Ensemble Kalman filter
§Replace xk, Pk from the full KF with an ensemble of xk,i,  i=1,É,Nens
§Add dynamic noise consistent with Qk to xk,i when propagating; add measurement noise consistent with Rk to measurements when updating, initial ensemble has a spread consistent with P0
§When needed in KF equations, Pk replaced with

§Replace matrix multiplications with sums of dot products
§Good for non-Gaussian distributions