Tracking and Data Fusion in Joint Sensing and Communication Networks

In this paper, we consider a beyond-5G multiple-input multiple-output (MIMO) joint sensing and communication (JSC) system where the base stations (BSs) act as monostatic radars by exploiting the signal transmitted to the user equipments (UEs) through a multi-beam radiation pattern. This generates a trade-off between the two functionalities, which should be investigated. This work aims to show the benefits of tracking algorithms to the root mean squared error (RMSE) of an object’s position estimation by reserving only a small fraction of the transmitted power for sensing. First, we compare the performance obtained with several tracking algorithms, such as the cubature Kalman filter (CKF), Gaussian mixture cardinalized probability hypothesis density (GMCPHD) and particle filter (PF), by varying the radar cross-section (RCS) of the object and the power devoted to sensing. Then, we consider a scenario where multiple monostatic JSC systems cooperate to improve the target position estimate accuracy via data fusion performed by tracking algorithms. Numerical results show that all tracking methods improve sensing performance under typical wireless communication scenarios and that cooperative sensing through data fusion boosts the whole system’s performance significantly, allowing the network designer to save resources for communication.