I cannot calculate the P matrix without any control input
Anyone can help me!
Without control input B = 0
And P=B*Q*B' = 0
Hence Predicting step is difficult
P=A*P*A'+B*Q*B' %predicting P
Yes you can; sometimes we suppose B=0, or u=0; in any case you can calculate it.
To understand better take a look at the example illustrated in the following link:
https://www.academia.edu/7547346/Subject_MI63_Kalman_Filter_Tank_Filling
Mohammad
P is the predicted (a priori) error covariance. If you don't have control input(u=0) it becomes= P(k+1)=AP(k)A'+Q(k)
where Q(k) is covariance matrix of the noise process w(k)
In fact, most of the dynamic systems do not consider input, so I don't guess you have a problem.
For a comprehensive study, you can look at Simon's book about Kalman filter.
" Optimal state estimation: Kalman, H infinity, and nonlinear approaches "
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix