. 1 = 1. 1997 . Resolution Threshold: inf j k . ( )^ ( ) This is the direct spatial analog of the Blackman-Tukey periodogram. Stoica and R. pj > = wavelength array length array beamwidth Inconsistency problem: Beamforming DOA estimates are consistent if n inconsistent if n > . but Slide L9–10 Lecture notes to accompany Introduction to Spectral Analysis by P.

Stoica and R. 1997 Slide L9–11 .1a( ) E jyF (t)j2 = 1=a ( )R. Moses. Both Beamforming and Capon are nonparametric approaches. Prentice Hall. They do not make assumptions on the covariance properties of the data (and hence do not depend on them).1a( ) n o Implementation: f^kg = the locations of the n largest peaks of ^ 1=a ( )R.1a( ): Performance: Slightly superior to Beamforming. Lecture notes to accompany Introduction to Spectral Analysis by P.Capon Method Filter design: min(h Rh) subject to h a( ) = 1 h Solution: h = R.1a( )=a ( )R.

Moses.Parametric Methods Assumptions: The array is described by the equation: y(t) = As(t) + e(t) The noise is spatially white and has the same power in all sensors: E fe(t)e (t)g = 2I The signal covariance matrix P = E fs(t)s (t)g is nonsingular. Stoica and R. Lecture notes to accompany Introduction to Spectral Analysis by P. MIN-NORM and ESPRIT methods of frequency estimation can be used. MUSIC. for DOA estimation. Then: R = E fy(t)y (t)g = APA + 2I Thus: The NLS. YW. Prentice Hall. almost without modification. 1997 Slide L9–12 .

Prentice Hall. Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R.1A ]y(t)k2 f ( ^) = N s t=1 1 N y (t) I .1A ]y(t) = N t=1 ^ = trf I . As(t)k2 min(t)g N f k g fs t=1 X | Minimizing f over s gives f ( s) {z } ^(t) = (A A).Nonlinear Least Squares Method 1 N ky(t) . Moses.1A ]Rg ^ Thus.1A y(t) t = 1 : : : N s X Then 1 N k I . this is precisely the form of the NLS method of frequency estimation.1A ]Rg X f kg For N . A(A A). f ^k g = arg max trf A(A A). A(A A). A(A A). 1997 Slide L9–13 =1 .

Nonlinear Least Squares Method Properties of NLS: Performance: high Computational complexity: high Main drawback: need for multidimensional search. Lecture notes to accompany Introduction to Spectral Analysis by P. Prentice Hall. Moses. Stoica and R. 1997 Slide L9–14 .

1997 Slide L9–15 .) = n. Moses.= y (M L) Also assume: M > n L > n ( ) m = M + L > 2n) rank ~ (A) = rank(A) = n " # " # (. and the SVD of . Prentice Hall.n ~ ~ Properties: A V2 = 0 V1 2 R(A) Then: rank |{z} |{z} Lecture notes to accompany Introduction to Spectral Analysis by P. = U1 U2 ] 0 n n 0 V1 g n .Yule-Walker Method e y(t) = y(t) = A s(t) + ~(t) ~ y(t) ~ e(t) A " # " # " # Assume: Then: E fe(t)~ (t)g = 0 e 4 E fy(t)~ (t)g = AP A ~ . Stoica and R.n 0 V2 g L n M . is .

YW-MUSIC DOA Estimator f^kg = where the n largest peaks of 1=~ ( )V2V2 ~( ) a ^^a ~( ). 1997 Slide L9–16 . using 1 N y(t)~ (t) ^ . Moses. (L 1). is the “array transfer vector” for y(t) a ~ at DOA ^ V2 is defined similarly to V2.=N y t=1 X Properties: Computational complexity: medium Performance: satisfactory if m 2n Main advantages: – weak assumption on e t – the subarray A need not be calibrated f ( )g Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. Prentice Hall.

Stoica and R.MUSIC and Min-Norm Methods Both MUSIC and Min-Norm methods for frequency estimation apply with only minor modifications to the DOA estimation problem. that is. 1997 Slide L9–17 . Lecture notes to accompany Introduction to Spectral Analysis by P. Spectral forms of MUSIC and Min-Norm can be used for arbitrary arrays Root forms can be used only with ULAs MUSIC and Min-Norm break down if the source signals are coherent. if rank (P ) = rank(E fs(t)s (t)g) < n Modifications that apply in the coherent case exist. Moses. Prentice Hall.

D= 0 e. 1997 Slide L9–18 ( ) = d sin( )=c . Stoica and R. Lecture notes to accompany Introduction to Spectral Analysis by P.i!c ( 1) 0 . Let m = # sensors in each subarray A1 = Im 0]A (transfer matrix of subarray 1) A2 = 0 Im]A (transfer matrix of subarray 2) Then A2 = A1D. Moses. where e. Prentice Hall...i!c ( n) ( ) = the time delay from subarray 1 to 2 6 4 3 7 5 subarray 2 for a signal with DOA = : where d is the subarray separation and is measured from the perpendicular to the subarray displacement vector.ESPRIT Method Assumption: The array is made from two identical subarrays separated by a known displacement vector.

so m m and .1 0]A = .1 . con't ESPRIT Scenario source θ known subarray 1 displac ement v ector subarray 2 Properties: Requires special array geometry Computationally efficient No risk of spurious DOA estimates Does not require array calibration Note: For a ULA. Moses.1 A1 = Im. the two subarrays are often the first m and last m array elements.ESPRIT Method.1 A2 = 0 Im.1]A Lecture notes to accompany Introduction to Spectral Analysis by P. Stoica and R. 1997 Slide L9–19 . Prentice Hall.

Master your semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master your semester with Scribd & The New York Times

Cancel anytime.