search

wvu-navLab / RobustGNSS

Robust GNSS Processing With Factor Graphs
https://www.researchgate.net/publication/320084321_Robust_Navigation_in_GNSS_Degraded_Environment_Using_Graph_Optimization
MIT License
126 stars 44 forks source link

Computation efficiency #5

Open albertchang1989 opened 4 years ago

albertchang1989 commented 4 years ago

The work is really impressive.

Do you compare the computation efficiency between the traditional EKF and your new proposed method?

watsonryan commented 4 years ago

Hi @albertchang1989 ,

For the software release in this repo., we have not conducted a comparison between the implemented approaches and an EKF. This is primarily because the implementations contained in this repo. are batch estimators, and we do not know of a fair way to conduct an efficiency comparison between an incremental (e.g., an EKF) and a batch estimator.

However, we have re-implemented all of the robust estimators in this repo. in an incremental nature through the utilization of the ISAM2 [1] algorithm in the following repo: https://github.com/wvu-navLab/ICE. Utilizing the software in this repo. a comparison could be conducted between the implemented approaches and an EKF. We have not done this comparison yet. If you were to use the provided software to conduct this computation efficiency comparison, we would definitely like to hear the results.

Ryan

[1] Kaess, Michael, et al. "iSAM2: Incremental smoothing and mapping using the Bayes tree." The International Journal of Robotics Research 31.2 (2012): 216-235.

albertchang1989 commented 4 years ago

Hi @albertchang1989 ,

For the software release in this repo., we have not conducted a comparison between the implemented approaches and an EKF. This is primarily because the implementations contained in this repo. are batch estimators, and we do not know of a fair way to conduct an efficiency comparison between an incremental (e.g., an EKF) and a batch estimator.

However, we have re-implemented all of the robust estimators in this repo. in an incremental nature through the utilization of the ISAM2 [1] algorithm in the following repo: https://github.com/wvu-navLab/ICE. Utilizing the software in this repo. a comparison could be conducted between the implemented approaches and an EKF. We have not done this comparison yet. If you were to use the provided software to conduct this computation efficiency comparison, we would definitely like to hear the results.

Ryan

[1] Kaess, Michael, et al. "iSAM2: Incremental smoothing and mapping using the Bayes tree." The International Journal of Robotics Research 31.2 (2012): 216-235.

thanks a lot！