Math Forum Discussions

SURESH THOMMANDRU

Posts: 5
Registered: 5/3/12

Tracking midpoint and bounding box
Posted: May 3, 2012 10:41 PM

Plain Text

In Kalman filter
state x[k] = Ax[k-1]
measurement z[k]=Hx[k]
Error Covariance P= AP[k-1]A' + Q[k-1]. -------1

from this
if we take state as [m m' w] m is mid point of line segment.
P[k-1], Q[k-1] as Posterior error covariance and process noise covariance.

Then how will get P[k] = [B1 B2 0; B2 B3 0; 0 0 B4] matrix from above equations 1.

where B1 = p11 + 2Tp12 + T*T P22 + q11.
B2 = p12 + Tp22 + q12.
B3 = p22 + q22 B4 = p33 + q33.

please give the solution(way) to understand this.