1: \begin{abstract}

2: %The multidimensional signals refer to the signal jointly from multiple dimensions, e.g., spatial, temporal, and frequency. Such signals inherit information from multiple dimensions and can be used to estimate unknown parameters in these dimensions, and generally involve the Kronecker product. 

3: This work studies the problem of jointly estimating unknown parameters from Kronecker-structured multidimensional signals, which arises in applications like intelligent reflecting surface (IRS)-aided channel estimation. Exploiting the Kronecker structure, we decompose the estimation problem into smaller, independent subproblems across each dimension. Each subproblem is posed as a sparse recovery problem using basis expansion and solved using a novel off-grid sparse Bayesian learning (SBL)-based algorithm. Additionally, we derive probabilistic error bounds for the decomposition, quantify its denoising effect, and provide convergence analysis for off-grid SBL. Our simulations show that applying the algorithm to IRS-aided channel estimation improves accuracy and runtime compared to state-of-the-art methods through the low-complexity and denoising benefits of the decomposition step and the high-resolution estimation capabilities of off-grid SBL.

4: %Our simulation results evaluate the decomposition step and off-grid SBL separately, highlighting the computational efficiency and denoising benefits of the decomposition step and the high-resolution estimation capabilities of off-grid SBL. Finally, applying our algorithm to IRS-aided channel estimation demonstrates improved accuracy and reduced runtime compared to the state-of-the-art methods, driven by the combined effects of decomposition and grid optimization. 

5: \end{abstract}

6: