Ellipsoidal Set Based Filter for Bearing-only Maneuvering Target Tracking

摘要

On a plane of Cartesian coordinate, a maneuvering tracking problem with bearing-only measurements using two static platforms is considered and an ellipsoidal set based filter algorithm for nonlinear state estimation is developed. The state transition equation is linear and the bearing-only measurement equation is nonlinear. Both of those equations are modeled with unknown but bounded noise terms. Using the innovation, a detection scheme has been developed to determine whether a maneuver is indeed occurring. If the target is under manoeuvre, we consider the impact of manoeuvre in target states as a changed bounded noise term in state transition equation. The ellipsoidal set based filter estimation defines an ellipsoidal set in position and velocity space which must contain the true target state. The ellipsoidal set based filter has been compared to the extended Kalman filter and extended optimal bounding ellipsoid, and the simulation results show that the ellipsoidal set based filter is successful in accommodating unknown maneuver models and unknown but bounded noise for maneuvering target tracking problems.