Hi all,

I am a newbie here and sorry if my question sounds stupid. Currently I am going through the tutorials at the getting started guide.

The system model for the robot is given by

x_k = x_{k-1} + v_{k-1} * cos(theta) * delta_t y_k = y_{k-1} + v_{k-1} * sin(theta) * delta_tIn the non-linear system model of EKF (http://people.mech.kuleuven.be/~tdelaet/bfl_doc/getting_started_guide/node20.html), it says that

The expected value is then calculated and returned by:

state(1) += cos(state(3)) * vel(1); state(2) += sin(state(3)) * vel(1); state(3) += vel(2); return state + AdditiveNoiseMuGet();in the ExpectedValueGet() function.

I want to know how delta_t is associated here.

Thank you in advance

Peshala

## EKF Getting Started Tutorial Example

Hi Peshala,

On Mon, Apr 19, 2010 at 04:52, peshala_24 [..] ... <peshala_24 [..] ...>wrote:

> Hi all,

>

> I am a newbie here and sorry if my question sounds stupid.

> Currently I am going through the tutorials at the getting started guide.

>

> The system model for the robot is given by

>

> x_k = x_{k-1} + v_{k-1} * cos(theta) * delta_t

> y_k = y_{k-1} + v_{k-1} * sin(theta) * delta_t

>

> In the non-linear system model of EKF (

> http://people.mech.kuleuven.be/~tdelaet/bfl_doc/getting_started_guide/no...),

> it says that

>

> The expected value is then calculated and returned by:

>

> state(1) += cos(state(3)) * vel(1);

> state(2) += sin(state(3)) * vel(1);

> state(3) += vel(2);

> return state + AdditiveNoiseMuGet();

>

> in the ExpectedValueGet() function.

>

> I want to know how delta_t is associated here.

>

Delta_t is just assumed to be one in this case.

Another view is to interpret vel as the velocity multiplied with the time

step.

For real applications it might be useful to provide a physical

interpretation of the velocity. In this case you will have to take into

account the period at which the algorithm is running.

Good luck.

@Peshala: I noticed that you asked a mail about using BFL in ros on the ros

mailinglist. Are you using the BFL-ros package of the examples?

_______________________________________________

I hereby promise not to top-post on the

BFL mailing list

BFL [..] ...

http://lists.mech.kuleuven.be/mailman/listinfo/bfl

## EKF Getting Started Tutorial Example

Hi Tinne De Laet,

Thanks for the feedback.

Yes I am using the BFL-ros package.

Peshala

Tinne De Laetwrote:On Mon, Apr 19, 2010 at 04:52, peshala_24 [..] ... <peshala_24 [..] ...>wrote:

> Hi all,

>

> I am a newbie here and sorry if my question sounds stupid.

> Currently I am going through the tutorials at the getting started guide.

>

> The system model for the robot is given by

>

> x_k = x_{k-1} + v_{k-1} * cos(theta) * delta_t

> y_k = y_{k-1} + v_{k-1} * sin(theta) * delta_t

>

> In the non-linear system model of EKF (

> http://people.mech.kuleuven.be/~tdelaet/bfl_doc/getting_started_guide/no...),

> it says that

>

> The expected value is then calculated and returned by:

>

> state(1) += cos(state(3)) * vel(1);

> state(2) += sin(state(3)) * vel(1);

> state(3) += vel(2);

> return state + AdditiveNoiseMuGet();

>

> in the ExpectedValueGet() function.

>

> I want to know how delta_t is associated here.

>

Delta_t is just assumed to be one in this case.

Another view is to interpret vel as the velocity multiplied with the time

step.

For real applications it might be useful to provide a physical

interpretation of the velocity. In this case you will have to take into

account the period at which the algorithm is running.

Good luck.

@Peshala: I noticed that you asked a mail about using BFL in ros on the ros

mailinglist. Are you using the BFL-ros package of the examples?

_______________________________________________

I hereby promise not to top-post on the

BFL mailing list

BFL [..] ...

http://lists.mech.kuleuven.be/mailman/listinfo/bfl

_______________________________________________

I hereby promise not to top-post on the

BFL mailing list

BFL [..] ...

http://lists.mech.kuleuven.be/mailman/listinfo/bfl

## Re: EKF Getting Started Tutorial Example

Hi Tinne De Laet,

Thanks for the feedback. Yes I am using the BFL-ros package.

Peshala

Tinne De Laetwrote:On Mon, Apr 19, 2010 at 04:52, peshala_24 [..] ... <peshala_24 [..] ...>wrote:

> Hi all, >> I am a newbie here and sorry if my question sounds stupid. > Currently I am going through the tutorials at the getting started guide. >> The system model for the robot is given by >> x_k = x_{k-1} + v_{k-1} * cos(theta) * delta_t > y_k = y_{k-1} + v_{k-1} * sin(theta) * delta_t >> In the non-linear system model of EKF ( > http://people.mech.kuleuven.be/~tdelaet/bfl_doc/getting_started_guide/node20.html<http://people.mech.kuleuven.be/%7Etdelaet/bfl_doc/getting_started_guide/node20.html>), > it says that >> The expected value is then calculated and returned by: >> state(1) += cos(state(3)) * vel(1); > state(2) += sin(state(3)) * vel(1); > state(3) += vel(2); > return state + AdditiveNoiseMuGet(); >> in the ExpectedValueGet() function. >> I want to know how delta_t is associated here. >

Delta_t is just assumed to be one in this case. Another view is to interpret vel as the velocity multiplied with the time step.

For real applications it might be useful to provide a physical interpretation of the velocity. In this case you will have to take into account the period at which the algorithm is running.

Good luck.

@Peshala: I noticed that you asked a mail about using BFL in ros on the ros mailinglist. Are you using the BFL-ros package of the examples? _______________________________________________ I hereby promise not to top-post on the BFL mailing list BFL [..] ... http://lists.mech.kuleuven.be/mailman/listinfo/bfl