Multisensor tracking of a maneuvering target in clutter using IMMPDA fixed-lag smoothing

We present a suboptimal fixed-lag smoothing algorithm for tracking a highly maneuvering target in a cluttered environment using multiple sensors. The tired-lag smoothing algorithm is developed by applying the basic interactin g multiple model (IMM) approach and the probabilistic data association (PDA ) technique to a state-augmented system. In the past this approach had been restricted to Markovian switching systems with no uncertainty regarding th e origin of the measurements (i.e., no clutter). The algorithm is illustrat ed,ia a highly maneuvering target tracking simulation example where two sen sors, a radar and an infrared sensor, are used and are assumed to operate i n a cluttered environment, Compared with an existing IMMPDA filtering algor ithm, the proposed smoothing algorithm achieves significant improvement in the accuracy of track estimation by introducing a small time lag between th e instants of estimation and latest measurements whereas the computational load increases linearly with lag. However, the tracking delay may lead to u ndesired effects on control loops in certain applications.