2 adaptive mixtures density estimation

... better than the Kernel density estimation method 20 0 Events 2, 000 Events 20 ,000 Events Kernel density estimate 7.0338e − 2. 8847e − 1.5 825 e − Modified TV MPLE 3.0796e − 2. 6594e − 8.9353e − 5.4658e ... pixels Color scale for these images 3. 622 7e − (b) 40 Events (c) Kernel density estimation (d) TV MPLE method 3. 622 7e − Color scale for these images (a) True density (e) Our modified TV MPLE method ... pp 25 3 26 5, 20 09 [4] B W Silverman, “Kernel density estimation using the fast fourier transform,” Applied Statistics, Royal Statistical Society, vol 31, pp 93–97, 19 82 [5] B W Silverman, Density...

Ngày tải lên: 21/06/2014, 16:20

... 20 2. 4.1 Estimation of Signal Subspace 22 2. 4 .2 Estimation of Noise Subspace 28 Data Generation and Performance Measures 38 2. 5.1 Data ... algorithm [ 62] for signal subspace estimation 24 2. 5 Oja’s algorithm [78] for signal subspace estimation 25 2. 6 PAST [109] for signal subspace estimation 26 2. 7 Yang ... determine y2 , we first get z2 by subtracting the component of x2 along y1 as H z2 = x2 − y1 x2 y1 (2. 3) where superscript ‘H’ denotes Hermitian transpose Then, y2 is obtained by normalizing z2 as y2 =...

Ngày tải lên: 14/09/2015, 08:44

... respectively, we get  A    k0 v  2 B    k1 v  k2 2 v C   2 v  2  v   2 v (59) where k0  3, k1  12, k2  15 for real-valued cases k0  2, k1  8, k2  for complex-valued cases ...  j  x y y   x    p /2  p p x  y2 2x  j x2  y 4 p p 2 p  z  x  jy   z p  z 2  Likewise, we can obtain operators [21 ]   p /2 p /2  (C.1) 2y p p  Here, we use the Brandword’s ...    u  2 v  k3 v  k4 2 v k0 v u   u   k1 v   k2 2 v  ( 62) where k0  3, k1  12, k2  45, k3  6, k4  15 for real-valued cases, k0  2, k1  8, k2  18, k3  4,...

Ngày tải lên: 19/06/2014, 12:20

... ASME, Vol 124 , 20 02, pp 10–18 Gan, W S & Kuo, S M (20 02) An integrated audio and active noise control headsets IEEE Transactions Consumer Electronics, Vol 48, No 2, May 20 02, pp 24 2 24 7 Guo, J ... 24 04 24 12 Sun, X & Kuo, S M & and Meng, G (20 06) Adaptive algorithm for active control of impulsive noise Journal of Sound and Vibration, Vol 29 1, No 1 2 2006, pp 516– 522 Sun, X & Kuo, S M (20 07) ... Applications of Adaptive Adaptive Advancements Recent Advancements in Active Noise Control Magnitude (dB) 20 20 P(z) S(z) −40 20 0 400 600 800 1000 120 0 1400 1600 1800 20 00 1400 1600 1800 20 00 Frequency...

Ngày tải lên: 19/06/2014, 19:20

... standard IEC 61400 -21 (20 08), along with harmonic measurement standard IEC 61000-4-7 (20 08), provides Technical and Regulatory Exigencies for Grid Connection of Wind Generation 25 the requirements ... requirements on the extent to which the system voltage can be allowed to deviate 22 Wind Farm – Technical Regulations, Potential Estimation and Siting Assessment from its nominal values (±10% for low voltage ... possibly be followed by a complete loss of power (blackout) (Jauch et al., 20 04; Chen & Blaabjerg, 20 09; Tsili & Papathanassiou, 20 09) Many faults in the power system are cleared by relay protections...

Ngày tải lên: 20/06/2014, 05:20

... in ) , Xp (l, in ) 122 (= NSD+NSP+NSG) F: Subcarrier frequency spacing 4. 125 MHz(= 528 MHz/ 128 ) TFFT: IFFT/FFT period 24 2.42ns(= 1/F) TCP: Cyclic prefix duration 60.61ns(= 32/ 528 MHz) TGI: Guard ... Error Rate 10 10 10 10 RRWF (p =2) RRWF (p=4) RRWF (p=6) RRWF (p=8) FIC RRWF Wiener filter 10 10 5 10 15 SNR (dB) 20 25 10 30 10 15 SNR (dB) 20 25 30 10 15 SNR (dB) 20 25 30 10 10 10 10 10 Bit Error ... environments IEEE Trans Microw Theory Tech 52( 9), 21 2 321 38 (20 04) doi:10.1109/ TMTT .20 04.834184 JJ van de Beek, O Edfors, M Sandell, SK Wilson, PO Bửrjesson, On channel estimation in OFDM systems, in Proceedings...

