103
edits
Changes
Jump to navigation
Jump to search
Line 63:
Line 63: − '''Notes'''+ Line 69:
Line 69: + − ''Diagram here''+ − − '''Other Resources'''
→Learn About Kalman Filtering & Sensor Fusion
[http://www.acfr.usyd.edu.au/pdfs/training/multiSensorDataFusion/dataFusionNotes.pdf Long, but well written intro to multi sensor fusion and Kalman filtering]
[http://www.acfr.usyd.edu.au/pdfs/training/multiSensorDataFusion/dataFusionNotes.pdf Long, but well written intro to multi sensor fusion and Kalman filtering]
===Notes===
The Kalman filter is linear quadratic estimator. We assume the system is linear, which should be good enough for our purposes. For non linear systems, do an extended Kalman filter. However, this takes much more computational resources and is much more unstable.
The Kalman filter is linear quadratic estimator. We assume the system is linear, which should be good enough for our purposes. For non linear systems, do an extended Kalman filter. However, this takes much more computational resources and is much more unstable.
A basic diagram of how this estimator works:
A basic diagram of how this estimator works:
[[File:Kalman_Filter_Estimator.png|center|400px]]
===Other Resources===
*More papers: http://www.tech-ex.com/article_images3/9/440249/1-11.pdf
*More papers: http://www.tech-ex.com/article_images3/9/440249/1-11.pdf
*“Welp Kalman filtering sometimes sucks” https://agrosy.informatik.uni-kl.de/fileadmin/Literatur/Blank10.pdf but I think we should use it anyway
*“Welp Kalman filtering sometimes sucks” https://agrosy.informatik.uni-kl.de/fileadmin/Literatur/Blank10.pdf but I think we should use it anyway