What is it about?

Estimation in nonlinear systems is very important in various areas such as signal processing, robotic control and so on, because almost all systems are inherently nonlinear. We considered a innovative method for the nonlinear filter with constant filter gains.

Featured Image

Why is it important?

The extended Kalman filter (EKF) is probably the most widely used estimation algorithm because of simplicity and ease of implementation for nonlinear systems. However, the EKF is only reliable for systems that are almost linear in state space because it's simply based on a first order Taylor expansion and substitutes Jacobian matrices for the Kalman filter equations. Moreover, there is no universal steady gain in the EKF which needs to calculate the differential Ricatti equation online. For the applications with limited computational power, the update of the filter gain at each step will be costly. To overcome these problems, we considered designing a nonlinear filter using a innovative method.


In this research, I applied the method to an estimation problem of an electric power system. I hope to apply the method to more real nonlinear systems.

Kazuo Komatsu
National Institute of Technology, Kumamoto College

Read the Original

This page is a summary of: A Nonlinear Filter with Constant‐Gains Using a Pseudo‐Formal Linearization Method Based on Chebyshev Expansion, IEEJ Transactions on Electrical and Electronic Engineering, May 2022, Wiley,
DOI: 10.1002/tee.23624.
You can read the full text:



The following have contributed to this page