Learning Target Dynamics While Tracking Using Gaussian Processes

Tracked targets often exhibit common behaviors due to influences from the surrounding environment, such as wind or obstacles, which are usually modeled as noise. Here, these influences are modeled using sparse Gaussian processes that are learned online together with the state inference using an exte...

Full description

Bibliographic Details
Published in:IEEE Transactions on Aerospace and Electronic Systems
Main Authors: Veibäck, Clas, Olofsson, Jonatan, Lauknes, Tom Rune, Hendeby, Gustaf
Format: Article in Journal/Newspaper
Language:English
Published: Linköpings universitet, Reglerteknik 2020
Subjects:
Gaussian processes
Computational modeling
Adaptation models
Target tracking
Estimation
Time-varying systems
Extended Kalman filter (EKF)
identification
online learning
sparse Gaussian process (GP)
Control Engineering
Reglerteknik
Norway
Tromso
Sea ice
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-168871
https://doi.org/10.1109/TAES.2019.2948699
Description
Summary:Tracked targets often exhibit common behaviors due to influences from the surrounding environment, such as wind or obstacles, which are usually modeled as noise. Here, these influences are modeled using sparse Gaussian processes that are learned online together with the state inference using an extended Kalman filter. The method can also be applied to time-varying influences and identify simple dynamic systems. The method is evaluated with promising results in a simulation and a real-world application. Funding Agencies|Vinnova Industry Excellence Center LINK-SICVinnova; European Union Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie GrantEuropean Union (EU) [642153]; Research Council of Norway through the Centres of Excellence funding scheme [223254-NTNU-AMOS]; CENIIT program at Linkoping University [17:12]; FRAM Centre, Tromso, Norway, through the Project "Ground-based radar measurements of sea-ice, icebergs, and growlers"

Similar Items