This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

Scribd Selects Books

Hand-picked favorites from

our editors

our editors

Scribd Selects Audiobooks

Hand-picked favorites from

our editors

our editors

Scribd Selects Comics

Hand-picked favorites from

our editors

our editors

Scribd Selects Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

P. 1

00757221 Discrete Fractional Fourier Transform Based on Orthogonal|Views: 26|Likes: 0

Publicado porOm Prakash Suthar

Published by: Om Prakash Suthar on Sep 14, 2011

Direitos Autorais:Attribution Non-commercial

See more

See less

https://pt.scribd.com/doc/64913994/00757221-Discrete-Fractional-Fourier-Transform-Based-on-Orthogonal

09/14/2011

text

original

5, MAY 1999

1335

**Discrete Fractional Fourier Transform Based on Orthogonal Projections
**

Soo-Chang Pei, Senior Member, IEEE, Min-Hung Yeh, and Chien-Cheng Tseng, Member, IEEE

Abstract—The continuous fractional Fourier transform (FRFT) performs a spectrum rotation of signal in the time–frequency plane, and it becomes an important tool for time-varying signal analysis. A discrete fractional Fourier transform has been recently developed by Santhanam and McClellan, but its results do not match those of the corresponding continuous fractional Fourier transforms. In this paper, we propose a new discrete fractional Fourier transform (DFRFT). The new DFRFT has DFT Hermite eigenvectors and retains the eigenvalue-eigenfunction relation as a continous FRFT. To obtain DFT Hermite eigenvectors, two orthogonal projection methods are introduced. Thus, the new DFRFT will provide similar transform and rotational properties as those of continuous fractional Fourier transforms. Moreover, the relationship between FRFT and the proposed DFRFT has been established in the same way as the conventional DFT-tocontinuous-Fourier transform. Index Terms— Discrete Fourier transform, discrete fractional Fourier transform, Fourier transform, fractional Fourier transform.

I. INTRODUCTION HE FOURIER transform (FT) is one of the most frequently used tools in signal analysis [1]. A generalization of the Fourier transform—the fractional Fourier transform (FRFT)—has been proposed in [2] and [3] and has become a powerful tool for time-varying signal analysis. In time-varying signal analysis, it is customary to use the time-frequency plane, with two orthogonal time and frequency axes [4]. Because the successive two forward Fourier transform operations will result in the reﬂected version of the original signal, the FT can in the be interpreted as a rotation of signal by the angle time–frequency plane and represented as an orthogonal signal representation for sinusoidal signal. The FRFT performs a rotation of signal in the continuous time–frequency plane to any angle and serves as an orthonormal signal representation for the chirp signal. The fractional Fourier transform is also called rotational Fourier transform or angular Fourier transform in some documents. Besides being a generalization of the FT, the

Manuscript received December 2, 1996; revised February 19, 1998. This work was supported by the National Science Council, R.O.C., under Contract NSC 85-2213-E002-025. The associate editor coordinating the review of this paper and approving it for publication was Dr. Akram Aldroubi. S.-C. Pei is with the Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, R.O.C. (e-mail: pei@cc.ee.ntu.edu.tw). M.-H. Yeh is with the Department of Computer Information Science, Tamsui Oxford University College, Tamsui, Taipei, Taiwan, R.O.C. C.-C. Tseng was with the Department of Electronics Engineering, Hwa Hsia College of Technology and Commerce, Taipei, Taiwan, R.O.C. He is now with the Department of Computer and Communication Engineering, National Kaohsiung First University of Science and Technology, Taipei, Taiwan, R.O.C. Publisher Item Identiﬁer S 1053-587X(99)03244-4.

T

FRFT has been proved to relate to other time-varying signal analysis tools, such as Wigner distribution [4], short-time Fourier transform [4], wavelet transform, and so on. The applications of the FRFT include solving differential equations [2], quantum mechanics [3], optical signal processing [5], timevariant ﬁltering and multiplexing [5]–[8], swept-frequency ﬁlters [9], pattern recognition [10], and time–frequency signal analysis [11]–[13]. Several properties of the FRFT in signal analysis have been summarized in [9]. Many methods for realizing the FRFT have been developed, but most of them are to utilize the optical implementation [14], [15] or numerical integration. Because the FRFT is a potentially useful tool for signal processing, the direct computation of the fractional Fourier transform in digital computers has become an important issue. The ideal discrete fractional Fourier transform (DFRFT) will be a generalization of the discrete Fourier transform (DFT) that obeys the rotation rules as the continuous FRFT and provides similar results as the FRFT. In [16], a method for a numerical integration FRFT has been proposed, but the method does not obey the rotation rules, and the signal cannot be recovered from its inverse transform. In [17], Santhanam and McClellan have developed a discrete FRFT, but their method does not provide the same transforms to match those of the continuous case. In this paper, we present a new discrete fractional Fourier transform (DFRFT). This DFRFT is a generalization of the DFT and will provide similar transforms as those of the continuous case. The relationship between the DFRFT and the FRFT can also be established and discussed in detail. Moreover, the proposed DFRFT has important unitary and rotation properties. This paper is organized as follows. In Section II, the previous development of continuous and discrete fractional Fourier transforms are reviewed. The concept for developing the DFRFT to have similar results as the continuous corresponding case are described in Section III. Two acceptable solutions for the DFRFT are considered and proposed in Section III. Then, the relationships between the FRFT and the DFRFT can be established in Section IV. Finally, conclusions and discussions are made in Section V. II. PRELIMINARY A. Continuous FRFT The Fourier transform of a signal can be interpreted as a angle rotation of the signal in the time–frequency plane. The FRFT is then developed and treated as a rotation of signal