Ngày tải lên: 20/06/2014, 22:20

... in ) , Xp (l, in ) 122 (= NSD+NSP+NSG) F: Subcarrier frequency spacing 4. 125 MHz(= 528 MHz/ 128 ) TFFT: IFFT/FFT period 24 2.42ns(= 1/F) TCP: Cyclic prefix duration 60.61ns(= 32/ 528 MHz) TGI: Guard ... Error Rate 10 10 10 10 RRWF (p =2) RRWF (p=4) RRWF (p=6) RRWF (p=8) FIC RRWF Wiener filter 10 10 5 10 15 SNR (dB) 20 25 10 30 10 15 SNR (dB) 20 25 30 10 15 SNR (dB) 20 25 30 10 10 10 10 10 Bit Error ... environments IEEE Trans Microw Theory Tech 52( 9), 21 2 321 38 (20 04) doi:10.1109/ TMTT .20 04.834184 JJ van de Beek, O Edfors, M Sandell, SK Wilson, PO Bửrjesson, On channel estimation in OFDM systems, in Proceedings...

Ngày tải lên: 21/06/2014, 00:20

... then obtain I x2 ≈ I y2 ≈ Ω+ r Ω+ r I y ≈ y2d − Ω+ r (1 /2) κr x2 d − yd − L = I(1) ≈ Ω+ r (1 /2) κr (1 /2) κr r2d y = π r − κr , y2d y = π κ3 r − r , 12 (1 /2) κr d 2 ⎢ Σ(C) ≈ ⎣ r − κr 2 r , ⎡ − ⎢ ⎣ ... 10 Estimation error Estimation error 0.05 0 −0.05 −0.1 −0.15 −0 .2 −5 −5 x −0 .25 −5 x Adaptive radius r=4 r = 0.1 Adaptive radius r=4 r = 0.1 (a) (b) Estimation error 0 .2 −0 .2 −0.4 −0.6 −5 x Adaptive ... paraboloid fitting [22 24 ] and quadratic fitting [1, 25 27 ] Apart from these fitting techniques, other methods have been proposed, such as higher-order fitting [5, 28 ] and circular fitting [29 , 30].Cazals...

Ngày tải lên: 21/06/2014, 08:20

... Pd 0.9 cos2 φ = 0.4 0.3 0.4 cos2 φ = 0.3 cos2 φ = 0.7 0 .2 0.5 0 .2 0.1 0.1 10 15 20 25 cos2 φ = 0.7 10 15 25 SNR (dB) SNR (dB) KRAO: ρ = 0.9 KRAO: ρ = 1 .2 KRAO: ρ = 1.4 KRAO: ρ = 0 .2 KRAO: ρ = ... “Performance of a class of adaptive detection algorithms in nonhomogeneous environments,” IEEE EURASIP Journal on Advances in Signal Processing [20 ] [21 ] [22 ] [23 ] [24 ] [25 ] [26 ] [27 ] [28 ] Transactions ... factorization to M−1 /2 H: M−1 /2 H = H0 RH ( 42) EURASIP Journal on Advances in Signal Processing 0.9 0.8 0.7 0.7 0.6 0.6 0.5 0.5 Pd 0.8 Pd 0.9 0.4 0.4 0.3 0.3 0 .2 0 .2 0.1 cos2 φ = 0.1 10 15 20 cos2 = 0.8 10...

Ngày tải lên: 21/06/2014, 08:20

... −0 .2 t 0 .2 −0.4 0.4 −0 .2 (a) 0.4 t 0 .2 0.4 t 0 .2 0.4 t 0 .2 0.4 100 50 50 Ω(t) Ω(t) 0 .2 (b) 100 0 −0.4 −0 .2 t 0 .2 −0.4 0.4 −0 .2 (c) (d) 100 50 50 Ω(t) 100 Ω(t) t 0 −0.4 −0 .2 t 0 .2 −0.4 0.4 −0 .2 ... −0 .2 t 0 .2 −0.4 0.4 −0 .2 (a) 0.4 t 0 .2 0.4 t 0 .2 0.4 100 Ω(t) Ω(t) 0 .2 (b) 100 0 −100 −100 −0.4 −0 .2 t 0 .2 −0.4 0.4 −0 .2 (c) (d) 100 100 Ω(t) Ω(t) t 0 −100 −100 −0.4 −0 .2 t 0 .2 −0.4 0.4 −0 .2 (e) ... A4 n1 n2 ∗ wh (n1 T)wh (n2 T) − j (n1 T )2 − (n2 T )2 ⎛ × exp 2 j ∞ (n T)2k − (n2 T)2k ⎠ (t) (2k)! (A.7) For simplicity, all higher-order derivatives, except the fourth |φ(2k) (t)| for k > 2: order...

