Perception aware priority augmentation for age based scheduling in capacity limited cognitive radar networks

Abstract Capacity-limited cognitive radar networks require efficient coordination between sensing and communication for timely multi-target tracking. Existing Age of Information (AoI) and Age of Incorrect Information (AoII) scheduling frameworks assume uniform sensing reliability, leading to inefficient uplink resource use under degraded perception. This work proposes a perception aware augmentation of existing age-based priority indices, introducing calibrated CNN and LSTM reliability scores as a multiplicative weight on the predicted information-gain term, leaving AoI/AoII dynamics, capacity constraints, and IMM/Kalman filtering unchanged. This reshapes update ranking to jointly reflect timeliness and sensing reliability, without altering sensing, motion, filtering, or AoI/AoII evolution dynamics. The framework supports both centralized scheduling and distributed AoII-based triggering, preserving per-CPI capacity and long-run rate constraints. Under a simulated CRN with synthetic radar measurements and Poisson-deployed nodes, results show a 26.7% reduction in median fused position error against the strongest age-based baseline (Distributed-AoII), a 27.4% reduction in steady-state mean track age, and a 26.2% improvement in missed-target clearance, all under identical bandwidth constraints. Ablation studies confirm that gains arise solely from confidence-modulated prioritization.