On Oct 3, 3:36 am, Rafael Deliano <Rafael_DelianoENTFER...@t-
online.de> wrote:
> > Which books should i read to achieve that?

>
> In print ( $120 ), simple, 650+ pages: =20
> Zarchan, Musoff "Fundamentals of Kalman Filtering" AIAA 2000 ;
>    2. edition 2005: 760 pages, includes now alpha-beta filter (?),
>    all converted to Matlab (?).
>    First edition has code in Fortran in the text, lots of worked
>    out examples. Only reading won=B4t help, porting the code and
>    running the examples will.
>
> Good book, but perhaps not as an introduction:
> Grewal, Andrews "Kalman Filtering"
>    Prentice Hall 1993( Source in Fortan on disk )
>    2. expanded edition 2001 with matlab on CD (?)
>    Perhaps the 2. edition is more readable, has more examples.
>    It explains implementation details that will be useful in
>    the long run.
This is the 'other' book I talked about above.  The Zarchon and Musoff
book calculates the covariance matrix Q in a different way than the
Dan Simon book.  Both books can't be correct.  The resulting Kalman
gains will be different if the Q matrix is different from the Dan
Simon book.   I think the Dan Simon book is right and Zarchan and
Musoff is wrong.  What I did like about the Zarchan and Musoff book is
the concept of just having a bandwidth for the filter instead of using
the Riccati formula to compute the Kalman gains.   The user can easily
calculate new 'Kalman' gains by a formula that converts bandwidth into
gains.   The filtering will not be optimal but it will have 'inertia'
like the Kalman filters due to the state transition matrix.

This brings up another topic.  On the sci.engr.control group we have
had a discussion about modeling.   One must have values to put into
the state transition matrix.  Where do they come from?   Sure, you can
use a Kalman filter to filter an encoder but what about system that
isn't expected to do the same thing as it did the last iteration?    I
think one should understand modeling and system identification before
getting into Kalman filters.

Peter Nachtwey

> Which books should i read to achieve that?

Very simple, readable, but out of print :
Bozic "Digital and Kalman Filtering" Arnold 1979 
  First half is digital filters ( and alpha-beta filter ).
  Second half is introduction to scalar Wiener and then to "scalar" 
  Kalman. Then expands to "vector" Kalman which is the usual real 
  world one. Includes short sourcecode in Fortran. 
  $40 at www.abebooks.com, but thats a bit much for a 155 pages book. 
 
In print ( $120 ), simple, 650+ pages:  
Zarchan, Musoff "Fundamentals of Kalman Filtering" AIAA 2000 ;
   2. edition 2005: 760 pages, includes now alpha-beta filter (?),
   all converted to Matlab (?).
   First edition has code in Fortran in the text, lots of worked 
   out examples. Only reading won&#4294967295;t help, porting the code and
   running the examples will. 

Good book, but perhaps not as an introduction: 
Grewal, Andrews "Kalman Filtering" 
   Prentice Hall 1993( Source in Fortan on disk )
   2. expanded edition 2001 with matlab on CD (?)
   Perhaps the 2. edition is more readable, has more examples. 
   It explains implementation details that will be useful in 
   the long run. Starts where Bozic & Zarchan end. 

MfG JRD

In article <4702028f$0$90262$14726298@news.sunsite.dk>,
 "Beginner" <beginner@b.com> wrote:

> I would want to learn to use and understand Kalman filtering. Which books 
> should i read to achieve that?
> 
> 
> 
> My algebra math skills are at elementary linear algebra level.

take a look at "kalman filtering, theory & practice using matlab" by 
grewal & andrews, 2nd ed. i thought it was a pretty fair introduction. i 
learned how to do the computations from this.

my personal favorite is "time series analysis by state space methods" by 
durbin & koopman. i found it easy going, but i bring a strong math & 
statistics background to it. still, it was way more readable than the 
gelb book (mentioned by others), which i never made it through. i 
learned what was going on from this book.

vale,
   rip

-- 
NB eddress is r i p 1 AT c o m c a s t DOT n e t

On Oct 2, 1:33 am, "Beginner" <begin...@b.com> wrote:
> I would want to learn to use and understand Kalman filtering. Which books
> should i read to achieve that?
>
> My algebra math skills are at elementary linear algebra level.
I ditto the suggestions that sharpen your math skill.  If you don't
understand Kalman filters then what do you expect a Kalman filter to
do that a Butterworth or some other filter can't?

There are some simple forms of Kalman filter that you may be able to
understand quickly.  If these simple forms will work for you then that
level of understand and math is all you need.

Peter Nachtwey

On Oct 2, 5:11 pm, dspg...@netscape.net wrote:
> On Oct 2, 11:36 am, Tim Wescott <t...@seemywebsite.com> wrote:
>
> > Beginner wrote:
> > > I would want to learn to use and understand Kalman filtering. Which books
> > > should i read to achieve that?
>
> > > My algebra math skills are at elementary linear algebra level.
>
> I would recommend "Applied Optimal Estimation" by Gelb. It is done
> without too much math - it uses an algebraic approach.  The filter can
> be derived without the use of statistics, but too eventually
> understand the behaviour of the filter  -  a knowledge of statistics
> of linear systems would be an asset.
>
> Cheers,
> David
I have Applied Optimal Estimation.  It is difficult reading.  There is
no way a rooky is going to get anything from that book.
I would buy Dan Simon's Optimal State Estimation instead.  I have one
beef with this book, there is an errata and even the errate is wrong
the last time I looked.

I have another book on Kalman filtering and it differs greatly from
the Dan Simon book on how to calculate the covariance array Q. I know
the Dan Simon book is correct because I can verify the calculations of
the Kalman gains are correct.  My point is that it is difficult to
learn on your own because the books have so many flaws.  One must beat
their heads against the books and compare them for inconsistencies.
Then try sample problems to veryify which equations are correct.

I find the general situation pitiful.  :(

Peter Nachtwey

On Oct 2, 11:36 am, Tim Wescott <t...@seemywebsite.com> wrote:
> Beginner wrote:
> > I would want to learn to use and understand Kalman filtering. Which books
> > should i read to achieve that?
>
> > My algebra math skills are at elementary linear algebra level.
>

I would recommend "Applied Optimal Estimation" by Gelb. It is done
without too much math - it uses an algebraic approach.  The filter can
be derived without the use of statistics, but too eventually
understand the behaviour of the filter  -  a knowledge of statistics
of linear systems would be an asset.

Cheers,
David

Beginner wrote:
> I would want to learn to use and understand Kalman filtering. Which books 
> should i read to achieve that?
> 
> My algebra math skills are at elementary linear algebra level.
> 
You'll not get a full understanding of Kalman filtering without 
understanding statistics, which requires some calculus (and knowing 
differential equations wouldn't hurt).

Unless you've tried and failed at it already, I'd suggest you start by 
working on your math skills.

Failing that, there may be a Kalman filtering book out there that 
doesn't require advanced math, but it'd be at the recipe level -- I 
don't think you'd get a very deep understanding out of it.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html

On Oct 2, 4:33 am, "Beginner" <begin...@b.com> wrote:
> I would want to learn to use and understand Kalman filtering. Which books
> should i read to achieve that?
>
> My algebra math skills are at elementary linear algebra level.


Try this first  and then some of the external links


http://en.wikipedia.org/wiki/Kalman_filter