1053–587X/99$10.00 © 1999 IEEE

[6]. 2 shows the results of this DFRFT produced by a discrete rectangular window. Imaginary part: dashed line. with unit variance. In order to avoid ambiguity. the (8) is the conventional DFT of signal . 47. However. The rectangular window used . Thus. The real parts of the FRFT or DFRFT in this paper are plotted by solid lines. Fig. In [19]. NO. The weighting coefﬁcients are obtained from multiplying the phase term and the inner product of the input signal and the corresponding Hermite function. and the rotation output of the signal is also the weighting summation of Hermite functions. [9] csc if if if is not a multiple of (1) is a multiple of is a multiple of (2) where indicates the rotation angle of transformed signal for is the th-order normalized Hermite function the FRFT. B. Fractional Fourier transform of a rectangular window. The fractional power of the DFT kernel matrix [20] is calculated by (6) (6) where (7) Applying the above deﬁned kernel to the signal is computed as DFRFT of the signal . [3]. is the eigenfunction of the FRFT. and the English superscripts are used to denote the fractional power values of the DFT kernel matrices in this paper. x(t) = 1 when 02 t 2. and . to any angles in the time–frequency plane [9]. Using the kernel of the by angle is computed as FRFT. and x(t) = 0 otherwise. and the imaginary parts of the FRFT or DFRFT are indicated by dashed lines. Equation where is the linear (8) indicates that the DFRFT of the signal combination of the four major angular parts: the original signal . The fractional power of the DFT kernel matrix and the rotation angles in the DFRFT essentially mean the same thing. 5. Because where the normalized Hermite function with unitary variance. Fig. 2 are quite different from the results of Fig. it will have a different phase term in the weighting coefﬁcients. The discrete results shown in Fig. . the Greek subscripts (for example.1336 IEEE TRANSACTIONS ON SIGNAL PROCESSING. The conventional methods to compute the DFRFT in [17] and [20] are mainly on numerical computing the fractional power of the DFT kernel matrix. a circular reﬂected version of the signal ( ). 1 shows the FRFT of the rectangular window function [ for . elsewhere] for various angles. its DFT . The equivalence of (1) and (2) has been proved in [3]. 1. The th-order normalized Hermite function with variance is deﬁned as (3) is the th-order Hermite polynomial [18]. the FRFT of the signal (4) (5) Equation (5) indicates that the FRFT can be interpreted as a weighting summation of Hermite functions. MAY 1999 Fig. Real part: solid line. the two results . a rotation operation using Hermite functions has also been proposed. (2) provides an eigendecomposition representation of the FRFT kernel. the middle is not the intermediate state of transform result for and . and a circular reﬂected version of its DFT . The Old DFRFT The kernel matrix of the DFRFT is obtained by computing the fractional power for the DFT kernel matrix. otherwise here is deﬁned as [ ]. Furthermore. . the Hermite functions in [19] with a different scaling to those of the FRFT. 1. The transform kernel of continuous FRFT is deﬁned as [2]. and ) are used to denote the rotation angles in the time–frequency plane. VOL.

[22]. (9) The Fourier transform of (13) can be computed as (15) . a multiplicity function function is used to denote the DFT eigenvalue multiplicity for . . where associated continuous spread variance is the sampling intervals of signal. In [22]. Proposition 1: The eigenvalues of and its multiplicities are shown in Table I. it satisﬁes the commutative property (10) will also be the eigenvectors The eigenvectors of matrix of the DFT kernel matrix . Can the DFT have the eigenvectors with the similar shapes as the Hermite functions? These DFT eigenvectors are known as DFT Hermite eigenvectors in this paper. we get (12) is the th-order Hermite polynomial. is deﬁned. . Here. In [20]. and corresponding to one of the four eigenvalues—1. . a method for computing the DFT eigenvectors has been introduced.. B. Matrix commutes with the DFT kernel . The eigenvectors of the DFT matrix constitute four . we have known that the eigendecomposition of the DFT kernel matrix is not unique. Eigendecomposition of DFRFT The development of our DFRFT is based on the eigendecomposition of the DFT kernel. . . Proof: See [21]. . The continuous approximate form can be written as (13) Sampling by . Old DFRFT of a rectangular window function. . x(n) = 1 when 06 n 6. The parameter is the index for the DFT eigenvalue .: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1337 Fig. . . . . where is the spread variance of the Proof: It is assumed that DFT eigenvectors. . The eigenvalue multiplicities of DFT matrix indicate the ranks of the DFT eigensubspaces [21]. (13) will become (14) . and x (n ) = 0 otherwise. are . Proof: See [20]. but it cannot obtain the real-valued DFT eigenvectors. DFT Hermite Eigenvectors III. and then. This In Table I. and matrix is deﬁned as OF DFRFT The continuous FRFT has a Hermite function with unitary variance as its eigenfunction. . —respectively. 2. from the continuous FRFT in Fig. and many properties of the DFT matrix eigenvalues and eigenvectors have been discussed in [20] and [21]. we only summarize some results for our further development of DFRFT. The corresponding eigenfunction property for the DFT would be like (11) is the eigenvector of DFT corresponding to the where th-order discrete Hermite function.PEI et al. In the discussion in the previous subsection. DEVELOPMENT A. The output is quite different TABLE I MULTIPLICITIES OF THE EIGENVALUES OF DFT KERNEL MATRIX where matrix . Proposition 2: A matrix can be used to compute the real eigenvectors of the DFT matrix . but they correspond to different eigenvalues. 1. . If the Hermite function are sampled in this way. a novel matrix is introduced to compute the real and complete set of DFT eigenvectors very elegantly. 1. Proposition 3: DFT Hermite eigenvectors should have the . . and —and each is major eigensubspaces— .

