Gabor Frames on Local Fields of Positive Characteristic Firdous A. Shah∗ 6 ∗DepartmentofMathematics, UniversityofKashmir, SouthCampus, Anantnag-192101, Jammu 1 and Kashmir, India. E-mail: [email protected] 0 2 n Abstract: Gabor frames have gained considerable popularity duringthe past decade, primarily a due to their substantiated applications in diverse and widespread fields of engineering and sci- J ence. Finding general and verifiable conditions which imply that the Gabor systems are Gabor 0 framesisamongthecoreproblemsintime-frequencyanalysis. Inthispaper,wegivesomesimple 2 andsufficientconditionsthatensureaGaborsystem M T g =:χ (bx)g x−u(n)a ] u(m)b u(n)a m m,n∈N0 A to be a frame for L2(K). The conditions proposed are stated in terms of the Fourier transforms (cid:8) (cid:0) (cid:1)(cid:9) F of the Gabor system’s generating functions. . h t Keywords: Gabor frame; Local field; Fourier transform a m 2010 Mathematics Subject Classification: 42C15; 42C40; 42B10; 43A70; 46B15 [ 1 1. Introduction v 8 7 The notion of frame was first introduced by Duffin and Schaeffer [1] in connection with 2 some deep problems in non-harmonic Fourier series. Frames are basis-like systems that 5 0 span a vector space but allow for linear dependency, which can be used to reduce noise, . 1 find sparse representations, or obtain other desirable features unavailable with orthonor- 0 mal bases. The idea of Duffin and Schaeffer did not generate much interest outside non- 6 1 harmonic Fourier series until the seminal work by Daubechies, Grossmann, and Meyer [2]. : v They combined the theory of continuous wavelet transforms with the theory of frames to Xi introduce wavelet (affine) frames for L2(R). After their work, the theory of frames began r to be studied widely and deeply. Today, the theory of frames has become an interesting a and fruitful field of mathematics with abundant applications in signal processing, image processing, harmonic analysis, Banach space theory, sampling theory, wireless sensor net- works, optics, filter banks, quantum computing, and medicine and so on. An introduction to the frame theory and its applications can be found in [3, 4]. Gabor frames form a special kind of frames for L2(R) whose elements are gener- ated by time-frequency shifts of a single window-function or atom. More specifically, let g ∈ L2(R) and a,b ∈ R+, we use G(g,a,b) to denote the Gabor family or sys- tem {M T g : m,n ∈ Z} generated by g where T f(x) = f(x − na) is the trans- mb na na lation unitary operator and M f(x) = e2πimbxf(x) is the modulation unitary opera- mb tor. The composition M T is called the time-frequency shift operator. The system mb na {M T g : m,n ∈ Z} is called a Gabor frame if there exist constants A,B > 0 such that mb na A f 2 ≤ f,M T g 2 ≤ B f 2, for all f ∈ L2(R). (1.1) 2 mb na 2 m∈Zn∈Z (cid:13) (cid:13) XX(cid:12)(cid:10) (cid:11)(cid:12) (cid:13) (cid:13) (cid:13) (cid:13) (cid:12) (cid:12) (cid:13) (cid:13) 2 Gabor systems that form frames for L2(R) have a wide variety of applications. In practice, once the window function has been chosen, the first question to investigate for Gabor analysis is to find the values of the time-frequency parameters a,b such that {M T g} is a frame. A useful tool in this context is the Ron and Shen [5] criterion. mb na m,n∈Z By using this criterion, Gr¨ochenig et al.