Update quaternion ekf

[eisoku9618]

• クォータニオンを正規化する部分でバグがあったので直しました
• ジャイロセンサの信頼度を表現できるようにしました
• （リファクタリングしました）

デフォルトでは何の影響もありませんので，よろしくお願いします．

[k-okada]

テストコードできるとうれしいんだけどな。
％ 企業はテストコードからつくるようですよ

◉ Kei Okada

On Tue, Feb 23, 2016 at 4:37 PM, Eisoku Kuroiwa [email protected]
wrote:

• クォータニオンを正規化する部分でバグがあったので直しました
• ジャイロセンサの信頼度を表現できるようにしました
• （リファクタリングしました）

デフォルトでは何の影響もありませんので，よろしくお願いします．

You can view, comment on, or merge this pull request online at:

#955
Commit Summary

• [KalmanFilter/EKFilter.h] fix bug : we should normalize only
  rotation quaternion
• [KalmanFilter/EKFilter.h] acceleration reference is to handle in
  KalmanFilter.cpp
• [KalmanFilter/EKFilter.h] normalize rotation quaternion as soon as
  possible
• [KalmanFilter/EKFilter.h] Q should be gyro noise covariance in order
  to make it easy to tune parameters
• [KalmanFilter/EKFilter.h] use initializer list at EKFilter
• [KalmanFilter/EKFilter.h] add a magic comment to use a 2 space
  indentation
• [KalmanFilter/EKFilter.h] remove unused old comments
• [KalmanFilter/EKFilter.h] use const member functions
• [KalmanFilter/EKFilter.h] update calcF for readability
• [KalmanFilter/EKFilter.h] do not pass a member variable to member
  functions
• [KalmanFilter/EKFilter.h] use const reference parameters
• [KalmanFilter/EKFilter.h] use hrpUtil to get Euler Angles from
  Rotation Matrix
• [KalmanFilter/EKFilter.h] use rotation quaternion to rotate
  coordinate instead of rotation matrix
• [KalmanFilter/EKFilter.h] clean up redundant codes
• [KalmanFilter/EKFilter.h] use reference instead of returning value
• [KalmanFilter/EKFilter.h] refectering

File Changes

• M rtc/KalmanFilter/EKFilter.h
  https://github.com/fkanehiro/hrpsys-base/pull/955/files#diff-0 (202)

Patch Links:

• https://github.com/fkanehiro/hrpsys-base/pull/955.patch
• https://github.com/fkanehiro/hrpsys-base/pull/955.diff

—
Reply to this email directly or view it on GitHub
#955.

[k-okada]

Refer to this link for build results (access rights to CI server needed):
http://jenkins.jsk.imi.i.u-tokyo.ac.jp:8080/job/hrpsys-qnx/2565/
Test PASSed.

[haudren]

Please define a matrix7d for clarity

[haudren]

Hello,
I made a few remarks above, I hope they will not come out as too blunt. Could you fix these style issues? They really hurt readability.

Otherwise there is maybe room for improvement and refactoring in the computation of the covariance matrix.

Thanks

[haudren]

Oh and also, It would be best to squash some commits before merging, so that the commits that are obviously just correcting some mistakes do not appear in the final tree (e.g. 182f8c9, 1d24bd7...) and only the ones adding functionality are preserved.

ご迷惑にかけてしますが、よろしくお願いします。

[eisoku9618]

@haudren Thank you for reviewing !
I've fixed all the issues you mentioned.

[k-okada]

Refer to this link for build results (access rights to CI server needed):
http://jenkins.jsk.imi.i.u-tokyo.ac.jp:8080/job/hrpsys-qnx/2566/
Test PASSed.