Exploration of Radar Cross Section Models and Distributed Sensing Techniques in JCAS Cell-free Massive MIMO

Abstract: Joint Communication and Sensing (JCAS) technology enables the sharing of infrastructure, resources, and signals between communication and sensing. However, studying the performance and algorithms using appropriate target reflectivity models for detection poses a significant challenge. Moreover, the increasing demand for efficient sensing systems in large-scale environments necessitates the study of distributed sensing for handling extensive data collection and processing. This study investigates the impact of target mobility on the choice between the Swerling-I and Swerling-II models for target reflectivity and proposes a brief method for reflectivity models in multi-static sensing. This method constructed a dedicated decorrelation area for a single radar detector using its decorrelation angle. Multiple radar system constructs an intersection of these areas. For targets expected to remain in this area, the Swerling-I model is preferred, while for targets likely to move to the outside intersection, the Swerling-II model is more suitable. Additionally, this thesis proposes and derives the test statistics for the distributed sensing in JCAS cell-free massive MIMO (multiple-input multiple-output) systems, where only the statistical distribution of transmitted signals is known at the receiver access points for the sensing purpose. This thesis compares the sensing performance of the proposed distributed processing with the centralized processing. Moreover, the results of a power allocation algorithm that maximizes sensing performance are compared against a baseline algorithm that minimizes total power consumption. In terms of sensing performance via guaranteeing the quality of service of the communication, the results indicate that the sensing algorithm consistently outperforms the power-minimizing algorithm, regardless of the selected reflectivity model. Furthermore, the Swerling-II model performs relatively worse when the reflectivity of the target is low, but exhibits improved performance on a relatively high reflectivity target. Regarding distributed sensing, its implementation may lead to a deterioration in sensing performance. However, the results show that distributed sensing can approach the performance of centralized sensing when the target has high reflectivity. The major advantage of distributed sensing is the reduced fronthaul signaling load in a JCAS cell-free massive MIMO system.