事件(粒子物理)
估计员
协方差
卡尔曼滤波器
计算机科学
数学优化
高斯分布
数学
算法
统计
人工智能
量子力学
物理
作者
Hao Yu,Jun Shang,Tongwen Chen
出处
期刊:Automatica
[Elsevier]
日期:2021-01-01
卷期号:123: 109314-109314
被引量:20
标识
DOI:10.1016/j.automatica.2020.109314
摘要
This paper investigates the problem of remote event-based state estimation for a linear discrete-time plant. Both stochastic and deterministic event-based transmission policies are considered for the systems implemented with smart sensors, where local Kalman filters are embedded. Based on the concept of generalized closed skew normal distributions, the exact probability density functions of the remote event-based state estimation processes are provided. With the properties of smart sensors, the explicit form of the remote event-based state estimators can be derived, without involving numerical integration. In addition, in the case of scalar plants, the estimation and transmission performances under different kinds of event-based scheduling policies are compared theoretically. An important inequality on a truncated covariance of some particular multivariate Gaussian distribution is proved, which builds a bridge between performances of the two event-based policies. Based on this inequality, it is proved that for any considered stochastic event-based transmission policy, there always exists a deterministic counterpart that leads to better estimation performance using the same communication and computational resources. Numerical simulations are provided to illustrate the theoretical results.
科研通智能强力驱动
Strongly Powered by AbleSci AI