the better approximation (17) is. . i. . and the other is the numerical error in (20). taking at (21) Equation (14) becomes The proof is completed. and the samplings in Proposition 3 cannot guarantee that the numbers of the sign changes in this sampled th-order Hermite function are also . by replacing the continuous integral with the numerical integral. Thus. but this will not inﬂuence the development of the DFRFT. it is clear that there are two approximation errors in (17). we have . When the value of approaches inﬁnity. 5. In Propositions 4 and 5. the approximation in (17) becomes worse. the larger order is. is obtained by samProposition 4: If the sequence pling the unit variance Hermite function with the sampling . VOL. because larger is . the proof can be easily derived in the same way. for (23) for then it can be shown that the DFT of the . the decay rate the degree of Hermite polynomial is proportional to for of the Hermite function sufﬁciently large . we have (20) (24) for sufﬁciently large . the sampling interval we will prove that even such samplings of Hermite functions can only have approximate Hermite eigenvectors of the DFT. we obtain and letting This expression is valid for any both sides. approximated by can be From Proposition 3. Next. The variable in (15) can be replaced by equal to to have its discrete form. when order becomes large. MAY 1999 The sampling theorem [23] tells us the continuous scope of .1338 IEEE TRANSACTIONS ON SIGNAL PROCESSING.e. From the above proof. we have the approximation expression (19) is wide This approximation is valid because is for large . the is. Combining (19) and (20) very small when . The small aliasing will occur while is closer to . 47. Next. Therefore. Thus. The condition of for the two shapes of (14) and (16) being the same becomes This approximation is also valid because is is very large. Proof: Here. The sequence obtained from (12) is also equal to the sampling of the Hermite function with the unit variance by . This implies that the truncation error in (19) is larger for high-order . For By truncating the integral interval of Fourier transform from to . we only prove the case that odd. is even. However. and our goal is to adjust the variance have the same eigenvector shape as itself. frequency spectrum that DFT can cover is the continuous spectrum spacing (or resolution) of the DFT is . these functions are not bandlimited. One is the truncation error in (19). both errors approach zero. deﬁned in the range Proposition 5: If the sequence is obtained by shifting Hermite Gauss samples in the following way: If is even for (22) for and if is od.. we can ﬁnd that the DFT Hermite eigenvectors should have associate continuous spread variance . Thus. The th-order continuous Hermite function should have zeros [18]. NO. and the decay rate of Gauss function very fast. (16) adjusts the scaling of Hermite funcThe variance value to make the DFT tion. The sampled Hermite functions still can be used to construct the DFRFT kernel because they have the similar shapes and good approximations to the continuous Hermite functions. the slower decay rate the Hermite function will have. Thus. then it can be shown that interval If is even (17) If is odd (18) for sufﬁciently large .

. In addition. the DFT is computed from the eigenvectors of Hermite eigenvector matrix in the same DFT eigensubspace. Because of the above ambiguity. In (32). therefore. Combining . we can compute in its DFT eigensubspace to obtain a the projections of Hermite-like DFT eigenvector (32) mod . This indicates that there exists root ambiguity in deciding the fractional power of eigenvalues.PEI et al. Newly Developed DFRFT DFT (27) (28) is limited in the range where equality rewritten as DFT (29) is limited in the range where (28) and (29). In Propositions 4 and 5. we also prove the case that is even. The idea for developing our DFRFT is to ﬁnd the discrete form for (2). eigenvectors for the However. (26) Substituting (26) into (25) and using Proposition 4. the choices for the DFT eigenvectors to construct the DFRFT kernel are multiple and not unique. there exist two types of ambiguity in deciding the fractional power of the DFT kernel matrix. we know that there are several DFRFT kernel matrices that can obey the rotation properties. for the discrete case. Unfortunately. Using the . • Ambiguity in Deciding the Fractional Powers of Eigenvalues: We know that the square root of unity are 1 and 1 from elementary mathematics. (27) can be The fractional power of matrix can be calculated from its eigendecomposition and the powers of eigenvalues. the proof can also be easily derived. The normalized vectors for the samplings of Hermite functions are deﬁned as (31) . In addition. This assignment rule matches the multiplicities of the eigenvalues of the DFT kernel matrix in Table I. the Because matrix can have complete real orthogonal DFT eigenvectors. we should not expect that a ﬁnite vector can express the oscillation behavior of the very high-order Hermite function very well. the unit variance Hermite functions are in the following sampled with a period of discussions. is odd. Table II shows the eigenvalues assignment rules for the DFRFT. the eigenvectors can be used as bases for individual DFT eigensubspaces. it has been proved that the samplings of Hermite functions can have approximate DFT eigenvectors. The selections of the DFT Hermite eigenvectors are from low to high orders. In order to retain the eigenfunction property in (11). and is the eigenvector of matrix where .: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1339 Proof: Here. will be a DFT Hermite eigenvector. The While is given by DFT of the sequence DFT TABLE II EIGENVALUES ASSIGNMENT RULE OF DFRFT KERNEL MATRIX (25) Using the equality second term on the right side of (25) becomes . It is because the approximation error of the low DFT Hermite eigenvectors are small. we obtain DFT for for (30) The proof is completed. only DFT Hermite eigenvectors can be added. In the case of continuous FRFT. • Ambiguity in Deciding the Eigenvectors of the DFT Kernel Matrix: The DFT eigenvectors constitute four major eigensubspaces. we obtain C. the terms of the Hermite functions are summed up from order zero to inﬁnity.

