325 / 1971-01-01 00:00:00

Novel Probabilistic Data Association Filter For Maneuvering Target Tracking

maneuvering target tracking,probabilistic data association,rescursive least squares filter,fuzzy system

A novel probabilistic data association-fuzzy recursive least squares filter (PDA-FRLSF) is presented for tracking a maneuvering target in clutter environments. In the proposed method, the association probabilities of the current valid observations belonging to the target are calculated by the probabilistic data association (PDA) algorithm. These probabilities are used to weight the observations for producing a fused observation. Then the fused observation is applied to calculate its observation residue and heading change, and a fuzzy system is designed according to the relationship between them and the target’s maneuver. Observation residues and heading changes are employed as the input variables of the system while the filter’s fading factor as its output variable. The impact of the predicted innovation related with the fused observation on the current state estimate is adjusted by the fading factor for updating the target state. The proposed filter has the advantage that any assumptions of prior information on process noises and target dynamics are not required. Moreover, it does not need a maneuver detector even when tracking a maneuvering target. The performance of PDA-FRLSF is evaluated by using a simulation experiment, and it is found to be better than those of the PDA filter and the PDA-IMMF in tracking accuracy.