[6] have proved that the system {M T g} mb na m,n∈Z cannot be a frame for a > 0 and b integer greater than 1. Many results in this area, including necessary conditions and sufficient conditions have been established during the last two decades [7–10]. We refer the reader to the books [11, 12] for a comprehensive treatment of Gabor frames. A field K equipped with a topology is called a local field if both the additive and multiplicative groups of K are locally compact Abelian groups. For example, any field endowed with the discrete topology is a local field. For this reason we consider only non-discrete fields. The local fields are essentially of two types (excluding the connected local fields R and C). The local fields of characteristic zero include the p-adic field Q . p Examples of local fields of positive characteristic are the Cantor dyadic group and the Vilenkin p-groups. Local fields have attracted the attention of several mathematicians, and have found innumerable applications not only in the number theory, but also in the representation theory, division algebras, quadratic forms and algebraic geometry. As a re- sult, local fields are now consolidated as a part of the standard repertoire of contemporary mathematics. For more details we referr to [13]. The local field K is a natural model for the structure of Gabor frame systems, as well as a domain upon which one can construct Gabor basis functions. There is a sub- stantial body of work that has been concerned with the construction of Gabor frames on K, or more generally, on local fields of positive characteristic. Jiang et al.[14] con- structed Gabor frames on local fields of positive characteristic using basic concepts of operator theory and have established a necessary and sufficient conditions for the system M T g =: χ (bx)g x−u(n)a to be a frame for L2(K). Recently, Shah u(m)b u(n)a m m,n∈N0 [15] established a complete characterization of Gabor frames on local fields by virtue of (cid:8) (cid:0) (cid:1)(cid:9) two basic equations in the Fourier domain and show how to construct an orthonormal Gabor basis for L2(K). Recent results related to wavelet and Gabor frames on local fields of prime characteristic can be found in [16–20] and the references therein. In this article, we continue our investigation on Gabor frames on local fields and will present generalized inequalities for Gabor frames on local fields of positive characteristic via Fourier transform. The inequalities we proposed are stated in terms of the Fourier transforms of the Gabor system’s generating functions, and the inequalities are better than that of Li and Jiang [14]. Although we consider a one-dimensional case here, our results areeasily generalized tomulti-dimensional Gaborsystems onlocalfields ofpositive characteristic. The paper is organized as follows. In Section 2, we discuss some preliminary facts about local fields of positive characteristic and state the main results. Section 3 gives the proofs of the results. 