We study spatially structured stochastic networks coupling queueing communication dynamics with filtering-based state estimation. Network nodes form a stationary homogeneous Poisson point process $\Phi \subset \mathbb{R}^2$. Around each node, secondary agents evolve through local random motions, producing a dynamic marked point process. Shared wireless medium generates shot-noise interference, so communication and sensing performance depend on the same random field. Communication is modelled by spatially indexed queues whose service rates are monotone functionals of the instantaneous SINR. Sensing is described by a partially observed stochastic process whose observation noise covariance depends on the same interference field, leading to state-dependent filtering.

This setting creates a coupling between a queueing network in random environment and stochastic estimators driven by spatial shot noise. We define system-level performance metrics under the Palm distribution of $\Phi$. Our main result proves an association property between communication and sensing functionals at the typical node. Under general shot-noise interference models, the queue workload process and filtering error process are associated in the sense of increasing functionals. The proof uses coupling arguments, stochastic monotonicity, and comparison techniques for interacting particle systems in random environments.

The results indicate that regimes improving communication performance also enhance sensing accuracy. More broadly, the framework offers a probabilistic basis for analysing spatial networks with coupled service and estimation mechanisms, showing how stochastic geometry, interacting particle systems, and queueing theory can jointly address large-scale systems where geometry and flow dynamics are intertwined.