Evaluation of Unscented Kalman Filter and Extended Kalman Filter for Radar Tracking Data Filtering

With the increasing focus on green energy, wind turbines (WTs) have become common occurrences in most landscapes. The presence of WTs in the field of view of a radar will create very complicated clutter in a received signal. One of the major reasons for the complicated nature of WT clutter is the fact that it will consist of a wide band of Doppler frequencies. In addition, the Doppler band of frequencies keep changing based on wind speed and direction The tracking of dominant Doppler frequencies is one of the most important steps in the process of filtering WT clutter. In this work, we present an unscented Kalman filter (UKF) based solution to track the dominant Doppler frequencies in WT echoes. We have shown the efficiency of our algorithm by applying it to a wide range of real-measured data collected from various locations in Europe.INDEX TERMS Adaptive Kalman filtering, Adaptive signal processing, Doppler effect, filter, radar I. INTRODUCTION

show abstract

“…The discussion of numerical complexity and accuracy of the different versions of KF can be found e.g. in [43]- [45].…”

Section: B Dsc Prediction Using Ukfmentioning

confidence: 99%

Wind Turbine Micro-Doppler Prediction Using Unscented Kalman Filter

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Taking advantage of UKF, significant attention has been placed in applying the algorithm for USVs' navigational purposes with GPS/IMU. [4,6,9,10]. Other equally notable ways like Particle filter is ruled out of interest in limiting computational burden.…”

Section: B Literature Reviewmentioning

confidence: 99%

Adaptive Localisation for Unmanned Surface Vehicles Using IMU-Interacting Multiple Model

Gunawan

Liu

2020

2020 International Conference on System Science and Engineering (ICSSE)

View full text Add to dashboard Cite

Unscented Kalman Filter (UKF) remains to be a prevalent multi-sensor fusion method in many practices, including navigational tracking for Unmanned Surface Vehicles (USVs). This paper suggests that results from UKF fusion is unsatisfactory for USVs' relatively smooth path due to UKF's lack of versatility. Hence, it is proposed here that by replacing the UKF with Interacting Multiple Model (IMM), estimation results will better represent USV's movement. Furthermore, this paper proposes slight modification to the IMM in order to heighten the algorithm's confidence in switching modes. By exploiting angular velocity information from Inertial Measurement Unit (IMU), an independent mode probability can be obtained which is then injected into the IMM. Computer simulations based on maritime operations were done to show that the proposed IMU-based IMM is able to react faster to mode changes, giving more reliable outcomes.

show abstract

“…ML is in general very powerful and has been applied for different optimization purposes in nanophotonics, like numerical device modelling 40,41 , reverse engineering of resonators, data analysis. It can even be further enhanced when combined with artificial intelligence for automatic ML technologies [42][43][44][45][46][47][48][49] .…”

Section: Introductionmentioning

confidence: 99%

“…It can even be further enhanced when combined with artificial intelligence for automatic ML technologies. [42][43][44][45][46][47][48][49] Here we apply ML-based image processing to improve the capabilities of in situ EBL. We demonstrate the advantages of ML-enhanced in situ EBL by comparative studies on the deterministic integration of InGaAs QDs into circular Bragg grating (CBG) resonators.…”

Section: Introductionmentioning

confidence: 99%