3 2. Preliminaries on Local Fields Let K be a field and a topological space. Then K is called a local field if both K+ and K∗ are locally compact Abelian groups, where K+ and K∗ denote the additive and multiplicative groups of K, respectively. If K is any field and is endowed with the discrete topology, then K is a local field. Further, if K is connected, then K is either R or C. If K is not connected, then it is totally disconnected. Hence by a local field, we mean a field K which is locally compact, non-discrete and totally disconnected. We use the notation of the book by Taibleson [13]. Proofs of all the results stated in this section can be found in the book [13]. Let K be a local field. Let dx be the Haar measure on the locally compact Abelian groupK+. Ifα ∈ K andα 6= 0, thend(αx)isalsoaHaarmeasure. Letd(αx) = |α|dx. We call |α| the absolute value of α. Moreover, the map x → |x| has the following properties: (a) |x| = 0 if and only if x = 0; (b) |xy| = |x||y| for all x,y ∈ K; and (c) |x + y| ≤ max{|x|,|y|} for all x,y ∈ K. Property (c) is called the ultrametric inequality. The set D = {x ∈ K : |x| ≤ 1} is called the ring of integers in K. Define B = {x ∈ K : |x| < 1}. The set B is called the prime ideal in K. The prime ideal in K is the unique maximal ideal in D and hence as result B is both principal and prime. Since the local field K is totally disconnected, so there exist an element of B of maximal absolute value. Let p be a fixed element of maximum absolute value in B. Such an element is called a prime element of K. Therefore, for such an ideal B in D, we have B = hpi = pD. As it was proved in [13], the set D is compact and open. Hence, B is compact and open. Therefore, the residue space D/B is isomorphic to a finite field GF(q), where q = pk for some prime p and k ∈ N. Let D∗ = D\B = {x ∈ K : |x| = 1}. Then, it can be proved that D∗ is a group of units in K∗ and if x 6= 0, then we may write x = pkx′,x′ ∈ D∗. For a proof of this fact we refer to [13]. Moreover, each Bk = pkD = x ∈ K : |x| < q−k is a compact subgroup of K+ and usually known as the fractional ideals of K+. Let U = {a }q−1 be any fixed (cid:8) (cid:9) i i=0 full set of coset representatives of B in D, then every element x ∈ K can be expressed uniquely as x = ∞ c pℓ with c ∈ U. Let χ be a fixed character on K+ that is trivial ℓ=k ℓ ℓ on D but is non-trivial on B−1. Therefore, χ is constant on cosets of D so if y ∈ Bk, P then χ (x) = χ(yx),x ∈ K. Suppose that χ is any character on K+, then clearly the y u restriction χ |D is also a character on D. Therefore, if {u(n) : n ∈ N } is a complete u 0 list of distinct coset representative of D in K+, then, as it was proved in [13], the set χ : n ∈ N of distinct characters on D is a complete orthonormal system on D. u(n) 0 (cid:8) (cid:9) The Fourier transform fˆof a function f ∈ L1(K)∩L2(K) is defined by ˆ f(ξ) = f(x)χ (x)dx. (2.1) ξ ZK It is noted that ˆ f(ξ) = f(x)χ (x)dx = f(x)χ(−ξx)dx. ξ ZK ZK 4 Furthermore, the properties of Fourier transform on local field K are much similar to those of on the real line. In particular Fourier transform is unitary on L2(K). We now impose a natural order on the sequence {u(n)}∞ . We have D/B∼= GF(q) n=0 where GF(q) is a c-dimensional vector space over the field GF(p). We choose a set {1 = ζ0,ζ1,ζ2,...,ζc−1} ⊂ D∗ such that span {ζj}cj−=10 ∼= GF(q). For n ∈ N0 satisfying 0 ≤ n < q, n = a +a p+···+a pc−1, 0 ≤ a < p, and k = 0,1,...,c−1, 0 1 c−1 k we define u(n) = (a +a ζ +···+a ζ )p−1. (2.2) 0 1 1 c−1 c−1 Also, for n = b +b q +b q2 +···+b qs, n ∈ N , 0 ≤ b < q,k = 0,1,2,...,s, we set 0 1 2 s 0 k u(n) = u(b )+u(b )p−1 +···+u(b )p−s. (2.3) 0 1 s This defines u(n) for all n ∈ N . In general, it is not true that u(m+n) = u(m)+u(n). 