Kalman Filter
1.1 Linear Optimal Filtering
LOF是kalman filter的精髓,可以描述为:
- inputs: x[0], x[1], …, x[n]
- outputs: y[0], y[1], …, y[n]
- filter: w[0], w[1], …, w[n]
- objection function: en=yn−ŷ n , ŷ n spanned by {xn}
1.2 Orthogonality Principle
LOF得到的
ŷ n
垂直于
yn−ŷ n
1.4 State Model
- Hk,Fk,Gk are determinated and known, MATRIX
- Uk,Bk : White Noises of state and observation
这里B加粗 B 表示 B 是向量。
1.5 Objective And Hypothesis
Objective:
Obtain recursively linear estimators of Xk based on the observations up to the k instants ( {Yn}n=0,...,k ) and the signal statistical properties.
Hypothesis:
E[X0]=x¯0,E[Bk]=0,E[Uk]=0E⎡⎣⎢⎢⎡⎣⎢⎢BkX0Uk⎤⎦⎥⎥[BTlXT0UTl]⎤⎦⎥⎥RkQkδkl=⎡⎣⎢⎢Rkδkl000P0000Qkδkl⎤⎦⎥⎥=E[BkBTk]=E[UkUTk]={1,0,if k=l;otherwise.
这里要证明一下,为什么假设
E[Bk]=0,E[Uk]=0
会是『Without loss of generality 』:
假设
Bk=B′k+B0
,
Uk=U′k+U0
Remark
Stationary: 是指随机过程(不是随机变量)的联合概率密度不随时间改变而改变的状态。
Stationary Process: is a stochastic process whose joint probability distribution does not change when shifted in time. Consequently, parameters such as mean and variance, if they are present, also do not change over time. – From Wiki
须满足
Proof: Xk,Yk are non-stationary unless
- the matrices Hk,Fk,Gk,Rk,Qk are time invariant (noted by H, F, G, R and Q);
- F is a stable “filter”, i.e., all eigenvalues lies within the unit circle;
- x¯0 = 0 and the initial covariance P0 is a solution to the Lyapunov equation:
P=FPFT+GQGT
1.3 Notations
- yn : space spanned by {Yi}i=1,..,n ;
- X̂ (n+1|yn) : optimal linear estimate of Xn+1 knowing yn ;
- Ŷ (n|yn−1) : optimal linear predictor of Yn knowing yn−1 ;
- α[n] =Yn−Ŷ (n|yn−1) : innovation/error, the information we have gained.
Remark
- i = n, X̂ (n|yn−1) is the Kalman Filter (On line)
- i < n, X̂ (i|yn−1) is the Kalman Smoothing (Off line)
Proposition
- α[n]⊥yn−1↔E[α[n]YTk]=0, 1≤k<n
- α[n]⊥α[k]↔E[α[k]α[k]T]=0, k≠n
- Linear transformation
yn={Y1,...,Yn}↔{α[1],...,α[1]}
证明:
1.
这里有一些引理:
◻
2.
3.
Conclusion of Space Relationship
The Innovation Covariance
Proof the prediction of observation:
Proof: 既然 Ŷ (n|yn−1) 表示 optimal linear predictor of Yn knowing yn−1 , 那么它与 Y 的差必定与空间spanned by
{yn−1} 相互垂直.
◻