the eigenvectors obtained from matrix are directly considered to be the discrete Hermite functions. Equation (32) provides a method for ﬁnding DFT Hermite eigenvectors. Calculate the continuous samples of Hermite functions: Compute the eigenvectors of matrix : Using equation (32) to compute Hermite eigenvectors by projections: for to for to end end . This is because the eigenvectors located in different eigensubspaces will be always orthogonal. . VOL. So the DFT Hermite eigenvectors can be orthogonalized for every eigensubspace individually to obtain orthogonal eigenvectors in the whole eigenspace of DFT. It must be mentioned that the last eigenvalue and should be assigned to the is not assigned to to match the eigenvalue multiplicities value of the DFT matrix in Table I. and for even . However. The symbol notation in developing the two algorithms are as follows: continuous Hermite samples vector. and for is even. Therefore. where the eigenvector corresponding to the th continuous Hermite function. orthogonalized Hermite eigenvector.. NO. . . is the bottom of the page. In addition. 5. The Gram–Schmidt Algorithm (GSA): The Gram–Schmidt method [29] is a well-known orthogonalization approach for vectors. The DFT eigenvectors are orthogonalized for each eigensubspace. nonorthogonalized Hermite eigenvector. the convergence for the eigenvectors obtained from matrix are not so fast for the high-order Hermite functions by the LMS error criterion. It is easy to show that the eigenvectors located in different eigensubspaces will be orthogonal.1340 IEEE TRANSACTIONS ON SIGNAL PROCESSING. It is easy to verify that the angle rotation property of the DFRFT can be preserved only while the DFT eigenvectors are orthogonal. whereas is odd. This means that is replaced by in (34). (34a) . . the transform kernel for the rotation angle can be computed as in (34a). MAY 1999 The transform kernel of the DFRFT can be deﬁned as (33) for (odd) (34) for (even) where . Unfortunately. . is the normalized eigenvector corresponding to the th-order Hermite function. . 47. An eigenbased method for computing the DFRFT to have similar continuous results has been proposed by us in [24]–[27]. . shown at .. Therefore. the DFT Hermite eigenvectors obtained from (32) cannot constitute an orthogonal basis for DFT eigenspace. vector orthogonalization is required for the DFT Hermite eigenvectors obtained from (32). The role of matrix in (32) is just as a tool to ﬁnd a complete set of real and orthogonal DFT eigenvectors. . Two methods for vector orthogonalization are proposed in the following of this paper. the eigenvalue–eigenfunction relation is retained in deﬁning the DFRFT. Such a method is called the method in this paper. . the eigenvectors obtained from the matrix are just discrete Mathieu functions [21]. Although the Mathieu functions can converge to a Hermite function [28]. In [25]. for and the eigenvalues of the DFRFT are . . The DFT Hermite eigenvectors in each DFT eigensubspace can be orthogonalized by the Gram–Schmidt method. Example 1— : The rotation angle is equal to . is deﬁned as follows: where For odd .

9 and 10. where . GSA. GSA. Based on the unitary property in the DFRFT and the transform results shown in Figs. The sampled discrete data then becomes . Frobenius norm of (35) samples of Hermite functions are plotted for . 3. 8–10. Norms of error vectors between DFT Hermite eigenvectors and samples of Hermite functions. The error vectors between samples of Hermite functions and DFT Hermite eigenvectors are deﬁned as (37) Example 2: In this example. we use the DFRFT to deal with the rectangular window function shown in Fig. 7. ( the DFRFT can be used for the chirp signal and chirp rate detection. 582] for the two given spaces. The continuous FRFT is an orthonormal signal decomposition for chirp signals [9]. D. otherwise. 3. Therefore. The minimizing calculating the singular value decomposition (SVD) of . ) are plotted Transform results of the impulse signal ( in Figs. and OPA. p. and will can be found by be our solution. 2. for minimize minimize where matrix. 1. 4–6 show the DFRFT of the rectangular window function for various angles for method. is the subject to (36) the . 8–10 for the method. In Fig.PEI et al. we can ﬁnd that the proposed DFRFT provides a similar orthonormal signal decomposition for discrete chirplike signals. . Because the . respectively. The is equal to 4/13. Orthogonal Procrustes Algorithm (OPA): A traditional mathematical problem known as the orthogonal Procrustes Algorithm can be used to ﬁnd the least Frobenius norm [30. . The norm of the error vectors between the DFRFT and samples of the FRFT are also shown in the titles of Figs. and OPA. Example 3: The DFRFT by the GSA for a chirp signal is computed in this example. 11. [ which is the same as the signal used in Fig. and the OPA minimizes the total errors between the samples of Hermite functions and orthogonal DFT Hermite eigenvectors. The chirp signal used here is equal . . and the number of points sampling interval is equal to 73. satisfy Algorithm Calculate the continuous samples of Hermite functions: Compute the eigenvectors of : Using equation (32) to compute Hermite eigenvectors by projections: for to Compute the SVD of . Figs. the solution will also . A more detailed theory and algorithm can be found in [7]. The corresponding samples of the continuous FRFT for the impulse signal are plotted in Fig. We can formulate our problem as the OPA to ﬁnd the least Frobenius norm between the samples of Hermite functions and orthogonal DFT Hermite eigenvectors. respectively. In Fig. the norms of error vectors between the computed DFT Hermite eigenvectors and . the last half of the indices in the DFRFT must also be treated as the negative end The GSA minimizes the errors between the samples of Hermite functions and orthogonal DFT Hermite eigenvectors from low to high orders. Properties of DFRFT The properties of the DFRFT are shown in Table III. E. ]. it can be to found that the transform results change from a chirp signal ) to an impulse-like function ( ). Implementation of the New DFRFT As in the case of DFT frequency domain.: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1341 Fig.

