Finite-Time Robust Distributed Estimate for Nonlinear Systems With Heterogeneous Sensors
针对异构传感器网络中的离散时间随机非线性系统，提出一种有限时间分布式状态估计算法，通过交互多模型和共识融合实现鲁棒估计，并证明误差有界。
This article proposes a finite-time distributed state estimation (DSE) algorithm for discrete-time stochastic nonlinear systems with heterogeneous sensors. Considering the network with heterogeneous sensors, the distributed estimate framework is designed by three phases, namely, priori prediction, measurement update, and consensus fusion. To obtain the accurate priori prediction results, the interactive multiple model (IMM) method is adopted to calculate the priori state value in the priori prediction phase. By introducing the measurement probability matrix, a novel heterogeneous measurement information fusion algorithm is designed. Then the measurement information of each sensor is used to update the priori prediction estimates to calculate the estimate results in the measurement update phase. Based on the consensus method, the estimate results of each sensor are fused with consensus weight to calculate the distributed state estimates of nonlinear systems in the consensus fusion phase. Besides, with finite consensus fusion steps, the bounds of the proposed distributed estimate algorithm are proved to be existed. Finally, distributed state estimate simulation example for nonlinear system is set to validate the performance.