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

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

### Publishers

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks 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

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

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

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

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

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

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

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

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

The complexity of . 5. DFRFT by GSA method of a rectangular window function. and it is the same as in the DFT case. This ﬁgure has a Fig. ej 20:1141k . The output is close to Fig. 7. DFRFT by OPA method of a rectangular window function. and the eigenvalues of DFRFT are also determined. 6. Then. 5. ( and the rotation angle are When the number of points determined. This concept is also applied in the time domain . If the rotational angles are adjusting. This ﬁgure has a closer Fig. and x(n) = 0 otherwise. NO. x (n ) = 1 when 06 n 6. x(n) = 1 when 06 n 6. the following method for implementing the DFRFT can be applied: (38) (39) frequency. 47. VOL.1342 IEEE TRANSACTIONS ON SIGNAL PROCESSING. DFRFT by GSA of a chirp signal. an impulse-like output is obtained. k = 032 1 1 1 32. MAY 1999 Fig. closer output to the continuous FRFT than Fig. 4. and x (n ) = 0 otherwise. When = 3=7. DFRFT by S method of a rectangular window function. computing the DFRFT is . 4. and x (n ) = 0 otherwise. the continuous FRFT in Fig. x (n ) = 1 when 06 n 6. the computation of the DFRFT can be implemented only by a transform kernel matrix multiplication. the DFT Hermite eigenvectors can be computed a priori. 1. identity transform) and any angle transform domains. 4. output to the continuous FRFT than Fig.

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

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

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

If the Hermite eigenvectors in the circular rotation are from order 0 to inﬁnity. Three different cases of rotation concept for DFRFT in continuous time-frequency plane. the elliptical rotation will of signal become the circular rotation. • Elliptical Rotation: where (54) (55) (56) spread variance changes in these two schemes. (57) where tion factor is equal to .1346 IEEE TRANSACTIONS ON SIGNAL PROCESSING. 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. 47. and the post-phase compensa- is the transform kernel of elliptical rotation. The transform kernels of the circular and elliptical rotations deﬁned above are inﬁnite numbers of sums for Hermite eigenvectors. (52) (53) is the transform kernel of circular rotation. While . MAY 1999 Fig. 13. 5. and where is deﬁned as in (43). 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. is the circular rotation of signal . (59) . the circular rotation is just the FRFT of the signal. is the elliptical rotation . 12. (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. NO. VOL.

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

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

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

Download

by Anthony Doerr

by Graeme Simsion

by Lindsay Harrison

by Brad Kessler

by David McCullough

by Jaron Lanier

by Miriam Toews

by Meg Wolitzer

by Alexandra Horowitz

by Ernest Hemingway

by Lewis MacAdams

by Katie Ward

by Nic Pizzolatto

by Rachel Kushner

by Aravind Adiga

by David Gordon

by Maxim Biller

by Åke Edwardson

On Writing

The Shell Collector

I Am Having So Much Fun Here Without You

The Rosie Project

The Blazing World

Quiet Dell

After Visiting Friends

Missing

Birds in Fall

Who Owns the Future?

All My Puny Sorrows

The Wife

A Farewell to Arms

Birth of the Cool

Girl Reading

Galveston

Telex from Cuba

Rin Tin Tin

The White Tiger

The Serialist

No One Belongs Here More Than You

Love Today

Sail of Stone

Fourth of July Creek

The Walls Around Us

Are you sure?

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

CANCEL

OK

scribd