Kalman Filtering common sense

1. Common state prediction equation, which has several? (★)
   prediction equation is your reaction to the platform of a priori knowledge, common are:

• Platform to maintain the same position, then it is $x_{t+1} = X_{t}$.
• If uniform rectilinear motion (in front), $X = [x, v]^T$； $X_{t+1} = FX_{t}$, among them $F = [1,1; 0, 1]$;
• If it is uniformly accelerated linear motion, the situation is a little more complicated; $X_{t+1} = f(X_{t})$ , When the time interval as a constant, f is a linear function remains, X_ {t + 1} =

2. How predictive equation covariance matrix to understand? (★)
   after passing through the above prediction equation, the variable is still in line with normal distribution.
   $\hat{P} = FPF^T + Q$
   F. Univariate normal situation and propagation error is the same, such as: $(u,\sigma^2)->(2u, 4\sigma^2)$. Q is the amount of error to bring other, there is some state equations (third), and some do not (first, second).