DFRFT by GSA method of a rectangular window function. and x (n ) = 0 otherwise. 5. If the rotational angles are adjusting. and it is the same as in the DFT case. This ﬁgure has a Fig. 5. 4. The complexity of . k = 032 1 1 1 32. ( and the rotation angle are When the number of points determined. VOL. 1.1342 IEEE TRANSACTIONS ON SIGNAL PROCESSING. and x (n ) = 0 otherwise. 6. the following method for implementing the DFRFT can be applied: (38) (39) frequency. x (n ) = 1 when 06 n 6. and x(n) = 0 otherwise. DFRFT by GSA of a chirp signal. the DFT Hermite eigenvectors can be computed a priori. 4. Then. the continuous FRFT in Fig. 7. This concept is also applied in the time domain . NO. When = 3=7. closer output to the continuous FRFT than Fig. DFRFT by S method of a rectangular window function. the computation of the DFRFT can be implemented only by a transform kernel matrix multiplication. MAY 1999 Fig. x (n ) = 1 when 06 n 6. This ﬁgure has a closer Fig. identity transform) and any angle transform domains. The output is close to Fig. 47. and the eigenvalues of DFRFT are also determined. computing the DFRFT is . an impulse-like output is obtained. ej 20:1141k . 4. x(n) = 1 when 06 n 6. output to the continuous FRFT than Fig. DFRFT by OPA method of a rectangular window function.

and the output becomes a constant. DFRFT by GSA of an impulse function. we can obtain is also a DFRFT eigenvector. we stress the major reason for this mismatch. DFRFT by S method of an impulse function. Since any eigenvector in Therefore. Fig. x(0) = 1. 6 Fig. the inner products of signal and eigenvectors. It can be seen that chirp-like outputs are obtained while the angle 6= =2. A rigorous discussion for the mismatches of [17] has been presented in [16]. Proposition 6: The method in (6) assigns all the eigenvectors of the DFRFT matrix to only four different eigenvalues. DFRFT by OPA of an impulse function. When = =2. 8.: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1343 Fig. Proof: Let be any DFT eigenvector located in the . only should be recomputed. and the output becomes a constant. If the rotation angle is changed. and x(n) = 0 otherwise. . 6 (40) The deﬁnitions of matrix and matrix are the same as those . This is the major reason for the mismatches in [17]. but it cannot have similar results as in the continuous case. the DFRFT reduces to DFT. The coefﬁcients ’s are in (33). x(0) = 1. and x(n) = 0 otherwise. x(0) = 1. 9.PEI et al. Here. and the output becomes a constant. When = =2. When = =2. and they can be computed in advance. It can be seen that chirp-like outputs are obtained while the angle = =2. Discussion The method in [17] obeys the rotation properties. the diagonal matrix F. 10. . It can be seen that chirp-like outputs are obtained while the angle = =2. the DFRFT reduces to DFT. and the value is the eigenvalue for this eigenvector . the DFRFT reduces to DFT. and x(n) = 0 otherwise. Applying the transform eigensubspace kernel deﬁned in (6) to the eigenvector .

( one of the four eigensubspace has the same eigenvalue. and the frequency resolution is [23]. X is even. our DFRFT can be used for any rotation angle and can provide very similar results as continuous cases. On the other hand. and there are some approximation errors in the DFT Hermite approaches inﬁnity. it is also is the number of points of the discrete the time resolution. the method assigns the DFRFT eigenvalues to only four values. X is odd. 47. The development of the new DFRFT is based on the same idea as (2). the approximation eigenvectors. which satisﬁes the eigenvalue and eigenfunction relationship as a continuous FRFT kernel. Thus. x(0) = 1. the periods in the transform results change for different rotation angles. MAY 1999 Fig. VOL. The DFRFT developed in this paper is based on the mimicking of the eigenvalue–eigenfunction relationship that the continuous counterpart of this transform has with the normalized unit variance Hermite functions. NO. an alternative DFRFT has recently been proposed. While errors of the DFT Hermite eigenvectors will be reduced. the DFRFT can be used to give the similar continuous transform results within the accuracy of the discrete ﬁnite vector approximation. the signal and transform results in [31] are discrete and periodic. 5. In [31]. . It should be noted that the number of eigenfunctions for the continuous FRFT in (2) is inﬁnite. the number of DFT Hermite eigenvectors is only ﬁnite. However. and the rotation angle in [31] is valid for a certain discrete set of rotation angles. Therefore. Then. we will establish the relationship between the FRFT and the DFRFT. In Proposition 3. is the total recorded signal duration and is signal.b)-element in DFRFT kernel matrix F [vn ] = e0jn vn where vn is Hermite-like function N F [(k)] 2 10j2cot ej cot If x is even. and more DFT Hermite eigenvectors are used to compute the DFRFT. we will discuss the transform range and resolution for the DFRFT (see Table IV). Moreover. (41) is the eigenvector corresponding to the th-order where Hermite function. In the conventional DFT analysis. Here. and transform range in the time domain as well. we will of the DFT Hermite eigenvector is compute the FRFT for a Hermite function with any variance . b) is the (a. and x(t) = 0 otherwise. Transform Range and Resolution of DFRFT In this subsection. b) = F (b. we will review and understand the transform range and resolution for the is the conventional DFT. The the overall frequency range that the DFT can represent is equal to . A. In the DFRFT of [17]. Samples of continuous FRFT of an impulse function. TABLE III PROPERTIES OF DFRFT Unitary Angle additivity Time Inversion Periodicity Symmetry Eigenfunction Impulse Transform Parity F )3 = (F )01 = F0 F F =F F x(0n) = X (0n) F =F F (a. sampling interval for the original continuous signal. RELATIONSHIP BETWEEN FRFT AND DFRFT In this section. the proposed DFRFT has assigned a different eigenvalue for each eigenvector. To begin with. a) where F (a. there will be four different eigenvalues for all the eigenvectors. IV. the transform range and resolution of the DFT have been well discussed [23].1344 IEEE TRANSACTIONS ON SIGNAL PROCESSING. If x is odd. it has been proved that the spread variance . In the following discussion. It is based on retaining the property of the DFT that a sampled periodic function transforms into a periodic function. 11.

