A hybrid parametric, non-parametric approach to Bayesian target tracking

A Bayesian tracker, based on a hybrid of parametric and non-parametric techniques, has been successfully implemented and demonstrated on a simulator with a scenario comprising a single weaving target and a single radar. Its performance has been found significantly much better, for position estimates, than a basic Kalman filter equivalent and otherwise comparable. However, it requires much more computation time. At this stage, it shows much promise for the more difficult scenarios which are not handled well by Kalman filter trackers