J. Dang (PRC)
Motion objects tracking, Kalman filter and variance matrix
A new algorithm for tracking motion objects is proposed with the Kalman filters of weighting measurement variance matrix. This algorithm is based upon introducing the estimation of measurement variance matrix into Kalman filters and weighing on the estimation of measurement variance matrix in motion objects tracking systems to emphasize the function of new measurement in Kalman filter, to reduce estimation error of object motion state and to prevent the divergence of Kalman filter. The results of Mont Carlo simulation show that the algorithm proposed in this paper has adaptive ability for system measure error and modeling error. Even if object tracking systems have modeling error and time-vary measure error, this algorithm also can achieve preferable tracking precision on the position, velocity, and acceleration of motion objects.
Important Links:
Go Back
