Distributed Moving Horizon Estimation Over Sporadically Observing Sensor Networks: An L-Step Approach With Stability Guarantees

Distributed state estimation is essential for modern industrial systems, especially in sensor networks and multi-agent frameworks constrained by limited communication and computational resources. This article presents a novel ℓ-step Distributed Moving Horizon Estimation (DMHEℓ) algorithm that explicitly addresses sporadic measurements. The algorithm employs an information diffusion mechanism, enabling each sensor to leverage measurements from its ℓ-step neighbors. A comprehensive stability analysis proves that the estimation error remains uniformly bounded, thereby establishing convergence guarantees, regardless of the sporadic nature of the measurements. The effectiveness of DMHEℓ is demonstrated through simulations in a realistic industrial scenario involving an autonomous mobile robot. The results confirm that the method achieves high estimation accuracy while requiring significantly less communication and data exchange than existing distributed estimators for comparable accuracy levels.