yuzhou42 / eskf-attitude-estimation Goto Github PK

Dear NicoChou.

Thank you for the great matlab code for EKF_AHRS.
I have been studying EKF_ARHS, and found your code.

https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m

I seems to me that your matlab code is working correctly,
However, I have difficulty in interpreting your matlab code.
(Even though I am familiar with kalman filter)

I have following questions regarding following equations,
Is there any supporting material which describe the mathmetical equation for following sections as you did for error-state EKF?

https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L38
https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L41
https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L44

https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L78-L81

https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L87-L94

https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L117-L119

Hello，do you study MSCKF? I am learning the msckf and maybe we can have a discussion.Thank you very much!

NicoChou您好

您在误差重置过程中将误差置为零，
https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/df1152ab425aa22d4801c849742b86e3deeb6ec6/ESKF_AHRS.m#L59-L60
那
https://github.com/NicoChou/ESKF-Attitude-Estimation/blob/df1152ab425aa22d4801c849742b86e3deeb6ec6/propagateErrorStateAndCovar.m#L4-L7
这段代码进行误差传递有什么用呢

NicoChou 您好！

   EKF_AHRS.m文件中的113-115行,磁力计模型的量测矩阵时参考哪个文献得到的呀,怎么跟我查到的不大一样，期望得到您的回复。

yuzhou, 你好!

针对代码有下面两个疑问, 期待得到回复.

雅可比Ak中的角速度为什么没有减掉偏移量x(k,5~7)呢?
https://github.com/yuzhou42/ESKF-Attitude-Estimation/blob/master/EKF_AHRS.m#L88-L91

这里计算detZ时, 6个'预计观测' 的前面是否应该加上负号?
例如: - 2*(q(2)*q(4) - q(1)*q(3))
https://github.com/yuzhou42/ESKF-Attitude-Estimation/blob/master/calH.m#L28-L34

谢谢!

Hi, I was wondering about the unit of input data, mainly the Accelerometer, Gyroscope, and Magnetometer, for EKF_AHRS? besides, a number, 8.3, is used in yaw-related calculation, what's the meaning of this number? thanks a lot!

时间过去这么久，不知道你还关不关心这个问题。我发现你的过程协方差Q矩阵是有点问题的。
应该是：blkdiag( sigema_wn^2dt^2 sigema_wbn^2dt )
但是你写的是： blkdiag( sigema_wn^2dt sigema_wbn^2dt )

这样写误差会更小些。