sheraz.pervaiz,
There is nothing special about implementing a Kalman filter, except for the fact that floating point matrix manipulations are required.
.
For an introduction to the Kalman filter, read the following (in the order given):
.
h**p://www.cs.unc.edu/~welch/kalman/sorensonPaper.html
h**p://www.cs.unc.edu/~welch/kalman/kalmanIntro.html
Regards,
Kral
For Kalman filter, as far as I know, we need to make some assumptions on noise patterns (whitenoise, uncorrelated) and do state propagation and state update...
1. Filter design - find out the states of your system
2. Calculate Q and R. (Robert brown and Patrick Hwang book is a material good on this)
3. There is also lot in tuning.
But before all these your state-space design of the system should be available
in order to impliment Kalman filter in DSP, use the LMS or RLS algo which are present in MAtlab. Kalman and RLS are One-to-One corrospondance with each other. if u impliment RLS algo and change the Desired response or Misadjestment. u can achive the Kalman behaviour. Some of the examples are givven in Matlab. cheak the Simulink help. if u dont find any thing related to kalman, use the RLS algo and change the above mentioned parameters
for more detail Plz Study:
Adaptive Filter Theory (4th Edi)
by: Simon Haykin
chapter 10, chapter 5.(Real time applications topic)