Parameter Estimation of Multipath MIMO Channels in ISAC Systems Using Canonical Polyadic Decomposition

This work addresses joint communication and sensing in MIMO systems, where the received signal is modeled as a combination of multiple propagation paths. A Mode-Tensorized Canonical Polyadic Decomposition (MT-CPD) framework is proposed for estimating both the channel parameters and the characteristics of the observed targets. The method relies on representing the channel in a multidimensional tensor form by applying a mode tensorization procedure that introduces additional virtual modes, improving the separability of signal components. The proposed MT-CPD approach enables accurate extraction of spatio-temporal features of the MIMO channel and demonstrates enhanced robustness under low signal-to-noise ratio conditions. Simulation results indicate that MT-CPD outperforms conventional Canonical Polyadic Decomposition (CPD) in terms of estimation accuracy while maintaining low computational complexity, making it a promising candidate for next-generation Integrated Sensing and Communication (ISAC) systems, including 6G applications.