It is easy to check some special cases for (43): and . frequency ( The rotation concept of the DFRFT in the continuous is the time–frequency plane is plotted in Fig. but it actually performs elliptical rotation in the continuous time–frequency plane.PEI et al. B. For simpliﬁcation of analysis. the spread variance of the DFRFT for angle is equal to (47) The overall transform range the DFRFT can cover is equal to (50) In (34). whereas This means that the DFRFT is implemented in a circular rotation for the discrete case. of the FRFT of From Propositions 4 and 5. ( • The length of the time range is longer than that of the ). therefore. 12. From (45) and Fig. frequency ( • The length of the time range is shorter than that of the ). as drawn in Fig.: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1345 TABLE IV TRANSFORM RANGE AND RESOLUTION OF DFRFT Proposition 7: The continuous FRFT of a Hermite function with variance is equal to Moreover. (46) Thus. the DFRFT performs independent circular rotation in the discrete notation. • Circular Rotation: (51) . In this subsection. and is the where number of points for the discrete signal. Three cases can be realized for the signal rotation of the DFRFT in the continuous time–frequency plane. The symbol and are the two special cases. Angle actual angle for the elliptical rotation. The new variance is com- (43) Proof: This proposition can be easily proved by computing the FRFT of the normalized Hermite functions. the approximation errors of the DFT Hermite eigenvectors are ignored in the following discussion. the continuous time–frequency plane. only the last term can affect the envelope . but the DFRFT performs elliptical rotation in the continuous time–frequency plane while . The sampled spread variance of the DFRFT is vector in (12) can also be an approximate eigenvector in the fractional Fourier domain (angle ). it has been shown that there are approximation errors in the DFT Hermite eigenvectors. it has been known that the . • The lengths of time and frequency ranges are equal ). 12. we can substitute the variance in (43) by to get the spread variance of the DFRFT in the fractional Fourier domain (angle ) (48) Both (47) and (48) indicate the corresponding variance of DFT Hermite eigenvectors in the fractional Fourier domain. The overall transform range of the DFRFT can cover is equal to (45) Proof: From Proposition 3. Proposition 8: The resolution of the DFRFT with angular is equal to parameter (44) is the sampling interval of the signal. In (42). will be equal to 1 for any value of . the resolution of the FRFT can be obtained as (49) (42) where puted as . is used to denote the resolution of the FRFT. 12. it can be found that the signal rotation of the DFRFT is an elliptical rotation in . If . we will deﬁne circular and elliptical rotations clearly and establish the relationship between circular and elliptical rotations. Elliptical Rotation versus Circular Rotation The FRFT performs circular rotation in the continuous time–frequency plane.

VOL. (59) . While . the circular rotation is just the FRFT of the signal. (52) (53) is the transform kernel of circular rotation. is the elliptical rotation . and the post-phase compensa- is the transform kernel of elliptical rotation. the elliptical rotation will of signal become the circular rotation. Proposition 9: The circular rotation of signal can be implemented by an angle modiﬁcation of the elliptical rotation and post-multiplying a phase compensation factor. MAY 1999 Fig. NO. • Elliptical Rotation: where (54) (55) (56) spread variance changes in these two schemes. The transform kernels of the circular and elliptical rotations deﬁned above are inﬁnite numbers of sums for Hermite eigenvectors.1346 IEEE TRANSACTIONS ON SIGNAL PROCESSING. and where is deﬁned as in (43). (a) (b) (c) Fig. The critical point for which we care is the (58) can be written as the weighted Proof: The signal sum of the normalized Hermite functions. Results of Example 4 showing that the error norms between DFRFT and continuous FRFT samples are greatly reduced by angle modiﬁcation and post-phase compensation. 47. is the circular rotation of signal . 5. Three different cases of rotation concept for DFRFT in continuous time-frequency plane. If the Hermite eigenvectors in the circular rotation are from order 0 to inﬁnity. 12. (57) where tion factor is equal to . 13.

and the effects of angle modiﬁcation and postphase compensation will be very small. Step 1) Compute the modiﬁcation angle . 13(c) very well. The discrete post-phase compensation resolution is still factor for the DFRFT is the length of time range ( ) will almost equal the length ). V. The DFT Hermite eigenvectors can be calculated by the projection of samples of the unit variance Hermite functions in the DFT eigensubspaces through the help of the eigenvectors of the matrix . The length of the time range is almost equal to the length of the frequency range. and the other is OPA. The results shown in Fig. 13(a) is quite different from Fig. While the variance is far from unity. Step 3) Multiply the result obtained from Step 2 by the post-phase compensation factor shown in (62). whereas the OPA minimizes the total errors between the samples of the Hermite functions and the orthogonal DFT Hermite eigenvectors. (62) . The rotation angle . the sampling . 13(c) shows the sample values of the continuous FRFT . In Example 2. However. Fig. Therefore. The new transform retains the eigenvalue–eigenfunction relationship using the sampled version of the normalized unit-variance Hermite functions that the continuous FRFT has with the unit variance Hermite functions. the elliptical rotation can be implemented by circular rotation. for the indices . The correis equal to 1. we directly compute the DFRFT without angle modiﬁcation and postphase compensation. and Fig. the the unitary property in the DFRFT. . Fig. However.PEI et al. CONCLUSIONS The development of this DFRFT is based on the eigendecomposition of the DFT kernel matrix . the following steps must be used to make the results of discrete cases match those of the continuous case.: DISCRETE FRACTIONAL FOURIER TRANSFORM BASED ON ORTHOGONAL PROJECTIONS 1347 where . Two vector orthogonalization processes for the DFT Hermite eigenvectors are accordingly proposed in this paper: One is GSA. The in (58) is replaced by in (62) for preserving variable . the DFRFT in Example 2 is almost a circular rotation. 13(b) match the corresponding continuous FRFT cases shown in Fig. where . we again deal with the rectangular window shown in Example 2. With the help of the commutative matrix . such DFT Hermite eigenvectors cannot form an orthogonal basis for the DFT eigenspaces. Step 2) Calculate the -point DFRFT of the signal with the angular parameter . The GSA minimizes the errors between the samples of the Hermite functions and the orthogonal DFT Hermite eigenvectors from low to high orders. Fig. In this case. If variance of the continuous-time counterpart of the eigenvector will approximate to unity. we only compute the results of by the GSA. The operation in the DFRFT performs elliptical rotation in the continuous time–frequency plane for the case . Compute the FRFT of (circular rotation) for the both sides of (59). If the values of (57) are . it will become evaluated at the points (60) (61) can be Equation (61) indicates that the FRFT with angle implemented by a DFRFT with angle . Example 4: In this example. 13(b) is the DFRFT with angular parameter modiﬁcation and postphase compensation. which are equal to continuous counterpart of the variance in this example is . and the transform . The rotation will almost be of the frequency range ( a circular rotation. It must be noted that the postphase where compensation factors in (58) and (62) are different. 13(c) due to the elliptical rotation and not in the circular rotation in the continuous time–frequency plane. the modiﬁcation of the angle and postphase compensation discussed above are very critical. therefore.0488 in Example sponding variance 2. the complete real and orthonormal eigenvectors of the DFT kernel matrix can be computed. and the number of points in the interval is still signal becomes 37. 13(a) shows the DFRFT with angular parameter for the original signal. Using Proposition 9. Here.

