Real-time multi-agent position coordination in the presence of noise using a robust zeroing neural dynamics model

Multi-agent systems often encounter noise interference during real-time position coordination in complex environments. Traditional zeroing neural dynamics (ZND) models exhibit insufficient robustness when handling linear and quadratic noise. To address these challenges, this paper proposes a novel double-integral enhanced zeroing neural dynamics (NDIEZND) model. Rigorous theoretical analysis establishes that the proposed model ensures globally stable convergence under noisy conditions. Moreover, results from two sets of simulations and physical experiments show that the NDIEZND model achieves a convergence accuracy of 10⁻⁴ in position coordination tasks and significantly outperforms existing comparative models when handling linear noise, demonstrating its fast convergence and strong robustness.