Applications of nonlinear estimation techniques

The purpose of this paper is to provide an analysis on using two non‐conventional nonlinear estimating filters compared to the traditional linearized extended Kalman filter (EKF). This analysis will look at two state‐of‐the‐art applications and will provide insight to the problems associated with these applications.

The approach taken was to simulate both applications with three different filter designs: EKF, unscented Kalman filter, and particle filter. After results and explanations are given for both applications, then there is a comparison of results between the two applications to compare and contrast their findings.

This research shows how critical it is when selecting a filter for different applications. Not only is tuning the filter properly a necessity, but choosing a filter that is optimum for the application also greatly affects the accuracy and precision of the results.

As these filter methods are proven feasible for these applications, testing can move beyond simulation. Further research could compare other nonlinear filters to these results to determine if a better estimation technique exists.

This paper shows a lot of the issues one must face when choosing an estimation technique for their application as well as the impact the technique can have on the outcome.

This paper clearly describes the decision‐making criteria in regards to these two specific applications. These two applications are current technological problems that many are trying to solve. This paper shows where and why errors in calculations occur. It also offers insight into different ways to solve these problems when the specific application is taken into account.