44. T. 9. and it is an important tool for signal processing. 199–203. optical information processing. W. A. vol.” J. He was an Engineering Ofﬁcer in the Chinese Navy Shipyard from 1970 to 1971. 1998. R. Taipei. pattern recognition. G. H. From 1997 to 1998. and wavelets. Arikan.” IEEE Trans. 1992. G. Ozaktas. 33. Insel. From 1995 to 1997. Baltimore. the relationship between the FRFT and the . and A. 1992. The complexity for . 39. R. B. M. Dragonman. [23] A. in 1972 and 1975. J. pp. Taipei. vol. MAY 1999 Furthermore. B. pp. Almeida. “Generalized self-Fourier functions.-C. Math. Ltd. vol. Spence. May 1996. [6] H.” Signal [26] Process. vol. DFRFT can be established as follows. IEEE Int. 1416. and M..” in Proc. and Ph. NO. Barshan. Nov. “Relationships between the Radon–Wigner and fractional Fourier transforms. H. vol. pp. degree.S. in 1990 and 1995. Bourdeaux-Bartels. “Convolution and ﬁltering in fractional Fourier domain. degree in computer engineering from the National Chiao-Tung University. Weisburn. 1989. NJ: Prentice-Hall. Shenoy. New York: McGraw-Hill. Lohmann and B. 1994. Hlawatsch and G. [25] 22. vol. 547–559. 1989. 1. 13. “Two dimensional discrete fractional Fourier transform. vol. Van Loan. F. Onural. Mar.” J. Amer. [30] G. Santarsiero.” Appl.” Opt. pp.C. degree from the National Taiwan University (NTU). both from the National Taiwan University. Ozaktas.. Opt. Apr. H. [19] B. pp. Amer. vol. Friedberg. pp. 21–67. IEEE Int. pp. 45. 2141–2150. C. Chien-Cheng Tseng (S’90–M’95) was born in Taipei. The Fourier Transform and Its Applications.” IEEE Signal Processing Mag. 2522–2531. Erkaya..C. [9] L.” J.. Signal Processing. Matrix Computations. Soffer.” J.E. 1996. Discrete-Time Signal Processing. “Fractional Fourier transformations and their optical implementation—II. 1986. [10] D. pp. pp. N. Feb. Orthogonal Functions.D. in 1965.S. REFERENCES [1] R. Tatung Institute of Technology. pp. Shyu. B. Golub and C. 6th IEEE DSP Workshop. “Eigenvectors and functions of the discrete Fourier transform.. McBride and F. 1994. [2] A. vol. [18] G. NTU. “Discrete fractional Hartley and Fourier transforms.. pp. 10. all in electrical engineering. V. “Fractional Fourier transformations and their optical implementation—I. Linear Algebra. E. pp. Parks. 1994. 11.” J. and D. Yeh. H. Taipei.. and electronic commerce. Kutay. A.D. M.” Opt. M. vol. The DFRFT proposed in this paper not only supplies the similar transforms to match with those of the continuous case but also preserves the rotation properties. and R. “Convolution. C. 1982. He then received the M. Kutay.S. Taipei. M. A. 1953.. 1994. A. H. AU-20. M. Opt.. pp. N. Yeh. Opt. Bracewell. He was a Professor and Chairman with the Electrical Engineering Department. ﬁltering. 994–998. Feb. Englewood Cliffs. “Digital computation of the fractional Fourier transform. Ozaktas. and J. 1047–1049. pp.” Signal Process. Math. Taiwan. Presently.. ASSP-30. respectively. Mendlovic. Taiwan. Lohmann. F. [28] P. 303–309. O. 3. Ozaktas. 119–124. Oppenheim. Barshan. Arikan. A. 241–265. M. [8] A. pp. Opt. Inst. degree in electrical engineering in 1997.. Time–Freq. 1875–1881. Hsinchu. “Fractional Wigner distribution function. Taipei. [31] O. Morse and H. Mendlovic.. “Chirp ﬁltering in the fractional Fourier domain. 47. Sept. Lett. Taipei. 1997. Ozaktas. vol.E. “Effect of fractional Fourier transformation on time-frequency distributions belonging to the Cohen class. H. “Eigenvalue and eigenvector decomposition of the discrete Fourier transform. 11. Jan. he was an Associate Research Engineer at Telecommunication Laboratories. 1191–1194. Chunghwa Telecom Company. Interscience. July 15. pp. 536–539. 2nd ed. if the DFRFT becomes an elliptical rotation in the continuous time–frequency plane. Amer. Ozaktas. W. Time-Scale Anal. 1996. Mendlovic. VOL. M. [7] R. Apr. He received the B.” J. and G. W. Tamsui Oxford University College. he was an Assistant Professor of the Department of Electronics Engineering. [17] B.E. image processing. Hwa Hsia College of Technology and Commerce. June 1996. .O. 1989. M. pp. If the DFRFT performs a circular rotation of the signal in the . 