Page 8 - Demo
P. 8
Likelihood func@on
Normal distribu@on with covariance ΣD and mean KI.
11
−1
f(D|I)= exp(− (D−KI)TΣD(D−KI))
N/2 1/2 2 (2π) D ΣD
Sensors are independent -> ΣDis diagonal.
Σ =σ2δ D iij
Smoothing prior ( targets transport physics )
Normal distribu@on with covariance ΣI and mean μI.
11 N/21/22I I
−1
f(I)= exp(− (I−μ)TΣI (I−μ))
(2π) I ΣI
Σpis func@on of scale lengths {σR σZ } along R and Z direc@ons
⎛ (R−R)2 (Z−Z)2⎞ Σ (i,j)=σ2exp⎜− i j − i j ⎟
II⎜σ2σ2⎟ ⎝R Z⎠
If the scale lengths are long -> low prior probability for very dissimilar nearby currents.
Presenter : J.Romero. Sesion T07. Computational and theoretical Techniques. APS 2016.
8
