Tracking and data association
IEEE Transactions on Signal Processing
Hi-index | 0.00 |
A multirate Kalman filtering algorithm for target tracking with high-ordercorrelatednoise is proposed. The measurement signal is first split into subbands usinga filter bank.Then, the correlated noise in each subband is modeled using a first-order ARprocess and the AR parametersare identified online. Finally, a multirate Kalman reconstruction filter isused to obtain the state estimate.This method can be directly incorporated into the IMM algorithm, resulting inan effective tracking scheme.Simulations show that the new multirate processing scheme can significantlyimprove tracking performance.