Network-Level ISAC: An Analytical Study of Antenna Topologies Ranging From Massive to Cell-Free MIMO

A cooperative architecture is proposed for integrated sensing and communication (ISAC) networks, incorporating coordinated multi-point (CoMP) transmission along with multi-static sensing. We investigate how the allocation of antennas-to-base stations (BSs) affects cooperative sensing and cooperative communication performance. More explicitly, we balance the benefits of geographically concentrated antennas in the massive multiple input multiple output (MIMO) fashion, which enhance beamforming and coherent processing, against those of geographically distributed antennas towards cell-free transmission, which improve diversity and reduce service distances. Regarding sensing performance, we investigate three localization methods: angle-of-arrival (AOA)- based, time-of-flight (TOF)-based, and a hybrid approach combining both AOA and TOF measurements, for critically appraising their effects on ISAC network performance. Our analysis shows that in networks having <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i> ISAC nodes following a Poisson point process, the localization accuracy of TOF-based methods follows a ln<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i> scaling law (explicitly, the Cramér-Rao lower bound (CRLB) reduces with ln<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i>). The AOA-based methods follow a ln <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i> scaling law, while the hybrid methods scale as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</i> ln<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N+b</i> ln <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</i>, where <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</i> represent parameters related to TOF and AOA measurements, respectively. The difference between these scaling laws arises from the distinct ways in which measurement results are converted into the target location. Specifically, when converting AOA measurements to the target location, the localization error introduced during this conversion is inversely proportional to the distance between the BS and the target, leading to a more significant reduction in accuracy as the number of transceivers increases. In contrast, TOF-based localization avoids such distance dependent errors in the conversion process. In terms of communication performance, we derive a tractable expression for the communication data rate, considering various cooperative region sizes and antenna-to-BS allocation strategy. It is proved that higher path loss exponents favor distributed antenna allocation to reduce access distances, while lower exponents favor centralized antenna allocation to maximize beamforming gain. Simulations confirm that cooperative transmission and sensing in ISAC networks can effectively improve non-cooperative sensing and communication performance The proposed cooperative scheme shows superior performance improvement compared to centralized or distributed antenna allocation strategies.