Ngày tải lên: 21/06/2014, 19:20

... bc ( 2 ) ⊗ a1 2 , , bc (θK ) ⊗ a1 φK · bc φ1 ⊗ a1 (θ1 ), bc 2 ⊗ a1 ( 2 ), , bc φK ⊗ a1 (θK ) H d1 θ1 , φ1 , d1 2 , 2 , , d1 θK , φK Λ · d2 (θ1 , φ1 ), d2 ( 2 , 2 ), , d2 (θK ... 101 RMSE (deg) 101 10−1 10 2 −5 10−1 10 SNR (dB) 15 20 25 p =2 p=3 p=4 10 2 −5 10 SNR (dB) 15 20 25 p =2 p=3 p=4 (a) (b) Figure 2: RMSE of 2- D DOA estimation from (55◦ , 100◦ ) for (a) elevation and ... Processing, vol 40, no 7, pp 1 828 –18 32, 19 92 [11] V S Kedia and B Chandna, “A new algorithm for 2- D DOA estimation, ” Signal Processing, vol 60, no 3, pp 325 –3 32, 1997 [ 12] A L Swindlehurst and T Kailath,...

Ngày tải lên: 21/06/2014, 19:20

... polygon V Adaptive Control 32 2.3 .2 Vector case: d > In case of d > 1, since θ and φk are vectors, we cannot directly obtain explicit solution of inequality (2. 8) Notice that Eq (2. 8) can be ... model (2. 2) involves various uncertainties, we rewrite it into the following form of linear regression (2. 3) by letting vk = f (φk , k ) + ε k From the a priori knowledge for model (2. 2), we ... the adaptive estimation and control problem for the semi-parametric model Finally, we give some concluding remarks in Section Semi-parametric Adaptive Estimation: Principles and Examples 2. 1...

Ngày tải lên: 21/06/2014, 19:20

... 1.4 Ri 1.4 1 .2 Ri 1 .2 1 0.8 Rs, Rr, Lm Rs, Rr, Lm 0.8 0.6 0.6 0.4 0.4 0 .2 0 .2 0 0.5 1.5 Time 2. 5 (a) model Figure Estimated parameters for DFIM 3.5 -0 .2 0. 02 0.04 0.06 0.08 0.1 0. 12 Time ... 1995) 22 Frontiers in Adaptive Control Recursive Least-Squares (RLS) In adaptive controller such as self-tuning regulator the estimated parameters are needed online The least-squares estimation ... R (39) On-line Parameters Estimation with Application to Electrical Drives ,a , , a , a , a , , a , , a ,a , , a ]T 1,0 2, 0 R ,0 1,1 2, 1 R ,1 1, n 2, n R, n θ = [a 29 (40) We can write (36) as...

Ngày tải lên: 21/06/2014, 19:20

... −∞ (Xi /hN ) −(t2 /2) e dt 2 √ h2 t N f (x) + O h3 t N X +∞ fX (x) −∞ (B .2) √ e−(t /2) dt 2 h2 N f (x) X = −∞ fX (x)dx + lim E pe,N N → +∞ −∞ fX (x)dx N x − Xi K NhN i=1 hN (C .2) N x − Xi E ... then R K(w)dw = 1) +∞ =E =E −∞ X1 − (x + y /2) x−y √ K √ dx d y hN / 2hN −∞ E A2 = E K Using (D .2) , (D.5), and (D. 12) , we obtain Var pe,N = E A2 − (E[A]) = h2 h3 N fX (0) − hN pe + N fX (0) hN p ... pp 27 66 27 78, 20 08 [3] M Rosenblatt, “Remarks on some nonparametric estimates of a density function,” The Annals of Mathematical Statistics, vol 27 , no 3, pp 8 32 837, 1956 [4] E Parzen, “On estimation...

Ngày tải lên: 21/06/2014, 23:20

... λH (25 ) |en | > λH , |en | ≤ λH |en | > λH , (26 ) Guang Deng ρ (x) ρ(x) 0.5 ρ (x) 1 .2 0.4 0.5 0.8 0 .2 0.3 −0.5 −1 −0.5 0.5 −1 −1 −0.5 0.5 −0 .2 −1 L2 Fair Huber L2 Fair Huber (a) −0.5 0.5 L2 Fair ... parameter estimation, ” IEEE Signal Processing Letters, vol 12, no 1, pp 21 24 , 20 05 [10] S Haykin, Adaptive Filter Theory, Prentice-Hall, Englewood Cliffs, NJ, USA, 4th edition, 20 02 [11] W A ... zero mean Gaussian distributions which Guang Deng −10 20 15 −15 10 20 25 −5 −10 −30 −15 20 −35 500 1000 1500 20 00 25 00 3000 3500 4000 Figure 2: Noise signal used in simulations −40 −45 is simulated...

