Computer Science · Artificial Intelligence
A Novel Hybrid Consensus-based High-degree Cubature Information Filter
Jun Liu, Yu Liu, Kai Dong, Ziran Ding, Qichao Li
See on the mapAbstract
In this note, the problem of distributed nonlinear state estimation with networked sensors is considered. Based on statistical linear regression, a more suitable strategy to compute the measurement information contribution of the fifth-degree cubature information filter is presented. Based on the new strategy for measurement update and hybrid consensus on both information and measurements, a novel hybrid consensus-based high-degree cubature information filter (HCHCIF) is proposed. The effectiveness and superiority of the proposed HCHCIF is validated by tracing a maneuvering target in a sparse sensor network with naive nodes. Simulation results show that the proposed HCHCIF outperforms the existing distributed algorithms in the aspects of estimation accuracy and consensus on estimates from different nodes.