PAC-ZNN for Robust Target Tracking in WSNs Against Complex Polynomial Noise

In wireless sensor networks (WSNs), angle of arrival (AOA) and time difference of arrival (TDOA) localization systems relying on distributed sensor measurements degrade significantly under high-order time-varying noise. Although traditional zeroing neural networks (ZNNs) handle dynamic localization tasks, their insufficient robustness against such high-order noise often compromises convergence stability and accuracy. To address this limitation, this paper proposes a polynomial anti-noise compensation ZNN (PAC-ZNN) incorporating a polynomial anti-noise compensation (PAC) term and a logarithmic mapping activation function (LMAF). Specifically, the PAC term mitigates the adverse effects of cumulative high-order noise, while the LMAF further enhances the convergence speed and stability of the system. The global convergence and robustness of the proposed PAC-ZNN are rigorously proven based on Lyapunov stability theory. Simulation results demonstrate that when applied to AOA and TDOA-based dynamic localization tasks, the proposed PAC-ZNN outperforms traditional ZNN-based solutions in terms of anti-noise capability, convergence efficiency, and localization precision under high-order noise conditions. Furthermore, it maintains robust tracking performance even under complex multipath environments, verifying its superior performance and practical application value.