Supports of representations in the Cohen class (Q657879)

This interesting paper establishes uncertainty principles for quadratic time-frequency representations of the Cohen class. The elements of the Cohen class are obtained by smoothing out the Wigner distribution \(W\), i.e., they are precisely those quadratic mappings \(Q_\sigma : \mathcal{S}(\mathbb{R}^d) \ni f \mapsto Q(f)\) satisfying \[ Q_\sigma(f) = \sigma \ast W(f), \] where \(W(f) : \mathbb{R}^{2d} \to \mathbb{C}\) denotes the Wigner distribution of \(f\). There is a qualitative uncertainty principle for the Wigner distribution stating that the support of \(W(f)\) is noncompact for any nonzero Schwartz function \(f\). The purpose of this paper is to establish this statement for large classes of convolution kernels \(\sigma\). Each of the following conditions turns out to be sufficient: (i) \(Q_\sigma\) fulfills Moyal's identity; (ii) \(\sigma\) is a compactly supported distribution such that the associated convolution operator is hypoelliptic; (iiii) the convolution operator associated to \(\sigma\) is a partial differential operator with constant coefficients.