Target Parameter Estimation with Deep Unfolding Networks for MIMO-OFDM Based Integrated Sensing and Communication Systems

The multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) technology shows great potential in integrated sensing and communication (ISAC) systems. To address the challenges posed by the high computational complexity and low estimation accuracy of traditional target parameter estimation methods in a MIMO-OFDM based ISAC system, this paper introduces a compressed sensing based approach for target parameter estimation. We first derive the channel model within the MIMO-OFDM system, and then reformulate the problem of target parameter estimation as a least absolute shrinkage and selection operator (LASSO) problem. Furthermore, we propose a model-driven network, termed FISTA-Net, which deeply unfolds the fast iterative shrinkage thresholding algorithm (FISTA) into a deep neural network to solve the LASSO problem. Experimental results show that the proposed FISTA-Net algorithm outperforms the existing methods in terms of estimation accuracy and computation efficiency for MIMO-OFDM based target parameter estimation.