0 But, if r,k ∈ N and 0 ≤ s < qk, then u(rqk+s) = u(r)p−k+u(s). Further, it is also easy 0 to verify that u(n) = 0 if and only if n = 0 and {u(ℓ)+u(k) : k ∈ N } = {u(k) : k ∈ N } 0 0 for a fixed ℓ ∈ N . Hereafter we use the notation χ = χ , n ≥ 0. 0 n u(n) Let the local field K be of characteristic p > 0 and ζ ,ζ ,ζ ,...,ζ be as above. 0 1 2 c−1 We define a character χ on K as follows: exp(2πi/p), µ = 0 and j = 1, χ(ζ p−j) = (2.4) µ 1, µ = 1,...,c−1 or j 6= 1. (cid:26) We also denote the test function space on K by Ω, i.e., each function f in Ω is a finite linear combination of functions of the form 1 (x−h),h ∈ K,k ∈ Z, where 1 is the k k characteristic function of Bk. Then, it is clear that Ω is dense in Lp(K),1 ≤ p < ∞, and each function in Ω is of compact support and so is its Fourier transform. For a given ψ ∈ L2(K), define the Gabor system G(g,a,b) := M T g =: χ (bx)g x−u(n)a : n,m ∈ N . (2.5) u(m)b u(n)a m 0 n o (cid:0) (cid:1) We call the Gabor system G(g,a,b) a Gabor frame for L2(K), if there exist positive numbers 0 < C ≤ D < ∞ such that for all f ∈ L2(K) 2 2 2 C f ≤ f,M T g ≤ D f . (2.6) 2 u(m)b u(n)a 2 (cid:13) (cid:13) mX∈N0nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) (cid:13) (cid:13) Before stating our res(cid:13)ult(cid:13)s, we introduce(cid:12) some notations. F(cid:12)or any(cid:13)g(cid:13)∈ L2(K) and a,b > 0. 5 We set ∆ (ξ) = gˆ ξ −bu(m) gˆ ξ −bu(m)+a−1u(k) , k mX∈N0(cid:12)(cid:12) (cid:0) (cid:1) (cid:0) (cid:1)(cid:12)(cid:12) (cid:12) (cid:12) α = ess sup∆ (ξ), k ∈ N , β = α , γ = essinf∆ (ξ), k k 0 k 0 ξ k∈N ξ X Λ (ξ) = gˆ ξ −bu(m) gˆ ξ −bu(m)+a−1u(k) , k kX∈N0 (cid:0) (cid:1) (cid:0) (cid:1) δ = esssup Λ (ξ) , k ∈ N , µ = δ . k k 0 k ξ k∈N (cid:12) (cid:12) X (cid:12) (cid:12) Based on these notations, the authors in [14] established the following result. Theorem 2.1. Let a,b > 0 and g ∈ L2(K). If α ,β and γ satisfy 0 β < γ ≤ α < ∞, 0 C D then system in (2.5) constitutes a Gabor frame for L2(K) with bounds 1 and 1, where a a C = γ −β and D = α +β. 1 1 0 Motivating by thefundament works in[14,15], we will give two new sufficient conditions of Gabor frame on local fields of prime characteristic in this paper. The conditions obtained are better than that of one in Theorem 2.1. Now, the first result of the paper is stated as follows. Theorem 2.2. Let a,b > 0 and g ∈ L2(K). If α ,γ and µ satisfy 0 µ < γ ≤ α < ∞, (2.7) 0 C then the Gabor system G(g,a,b) as defined in (2.5) is a frame for L2(K) with bounds 2 a D 2 and , where C = γ −µ and D = α +µ. 2 2 0 a Remark 1. It is easy to see that µ ≤ β, so the frame bounds in Theorem 2.2 are better than ones in Theorem 2.1. Next, we prove a more general result which includes not only the results of Theorem 2.1 and 2.2 as special cases, but also leads to a standard development of interesting generalizations of some well-known results. To do so, we set σ = esssup Λ . k ξ k∈N X(cid:12) (cid:12) (cid:12) (cid:12) Theorem 2.3. Let a,b > 0 and g ∈ L2(K). If α ,γ and σ satisfy 0 6 σ < γ ≤ α < ∞, (2.8) 0 then the Gabor system G(g,a,b) given by (2.5) constitutes a frame for L2(K) with bounds C D 3 3 and , where C = γ −σ and D = α +σ. 3 3 0 a a Remark 2. Sinceσ ≤ µ, theframeboundsinTheorem2.3arebetterthanonesinTheorem 2.2. 