Ellipsoidal Set Based Filter for Bearing-only Maneuvering Target Tracking

摘要

On a plane of Cartesian coordinate, a maneuveringtracking problem with bearing-only measurements usingtwo static platforms is considered and an ellipsoidal setbased filter algorithm for nonlinear state estimation isdeveloped. The state transition equation is linear and thebearing-only measurement equation is nonlinear. Both ofthose equations are modeled with unknown but boundednoise terms. Using the innovation, a detection scheme hasbeen developed to determine whether a maneuver isindeed occurring. If the target is under manoeuvre, weconsider the impact of manoeuvre in target states as achanged bounded noise term in state transition equation.The ellipsoidal set based filter estimation defines anellipsoidal set in position and velocity space which mustcontain the true target state. The ellipsoidal set based filterhas been compared to the extended Kalman filter andextended optimal bounding ellipsoid, and the simulationresults show that the ellipsoidal set based filter issuccessful in accommodating unknown maneuver modelsand unknown but bounded noise for maneuvering targettracking problems.