Ngày tải lên: 22/06/2014, 01:20

... Mobile Carphone Foreman Table tennis Bream Bus Maximum FSBM 26 5 26 5 26 5 26 5 26 5 26 5 26 5 4SS 19 21 21 19 19 24 24 DS 15 18 18 15 15 21 21 MVFAST 13 13 8 18 18 FAME 9 8 Table 3: Required operating ... memories Such T Dias et al 10 IR ‘0’ PC MB MEM MUX 16 R0 R4 R1 R5 R2 R6 R3 R7 ··· ··· ··· R20 R24 R28 R21 R25 R29 R 22 R26 R30 R23 R27 R31 16 16 ASR 16 MUX ··· 16 ··· 10 SADU 16 MUX MUX MUX ‘1’ ... ML310 development board Unit AMEP core Interface Slices LUTs BRAMs F (MHz) 20 52 14% 22 89 8% 20 8 67.18 20 7 1% 3 82 1% 156 .20 the MVFAST and FAME algorithms (744 × 16 bit) However, since most of the...

Ngày tải lên: 22/06/2014, 19:20

... SDM −100 20 0 20 0 −300 Power spectral density (dB) −100 Power spectral density (dB) SDM 2- bit ASDM −100 20 0 −300 Proposed 2- bit ASDM −300 2- bit ASDM −100 −150 20 0 25 0 Proposed 2- bit ASDM ... SDM −100 20 0 20 0 −300 Power spectral density (dB) −100 Power spectral density (dB) SDM 2- bit ASDM −100 20 0 −300 Proposed 2- bit ASDM −300 2- bit ASDM −100 −150 20 0 25 0 Proposed 2- bit ASDM ... −150 20 0 20 0 −300 100 101 1 02 103 Frequency (Hz) 104 25 0 105 1000 20 00 3000 4000 5000 6000 7000 8000 9000 Frequency (Hz) (a) (b) ×105 SDM Power 2 2-bit ASDM −1 2 1.5 0.5 −0.5 Proposed 2- bit...

Ngày tải lên: 22/06/2014, 19:20

... (VBS) 8.91 2. 51 12. 63 16.18 10 .26 1.51 16.60 12. 31 10 .23 1.18 12. 92 15.81 10 .22 1.17 12. 91 15. 82 9.98 -1 .21 13.18 15.49 28 .16 2. 03 16 .24 12. 58 28 .13 1.95 12. 92 15. 82 28.11 1.85 9.45 21 .61 27 .84 0.89 ... 19. 62 285.01 2. 89 21 .58 9.36 29 9.69 8.19 18.90 10.69 28 7.03 3. 62 17. 02 11.87 29 2.6 5.63 14.67 13.77 28 4.93 2. 86 15.53 13.01 25 0 .23 −9.67 16.60 12. 17 416.43 8 .20 23 .09 8.75 4 12. 54 7.19 20 .56 9. 82 ... Speedup 40.67 0.00 20 2.05 1.00 44.01 8 .21 23 .18 8. 72 42. 21 3.78 22 .24 9.09 42. 92 5.54 19 .22 10.51 42. 38 4 .21 17.51 11.54 43.3 6.47 15. 12 13.36 42. 09 3.49 14.73 13. 72 37.64 −7.44 15.91 12. 70 Carphone...

Ngày tải lên: 22/06/2014, 19:20

... 7.89 ¢ 10 2 σpre = 3.89 ¢ 10 2 σpre = σpre = 6.89 ¢ 10 2 σpre = 5.89 ¢ 10 2 σpre = 4.89 ¢ 10 2 2.89 ¢ 10 2 > σ n 2 σpre = 1.89 ¢ 10 2 < σn f (L) e(L) L 10 12 14 16 e2 (L) Lσn e1 (L) Figure 10: ... − hL k,i 2 + σn L (29 ) Van Duc Nguyen et al The auxiliary function f (L) (dB) The auxiliary function f (L) (dB)  1 2  3  4  5  6  5  10  15 20 25  30 10 12 14 16 L σpre = 7.89 ¢ 10 2 σpre = ... is, (s) 2 σn = σpre + Δσ ( 12) Store the estimated noise variance obtained from the sth EURASIP Journal on Wireless Communications and Networking experiment in a vector: 2( s) 2( 1) 2( 2) 2( s) ˇ...

Ngày tải lên: 22/06/2014, 22:20