Scholarlink

<p>The robust state estimation methods for nonlinear systems with Tobit-II type censored measurements and time-varying uncertainty parameters are proposed. Firstly, the state equation with model parameters uncertainties and the Tobit-II type censored measurement equation are reconstructed for nonlinear systems. Secondly, within the Bayesian framework, the state prediction and latent measurement prediction are computed by cubature sampling and lattice sampling, respectively. Then, the Square Root Cubature Tobit Kalman Filter (RCTKF) and the Square Root Lattice Tobit Kalman Filter (RLTKF) are proposed by <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(QR\)</EquationSource> </InlineEquation> decomposition of covariance matrix. Furthermore, the Square Root Lattice Smooth Variable Structure Tobit Filter (RLSVSTF) and the Square Root Cubature Smooth Variable Structure Tobit Filter (RCSVSTF) are further presented for model parameter uncertainties and censored measurements based on the Smooth Variable Structure Filter (SVSF). Finally, the proposed algorithms are validated in multiple scenarios, including a UAV cruising flight scenario and the Udacity real radar dataset, which demonstrating superior estimation accuracy and enhanced robustness.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Robust State Estimation for Nonlinear Systems with Censored Measurements and Model Parameters Uncertainties

  • 期刊论文

  • 发表日期：

    2025年7月17日

查看全文
  • Yuzhao Jiao,
  • Shuai Chang,
  • Zhiwu Chen,
  • Taishan Lou,
  • Xuetao Li,
  • Caoping Niu

摘要

The robust state estimation methods for nonlinear systems with Tobit-II type censored measurements and time-varying uncertainty parameters are proposed. Firstly, the state equation with model parameters uncertainties and the Tobit-II type censored measurement equation are reconstructed for nonlinear systems. Secondly, within the Bayesian framework, the state prediction and latent measurement prediction are computed by cubature sampling and lattice sampling, respectively. Then, the Square Root Cubature Tobit Kalman Filter (RCTKF) and the Square Root Lattice Tobit Kalman Filter (RLTKF) are proposed by \(QR\) decomposition of covariance matrix. Furthermore, the Square Root Lattice Smooth Variable Structure Tobit Filter (RLSVSTF) and the Square Root Cubature Smooth Variable Structure Tobit Filter (RCSVSTF) are further presented for model parameter uncertainties and censored measurements based on the Smooth Variable Structure Filter (SVSF). Finally, the proposed algorithms are validated in multiple scenarios, including a UAV cruising flight scenario and the Udacity real radar dataset, which demonstrating superior estimation accuracy and enhanced robustness.