Combinige generalized JDPA and FRLS filter for tracking multiple maneuvering targets

This paper proposes a generalized joint probabilistic (JPDA) filter for tracking multiple maneuvering targets in situations of observations with unknown random characteristics. In the proposed filter, the joint association probabilities in the standard JPDA filter are reconstructed by utilizing the generalized association probabilities of observations belonging to the targets. To calculate the generalized association probabilities, two measures of the uncertainty of statistical and fuzzy observations are defined. Using the measures, an adaptively additive fusion strategy is also proposed, which can process both statistical and fuzzy observations and keep the consistency of the estimated states with fuzzy observations. Then the fuzzy recursive least squares (FRLS) filter is adopted to update all tracks. The proposed filter has the advantage that the restrictive assumptions of statistical models for process noise and motion models are relaxed, and it does not need a maneuver detector when tracking multiple maneuvering targets. Moreover, it can adaptive adjust the weights of different types of observations in association decision according to the changes of observational environments. The performance of the proposed filter is evaluated by using the simulated data. It is found to be better than those of the traditional filters in tracking accuracy.