3. Proof of the Main Results In order to prove Theorems 2.2 and 2.3, we need the following lemma whose proof can be found in [3]. Lemma 3.1. Suppose that {f }∞ is a family of elements in a Hilbert space H such that k k=1 there exist constants 0 < A ≤ B < ∞ satisfying ∞ 2 2 2 A f ≤ f,f ≤ B f , 2 k 2 k=1 (cid:13) (cid:13) X(cid:12)(cid:10) (cid:11)(cid:12) (cid:13) (cid:13) (cid:13) (cid:13) (cid:12) (cid:12) (cid:13) (cid:13) for all f belonging to a dense subset D of H. Then, the same inequalities are true for all ∞ f ∈ H; that is, {f } is a frame for H. k k=1 In view of Lemma 3.1, we will consider the following set of functions: Ω0 = f ∈ Ω : suppfˆ⊂ K\{0} and fˆ < ∞ . ∞ n o (cid:13) (cid:13) Since Ω is dense in L2(K) and closed under the Fourie(cid:13)r t(cid:13)ransforms, the set Ω0 is also dense in L2(K). Therefore, it is enough to verify that the system G(g,a,b) given by (2.5) is a frame for L2(K) if the results of Theorems 2.2 and 2.3 hold for all f ∈ Ω0. Assume that f ∈ L2(K) and h ∈ Ω0, then by periodization, we have h(ξ)f(ξ)χ a(ξ −ω) dξ = h ξ +a−1u(k) f ξ +a−1u(k) χ a(ξ −ω) dξ k k ZK (cid:0) (cid:1) ZGa−1 kX∈N0 (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) Since h lies in Ω0, so it is a bounded and compactly supported, therefore the number of k in the above sum is finite. Thus, we can say the series h(ξ)f(ξ)χ a(ξ −ω) dξ (3.1) k kX∈N0ZK (cid:0) (cid:1) is convergent to a periodic function H(ξ) ∈ L2(Ga−1), where H(ξ) = h ξ +a−1u(k) f ξ +a−1u(k) , and G = {x ∈ K : |x| ≤ |a|}. a kX∈N0 (cid:0) (cid:1) (cid:0) (cid:1) 7 Proof of Theorem 2.2. For any f ∈ L2(K), there exists a function sequence {f }∞ ⊂ j j=1 Ω0, such that fˆ −fˆ → 0 as j → ∞, and supp fˆ ⊂ Bj j j 2 (cid:13) (cid:13) since Ω0 is dense in L(cid:13)(cid:13)2(K). F(cid:13)(cid:13)or fixed m ∈ N0, define the functional 2 P (h) = h,g 2 = hˆ,gˆ , h ∈ L2(K). (3.2) m m,n m,n nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) Since the Fourier transform of gˆ (ξ) = χ a ξ −bu(m) gˆ ξ −bu(m) , m,n n (cid:16) (cid:17) (cid:0) (cid:1) (cid:0) (cid:1) therefore, by using equation (3.1), we are able to express (3.2) as 2 ˆ P (f ) = f ,gˆ m j j m,n nX∈N0(cid:12)D E(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) ˆ ˆ = f ,gˆ f ,gˆ j m,n j m,n nX∈N0D ED E ˆ = f (ξ)gˆ ξ −bu(m) χ a ξ −bu(m) dξ j n nX∈N0ZK (cid:0) (cid:1) (cid:16) (cid:0) (cid:1)(cid:17) ˆ × f (ω)gˆ ω −bu(m) χ a ω −bu(m) dω j n ZK (cid:16) (cid:17) (cid:0) (cid:1) (cid:0) (cid:1) 1 = fˆ ξ +a−1u(k) gˆ ξ −bu(m)+a−1u(k) fˆ(ξ)gˆ ω −bu(m) dξ. j j a kX∈N0ZK (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) Let 2 P(f) = f,g = P (f), (3.3) m,n m mX∈N0nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) mX∈N0 then (cid:12) (cid:12) 1 P(f ) = fˆ ξ +a−1u(k) gˆ ξ −bu(m)+a−1u(k) fˆ(ξ)gˆ ω −bu(m) dξ j j j a kX∈N0mX∈N0ZK (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) = Q (f )+Q (f ), (3.4) 1 j 2 j where 1 2 ˆ Q (f ) = f (ξ)gˆ ω −bu(m) dξ (3.5) 1 j j a mX∈N0ZK (cid:12)(cid:12) (cid:0) (cid:1)(cid:12)(cid:12) (cid:12) (cid:12) 1 Q (f ) = fˆ ξ +a−1u(k) gˆ ξ −bu(m)+a−1u(k) fˆ(ξ)gˆ ξ −bu(m) dξ. 2 j j j a Xk∈NmX∈N0ZK (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) (3.6) 8 Since α < ∞, the series Q (f ) is convergent and 0 1 j γ 2 α 2 ˆ 0 ˆ f ≤ Q (f ) ≤ f , j 1 j j a 2 a 2 (cid:13) (cid:13) (cid:13) (cid:13) or equivalently (cid:13) (cid:13) (cid:13) (cid:13) γ(cid:13) (cid:13) α (cid:13) (cid:13) 2 0 2 f ≤ Q (f ) ≤ f . (3.7) a j 2 1 j a j 2 (cid:13) (cid:13) (cid:13) (cid:13) Next, we claim that Q (f ) is ab(cid:13)sol(cid:13)utely convergent. (cid:13)To(cid:13)prove this, we set 2 j 1 Q∗(f ) = fˆ ξ +a−1u(k) gˆ ξ −bu(m)+a−1u(k) fˆ(ξ)gˆ ξ −bu(m) dξ . 