Checking potencial problem in function F_update.m, relates to #101

[rodralez]

F_update.m
The state transition matrix in the program is 15 by 15, which is different from the original paper,this in itself has nothing to say.

In the program： F = [F11 F12 F13 DCMbn Z ; F21 F22 F23 Z -DCMbn ; F31 F32 F33 Z Z ; Z Z Z Fgg Z ; Z Z Z Z Faa ; ];

G = [DCMbn Z Z Z ; Z -DCMbn Z Z ; Z Z Z Z ; Z Z Fbg Z ; Z Z Z Fba ; ];

why not:

F = [F11 F12 F13 -DCMbn Z ; F21 F22 F23 Z DCMbn ; F31 F32 F33 Z Z ; Z Z Z Fgg Z ; Z Z Z Z Faa ; ];

G = [-DCMbn Z Z Z ; Z DCMbn Z Z ; Z Z Z Z ; Z Z Fbg Z ; Z Z Z Fba ; ];

Only in this way it is reasonable. Is there any special treatment or transformation? It is also the latter rather than the former in corresponding paper about NaveGo.

[rodralez]

This is an outdated version of the function. Please, check the last version.

[scott198510]

@rodralez The last version is :

F = [F11 F12 F13 DCMbn Z ; F21 F22 F23 Z DCMbn ; F31 F32 F33 Z Z ; Z Z Z Fgg Z ; Z Z Z Z Faa ; ];

G = [DCMbn Z Z Z ; Z DCMbn Z Z ; Z Z Z Z ; Z Z Fbg Z ; Z Z Z Fba ; ];

This is not right, no minus sign in the last verison.

The right one is :

F = [F11 F12 F13 -DCMbn Z ; F21 F22 F23 Z DCMbn ; F31 F32 F33 Z Z ; Z Z Z Fgg Z ; Z Z Z Z Faa ; ];

G = [-DCMbn Z Z Z ; Z DCMbn Z Z ; Z Z Z Z ; Z Z Fbg Z ; Z Z Z Fba ; ];

[rodralez]

I had to add again a negative sing to the 2do DCMbn in matrices F and G because MPU-6000 example stopped working. If you try this example with your proposed matrices F and G, the INS/GNSS diverges badly.

Please, try the develop branch.

[scott198510]

@rodralez Can you tell me some explanations?

I have read the books as you have referred to.

Both of them about F & G are different as your code.

[rodralez]

In this case, Farrell's model shows to be working. You may compare both models, Farrell's and Groves', and thus determine what's going on.

[scott198510]

@paolozoccarato If you try this example with your proposed matrices F and G, the INS/GNSS diverges badly.

I have used the model as my proposed matrices, the result is a little different, but it doesn't diverge.