67. He received the B. 1980. Circuits Syst. vol. Taiwan. His main research interests are in fractional Fourier transforms. M. 205–207. NJ: Prentice-Hall.S. and L. Symp.. This DFRFT provides a method for implementing the discrete FRFT. pp.” IEEE Trans. II. F. . Signal Processing. Soo-Chang Pei (SM’89) was born in Soo-Auo. vol. [11] H. Tamsui. and laser holography. degrees from the National Taiwan University. His research interests include digital signal processing. Gori.. A. Applicat. [21] J. degrees from the University of California. Press. 1993.” IEEE Signal Processing Lett. 1998. 1995. 1972. An angle modiﬁcation and a postphase compensation in the DFRFT for elliptical rotation are required to match the results that are similar for the continuous FRFT.” J. Opt. W. Taoyuan.” IEEE Trans. Englewood Cliffs. and multiplexing in fractional Fourier domains and their relationship to chirp and wavelet transforms.S. vol. He is currently an Assistant Professor with the Department of Computer and Communication Engineering. J. [5] H. “On Namias’ fractional Fourier transforms. Methods of Theoretical Physics. “The discrete fractional Fourier transform. M. From 1971 to 1975. A.S. vol.. Santhanam and J. and Ph. Pei and M. which is the same as that implementing the DFRFT is of the DFT. “The fractional Fourier transform and time-frequency representation. 42. 66–74. H. pp. . Bitran. 40–41. vol. 665–675. [4] F. New York: McGraw-Hill. pp. respectively. 34. 159–175. 1987. Cincotti. Taiwan. Ozaktas. H. time–frequency analysis. “Fractional Fourier domains. Sansone. Pei. in 1949. Dr. Mendlovic. W.1348 IEEE TRANSACTIONS ON SIGNAL PROCESSING. [20] B.” in Proc. [14] D. he was a Research Assistant at UCSB. 1996.” IEEE Trans. [12] D. pp. He is currently an Assistant Professor with the Department of Computer Information Science. in 1988. Mar. [16] H. vol. pp. [24] S. Opt. “The fractional order Fourier transform and its application to quantum mechanics. A. [29] S. “Improved discrete fractional Fourier transform.O. 1993.” IEEE Trans. Signal Processing. Y. McClellan. in 1964. from 1981 to 1983. vol. Bozdagi. Rev. Feschbach. Santa Barbara (UCSB). 1994. in 1970 and the M.-H. “Fractional correlation. Lohmann. D. Kerr. he is Professor and Chairman of the Electrical Engineering Department. 3084–3091. 42. Circuits Syst. Taiwan. vol. K. in 1987. “Linear and quadratic timefrequency signal representations. Tseng. Feb. and the M. pp. with honors. Parks. and L. Opt. “Discrete fractional Fourier transform. New York.. Appl. Acoust.” IMA J. vol. MD: Johns Hopkins Univ. 25. degree in computer science and information engineering in 1992 and the Ph.C. Taipei. Soc. Speech. 1995. On the other hand. [13] H. Dorsch. Aademir. 5.E. 7599–7602. C.O.. Steiglitz. M. A. Mendlovic and H. and M. C. 46. Phys. H.” IEEE Trans. 474–478. pp. Namias. 99–108. National Kaohsiung First University of Science and Technology. Kutay. and D. 25.D. 1798–1801.” Appl. 1959. Amer. A. Min-Hung Yeh was born in Taipei. R. pp. 25–31. Signal Processing. vol.. 10. [22] G. Taipei. 1996. M. “Separation of transient signals.. McClellan and T. He received the B. time–frequency plane. Amer. p. W. His research interests include digital signal processing. “The discrete rotational Fourier transform. Soc. Soc. A. Pei is a Member of Eta Kappa Nu and the Optical Society of America. and O. Ozaktas. Symp. 15–16. from the Tatung Institute of Technology. June 1994. Soc. vol. Audio Electroacoust. Soc. [27] S. Dickinson and K. [3] V. [15] .” in Proc. Oct..

3 Equality

Venus & Adonis - Shakespeare

Troilus & Cressida - Shakespeare

Tempest - Shakespeare

Hamlet by William Shakespeare

Phoenix & Turtle - Shakespeare

Passionate Pilgrim - Shakespeare

Othello - Shakespeare

Merry Wives of Wndsor - Shakespeare

Merchant of Venice - Shakespeare

Lucrece - Shakespeare

Love's Labours Lost - Shakespeare

Julius Caeser - Shakespeare

Comedy of Errors - Shakespeare

As You Like It - Shakespeare

Antony & Cleopatra - Shakespeare

All is Well That Ends Well - Shakespeare

UBUNTU8.04TUT

PC Magazine - Linux Solutions Malestrom

Linux Sys Calls

Linux Drivers

Linux Unix Concepts

Intro Linux Device Drivers

Embedded Linux Kernel and Drivers

Arcom Embedded Linux Technical Manual

- Leia e imprima sem anúncios
- Download to keep your version
- Edit, email or read offline

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

CANCEL

OK

You've been reading!

NO, THANKS

OK

scribd