The first chapter of Peter Maybeck's book is an excellent place to start.
It's free on-line. If you want to know more after reading that, I strongly
recommend Peter Joseph's paper. The introduction is available on-line and
he will email you the first four chapters for free. He will email you the
more advanced chapter for less than $10, which is a very good deal.
Joseph's paper was much easier for me to understand than the higher (>1)
chapters of Maybeck's book, but Maybeck's book is available to buy also.
Maybeck's first chapter: http://www.cs.unc.edu/~welch/kalman/maybeck.html
Joseph's first chapter:
Assorted other Kalman filter stuff:
I agree that the Welch-Bishop and the Maybeck sources are
probably the best available.
I studied and I studied the literature and it didn't make any sense
until I reread the Maybeck chapter. The most important part for
me was to add two normal graphs like described by Maybeck.
As soon as I did that to compute the new mean and sigma in
a recursive form, then I immediately saw the functionality of
the kalman filter in 1D.
So, I would say that you first understand how to add two normal
distributions in batch form and then in recursive form. Maybeck's
paper demonstrates that. The correlation to the multi-dimensional
matrix form Kalman filter then becomes obvious.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.