2 j a j j Xk∈NmX∈N0(cid:12)(cid:12)ZK (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) (cid:12)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) Note that 1 2 2 gˆ ξ−bu(m)+a−1u(k) gˆ ξ−bu(m) ≤ gˆ ξ −bu(m)+a−1u(k) + gˆ ξ −bu(m) , 2 (cid:18) (cid:19) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) (cid:0) (cid:1) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) h(cid:12)ence we have (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1 Q∗(f ) ≤ fˆ ξ +bu(m)+a−1u(k) fˆ ξ +a−1u(k) gˆ(ξ) 2dξ. 2 j a j j Xk∈NmX∈N0ZK (cid:12) (cid:0) (cid:1) (cid:0) (cid:1)(cid:12)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) Since each f is bounded and compactly supported on Bj, and in fact they belongs to Ω0, j hence there exist a constant M > 0 such that 2 ∗ ˆ 2 Q (f ) ≤ M f g < ∞, 2 j j ∞ (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) which proves our claim that Q (f ) is abso(cid:13)lut(cid:13)ely(cid:13)co(cid:13)nvergent. 2 j Using CauchySchwarz inequality, we obtain 1 Q (f ) = fˆ ξ +a−1u(k) fˆ(ξ)Λ (ξ)dξ 2 j j j k a (cid:12) (cid:12) (cid:12)(cid:12) Xk∈NZK (cid:0) (cid:1) (cid:12)(cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1 ≤ fˆ ξ +a−1u(k) Λ (ξ) 1/2 fˆ(ξ) Λ (ξ) 1/2 dξ j k j k a Xk∈NZK n(cid:12) (cid:0) (cid:1)(cid:12)(cid:12) (cid:12) on(cid:12) (cid:12)(cid:12) (cid:12) o (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1/2 1/2 1 2 2 ≤ fˆ ξ +a−1u(k) Λ (ξ) dξ fˆ(ξ) Λ (ξ) dξ . j k j k a Xk∈N(cid:26)ZK (cid:12) (cid:0) (cid:1)(cid:12) (cid:12) (cid:12) (cid:27) (cid:26)ZK (cid:12) (cid:12) (cid:12) (cid:12) (cid:27) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (3.8) (cid:12) (cid:12) (cid:12) (cid:12) It is easy to verify that fˆ ξ +a−1u(k) = fˆ(ξ), and Λ ξ −a−1u(k) = Λ (ξ), ∀ k ∈ N. j j k k (cid:0) (cid:1) (cid:0) (cid:1) 9 Thus, we have 1 2 µ ˆ 2 Q (f ) ≤ f (ξ) dξ δ = f , 2 j a j k a j 2 (cid:12) (cid:12) ZK (cid:12) (cid:12) Xk∈N (cid:13) (cid:13) (cid:12) (cid:12) (cid:12) (cid:12) or equivalently, (cid:12) (cid:12) (cid:12) (cid:12) (cid:13) (cid:13) µ µ 2 2 − f ≤ Q (f ) ≤ f . (3.9) a j 2 2 j a j 2 (cid:13) (cid:13) (cid:13) (cid:13) It follows from (3.7) and (3.9) tha(cid:13)t (cid:13) (cid:13) (cid:13) γ −µ α +µ 2 0 2 f ≤ P(f ) ≤ f . a j 2 j a j 2 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) Letting j → ∞ in above inequality, we obtain γ −µ α +µ 2 0 2 f ≤ P(f) ≤ f , a 2 a 2 or (cid:13) (cid:13) (cid:13) (cid:13) C (cid:13) (cid:13) (cid:13)D(cid:13) 2 2 2 2 2 f ≤ f,g ≤ f , a 2 m,n a 2 (cid:13) (cid:13) mX∈N0nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) (cid:13) (cid:13) (cid:13) (cid:13) (cid:12) (cid:12) (cid:13) (cid:13) where C = γ −µ and D = α +µ. This completes the proof of Theorem 2.2. 