Information Fusion Filtering Scheme for Multisensor Systems With Dual-Unknown Inputs

An estimating technique for multi-sensor systems with dual-unknown inputs is presented in this work. The approach separates the estimate procedures of system states and unknown inputs, carrying out state estimation and estimation of the dual-unknown inputs individually for multi-sensor systems with different unknown inputs in the state and observation equations. The state estimates from individual sensors are fused using the Linear Minimal Variance Information Fusion Criterion (LMVIFC) to produce linearly minimal variance unbiased state estimates after local filtering algorithms are created for each sensor. In order to guarantee the estimates’ objectivity and minimal variation, each sensor simultaneously estimates the unknown inputs in its observation equation independently. The associated estimates are then fused using LMVIFC. Lyapunov stability theory is used to theoretically demonstrate the filter’s stability under dual-unknown inputs. Lastly, Monte Carlo simulations are used to assess and contrast the suggested approach with other conventional techniques. The outcomes demonstrate that even in complex scenarios with two distinct kinds of unknown inputs, the suggested approach can precisely and consistently track system states. In a hypersonic aircraft control system, its advantage is further confirmed.