Robust Fusion Estimation Under Data-Driven Transmission Strategy for Multisensor Systems With Random Packet Drops

One of the critical challenges in multisensor fusion for vehicle attitude estimation is dealing with process and measurement noise, model uncertainty, and network channel failures. This article proposes a robust fusion estimator for multisensor systems (MSSs) based on constrained communication rates and random packet drops to address these issues. A robust state estimator (RSE) based on sensitivity penalty is used for local estimation. The estimation error pseudo-cross-covariance matrix of the fusion center, considering both the transmission strategy and packet drops, is derived. The fusion estimator employs a data-driven transmission strategy that guarantees comparatively good estimation performance while effectively reducing sensor energy loss and network congestion. It investigated the conditions under which the estimation error of the fusion estimator is uniform boundedness. The simulation is performed with a tractor-car system, and the effectiveness of the derived method is verified by fusion estimation in different cases.