2 2 0 Proof of Theorem 2.3. Similar to the proof of Theorem 2.2, (3.5)–(3.8) hold. It follows from (3.8), the Cauchy-Schwarz inequality that 1/2 1/2 1 2 2 Q (f ) ≤ fˆ ξ +a−1u(k) Λ (ξ) dξ fˆ(ξ) Λ (ξ) dξ 2 j j k j k a (cid:12)(cid:12) (cid:12)(cid:12) (Xk∈NZK (cid:12)(cid:12) (cid:0) (cid:1)(cid:12)(cid:12) (cid:12) (cid:12) ) (Xk∈NZK (cid:12)(cid:12) (cid:12)(cid:12) (cid:12) (cid:12) ) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1/2 1/2 1 2 2 = fˆ(ξ) Λ ξ −a−1u(k) dξ fˆ(ξ) Λ (ξ) dξ j k j k a (Xk∈NZK (cid:12) (cid:12) (cid:12) (cid:0) (cid:1)(cid:12) ) (Xk∈NZK (cid:12) (cid:12) (cid:12) (cid:12) ) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) 1/2 1/2 1 2 2 ˆ ˆ = f (ξ) Λ (ξ) dξ f (ξ) Λ (ξ) dξ j k j k a (ZK (cid:12) (cid:12) Xk∈N(cid:12) (cid:12) ) (ZK (cid:12) (cid:12) Xk∈N(cid:12) (cid:12) ) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) (cid:12) σ 2 ≤ f , (3.10) a j 2 (cid:13) (cid:12) (cid:13) (cid:12) which implies that σ σ 2 2 − f ≤ Q (f ) ≤ f . (3.11) a j 2 2 j a j 2 (cid:13) (cid:13) (cid:13) (cid:13) Combining (3.7) with (3.11), we o(cid:13)bta(cid:13)in (cid:13) (cid:13) γ −σ α +σ 2 0 2 f ≤ P(f ) ≤ f . a j 2 j a j 2 (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) (cid:13) 10 By taking j → ∞ in above relation, we get γ −σ α +σ 2 0 2 f ≤ P(f) ≤ f , a 2 a 2 (cid:13) (cid:13) (cid:13) (cid:13) or (cid:13) (cid:13) (cid:13) (cid:13) C D 3 2 2 3 2 f ≤ f,g ≤ f , a 2 m,n a 2 (cid:13) (cid:13) mX∈N0nX∈N0(cid:12)(cid:10) (cid:11)(cid:12) (cid:13) (cid:13) (cid:13) (cid:13) (cid:12) (cid:12) (cid:13) (cid:13) where C = γ −σ and D = α +σ. This completes the proof of Theorem 2.3. 3 3 0 References [1] R. J. Duffin and A. C. Shaeffer, A class of nonharmonic Fourier series, Transactions of the American Mathematical Society, vol. 72, pp. 341-366, 1952. [2] I. Daubechies, A. Grossmann and Y. Meyer, Painless non-orthogonal expansions, Journal of Mathematical Physics, vol. 27, no. 5, pp. 1271-1283, 1986. [3] O. Christensen, An Introduction to Frames and Riesz Bases, Birkh¨auser, Boston, 2015. [4] L. Debnath and F. A. Shah, Wavelet Transforms and Their Applications, Birkh¨auser, New York, 2015. [5] A. Ron and Z.Shen, Weyl-Heisenberg frames and Riesz bases in L2(Rd), Duke Mathemat- ics Journal, vol. 89, pp. 237- 282, 1997. [6] K. Gro¨chenig, A. J. Janssen, N. Kaiblinger and GE. Pfander, Note on B-splines, wavelet scaling functions, and Gabor frames, IEEE Transactions and Information Theory, vol. 49, no. 12, pp. 3318-3320, 2003. [7] P. G. Casazza and O. Christensen, Weyl-Heisenberg frames for subspaces of L2(R), Pro- ceedings of American Mathematical Society, vol. 129, pp. 145-154, 2001. [8] K. Wang, Necessary and sufficient conditions for expansions of Gabor type, Analysis in Theory and Applications, vol. 22, pp. 155-171, 2006. [9] X. L. Shi and F. Chen, Necessary conditions for Gabor frames, Science in China : Series A. vol. 50, no. 2, pp. 276-284, 2007. [10] D. Li, G. Wu and X. Zhang, Two sufficient conditions in frequency domain for Gabor frames, Applied Mathematics Letters, vol. 24, pp. 506-511, 2011. [11] K. Gro¨chenig, Foundation of Time-Frequency Analysis, Birkh¨auser, Boston, 2001. [12] H. G. Feichtinger and T. Strohmer, Advances in Gabor Analysis, Birkh¨auser, Boston, 2003. [13] M. H. Taibleson, Fourier Analysis on Local Fields, Princeton University Press, Princeton, NJ, 1975. [14] D. Li and H. K. Jiang, Basic results Gabor frame on local fields, Chinese Annals of Mathematics: Series B, vol. 28, no. 2, 165-176, 2007.