Hybrid Unscented Kalman Filter with Rare features for Underwater Target tracking using Passive Sonar Measurements

A Novel estimator called as Hybrid Unscented Kalman Filter(HUKF) is developed in the paper to tackle the issue of passive target tracking in underwater scenarios using bearing-only measurements (captured by a towed array). As the name indicates the algorithm is a hybrid one obtained by combining three existing algorithms namely UKF, Integration technique and Pre-processing mechanism to yield much better performance than any of them individually. The sensor noise reduction in spatial measurements caused by integration technique together with noise reduction in time measurements caused by Pre-processing mechanism in successive time iterations cumulatively contribute to the improvement in performance. Moreover the montecarlo simulations in Matlab(R2009a) provide evidence that, HUKF also display attractive features like Optimal Performance(produce less estimation errors at less computational complexity), Efficient Tracking in critical conditions(long ranges/severe noise) and Non-Divergence even with nasty initialization. Lastly the algorithm can be helpful in promoting TA usage on regular basis despite of the associated maneuvering issues.