Title :
Tracking of Multiple Target Types with a Single Neural Extended Kalman Filter
Author :
Kramer, Kathleen A. ; Stubberud, Stephen C.
Author_Institution :
Dept. of Eng., San Diego Univ., CA
Abstract :
The neural extended Kalman filter is an adaptive state estimation routine that can be used in tracking systems to aid in the tracking through maneuvers. A neural network is trained using a Kalman filter training paradigm that is driven by the same residual as the state estimator and approximates the difference between the a priori model used in the prediction steps of the estimator and the actual target dynamics. This approach has been shown to provide improved tracking capabilities in maneuver tracking. An important benefit of the technique is its versatility because little if any a priori knowledge is needed of the dynamics of the target. This allows the neural extended Kalman filter to be applied to a generic tracking system that will encounter various classes of targets. Here, the neural extended Kalman filter is applied simultaneously to three separate classes of targets each with different maneuver capabilities. The results of the analysis show that the approach is well suited for use within a tracking system without prior knowledge of the targets´ characteristics
Keywords :
Kalman filters; learning (artificial intelligence); neural nets; state estimation; target tracking; Kalman filter training; adaptive state estimation; maneuver tracking; multiple target tracking; neural extended Kalman filter; neural network; tracking capability; tracking systems; Adaptive systems; Intelligent structures; Intelligent systems; Motion measurement; Neural networks; Noise measurement; Performance analysis; Standards development; State estimation; Target tracking; Kalman filters; multiple targets; neural networks; target tracking;
Conference_Titel :
Intelligent Systems, 2006 3rd International IEEE Conference on
Conference_Location :
London
Print_ISBN :
1-4244-01996-8
Electronic_ISBN :
1-4244-01996-8
DOI :
10.1109/IS.2006.348463