An efficient estimation approach for moving object tracking with correlated measuring noises

Due to the large deviations in the position data from GPS receiver, the precious positions of the moving objects must be calculated by Kalman filter. But the system measuring noise is colored rather than white noise, the basic Kalman filter cannot be used directly. After the positioning error is fit to be a 2-order Markov process, a new algorithm is designed to eliminate the noises' correlations, using the subtraction values between the adjacent measurements to be the new measurements. It can avoid higher-order matrix inversing, and certainly, the computation is simplified greatly and the less hardware resource is needed. The new system equations are rebuilt, and the implementation procedure is discussed in detail. The simulation result shows that the positioning and tracking precision is increased effectively and the real-time tracking requirement is available.