vel_update

[YSatShinshu]

Dear NaveGo-Team, thank you for your very helpful work.

I have a question about function "vel_update".

Why do you not multiply the Coriolis acceleration by the velocity in your velocity update formula?

Here is the location (vel_update.m, line 44-46)

coriolis = (omega_en_n + 2 * omega_ie_n); % Coriolis

fn_c = fn - coriolis + gn; % Corrected specific force in nav-frame

I'm sorry for my lack of study, but it seems to be multiplied in the reference, so please tell me.

Thank you.

[rodralez]

Hello,

Thank you for pointing out this BIG bug.

I had to modified the following functions:

ins-gnss/ins_gnss.m
ins/coriolis.m
ins/earth_rate.m
ins/transport_rate.m
ins/vel_update.m
simulation/acc_gen.m
simulation/gyro_gen.m

The transport and Earth rates were incorrectly defined in NaveGo as well.

Since these turn rates and Corolioris can be neglected in navigation systems that are based on tactical-grade sensors and inferior categories, NaveGo's examples are unaffected by these changes.

Kind regards.

[YSatShinshu]

As far as I can confirm, there is no problem.

On a different note, NaveGo uses the previous value for the speed multiplied by Coriolis. In the examples I know, it uses the velocity derived by interpolation or extrapolation, and I'm not sure why. Do you think it is OK to use the previous value? Please let me know your opinion.

[rodralez]

I have tested velocity interpolation for updating the velocity (vel_update.m) but NaveGo performance does not change, at least using tactical- and automotive-grade IMUs.

[venkat057]

As far as I can confirm, there is no problem.

On a different note, NaveGo uses the previous value for the speed multiplied by Coriolis. In the examples I know, it uses the velocity derived by interpolation or extrapolation, and I'm not sure why. Do you think it is OK to use the previous value? Please let me know your opinion.

Hi,
Velocity derived by interpolation or extrapolation gives better accuracy only when vehicle dynamics are high like vehicle speed is high. Since in real-time by the time kf compute the position, velocity the vehicle is already moved to a new position, velocity. So the kf outputs are delayed/old by kf execution time.
Can you send me link of this Velocity derived by interpolation or extrapolation example.

[YSatShinshu]

Hello,

Sorry, because the example I have is a paid one, I send you a link to the sales page. I use "INS Toolbox 3.0@MATLAB"
GPsoft LCC.
https://gpsoftnav.com/products/

Velocity derived by interpolation or extrapolation gives better accuracy only when vehicle dynamics are high like vehicle speed is high. Since in real-time by the time kf compute the position, velocity the vehicle is already moved to a new position, velocity. So the kf outputs are delayed/old by kf execution time.

Thanks for giving me advice. Is this generally known? If you know any references, please tell me.

Best regards.

[rodralez]

Velocity update code is based on Eq. 5.53 and 5.54, page 179, from Paul D. Groves, Principles of GNSS, Inertial, and Multisensor Integrated Navigation Systems. Second Edition. I have realized that previous gravity, previous Transport rate and previous Earth rate should be used instead of current values. This error has to be fixed.