2010 · 11 citations · 17 references
ComputingAutomotive EngineeringAutonomous SystemsAutonomous VehiclesSensor FusionTransportationIntelligent Transportation SystemsKalman FilteringKinematics Of Vehicle MotionNavigation SystemsSensor Fusion For LocalizationVehicle PerceptionAd Hoc NetworksExtended Kalman FilterPositioning In Vehicular NetworksUnscented Kalman FilterUnscented Kalman FilteringVehicle NavigationVehicle Position EstimationVehicular Ad Hoc Networks
In this paper, we propose a cooperative approach based on unscented Kalman filter (UKF) for vehicle positioning in Vehicular Ad hoc Networks (VANET). In our system, various kinematic parameters (e.g., position, speed, heading, and acceleration) of a vehicle are considered as a multi-dimensional data. Accordingly, the kinematic parameters of all the vehicles in the cluster can form a multi-layer and multi-dimensional information (MLMDI) database. Due to the motion characters, most kinematic parameters vary nonlinearly. We specially introduce the UKF to fuse the MLMDI data from different information sources, since UKF has an advantage to reckon the statistics of a random variable undergoing a non-linear transformation compared with extended Kalman filter (EKF). Simulation results show our approach can get more accurate, reliable and computationally efficient than GPS/DR system and the Extended Kalman Filter (EKF) based solution.
17
Unscented Filtering and Nonlinear Estimation
Simon Julier, Jeffrey Uhlmann · Proceedings of the IEEE · 2004 · 6.5K citations
Computing, Data Science, Forecasting +14
The unscented Kalman filter for nonlinear estimation
Eric A. Wan, Rudolph van der Merwe · 2002 · 3.9K citations