A recursive procedure to obtain a class of orthogonal polynomial wavelets (Q2479463)

The authors discuss a multiresolution analysis scheme in \(L^{2}( a,b) \) for an interval \(( a,b) \) introduced by \textit{M. A. Skopina} [Sb. Math. 192, No. 3, 433--454 (2001; Zbl 1021.42026)]. This scheme is based on a Meyer-type wavelet \(\psi \) in \(L^{2}({\mathbb R}) ,\) with corresponding scaling function \(\varphi \), and a sequence of orthogonal polynomials \(p_{n},n\in{\mathbb N}_{0},\) over the interval \( ( a,b)\). The scaling spaces \(V_{j},\) \(j\in {\mathbb N}\) are the linear spaces generated by \[ R_{-j,k} :=\sum_{l\in Z}\widehat{\varphi }_{-j,k}( 2\pi l) P_{\sigma ( l) } \tag{1} \] for \(k=0,\dots,2^{j}-1,\) where \(\widehat{\varphi }_{j,k}\) is as usual the function defined by \( 2^{-j/2}\widehat{\varphi }( 2^{-j}x-k) ;\) the sum is actually finite since \(\widehat{\varphi }\) has compact support. Here \( P_{\sigma ( l) }\) are complex-valued polynomials constructed from the orthogonal polynomials \(p_{n}\) in the following way: define \(\sigma ( l) =2l\) for \(l\geq 0\) and \(\sigma ( l) =2| l| -1\) for \(l<0,\) and set \[ P_{\sigma ( l) }= \frac{p_{\sigma ( l) }( x) }{k_{\sigma ( l) }}+( -1) ^{\text{sign}( l) }i \frac{p_{\sigma ( l) -1}( x) } {k_{\sigma ( l) -1}} \] where \(\text{sign}( l) \) is the signum of \(l\) (with \(sign( 0) =0)\) and \(k_{n}\) is the \(L^{2}\)-norm of \(p_{n}.\) Wavelet spaces are defined in a similar way, generated by functions \(S_{-j,k}\) , \(k=0,\dots,2^{j}-1,\) by replacing the coefficients \(\widehat{\varphi }_{-j,k}( 2\pi l) \) in formula (1) by \(\widehat{\psi }_{-j,k}( 2\pi l) \). The main results in the paper are two-scale relations and decomposition and reconstruction formula for the Skopina system. Finally conjugation and zero moments properties are discussed.