Vehicle Position Estimation using GPS/CAN Data based on Nonlinear Programming

Lenka Pavelkov√°


nonlinear systems, state-space model, filtering, bounded noise, incomplete data, vehicle position estimation


The paper solves a problem of the estimation of the moving vehicle position. The position is measured by global position system (GPS) but outages sometimes occur in the measurements. During these outages, the actual position is estimated using data from vehicle sensors. A moving vehicle is described by a discrete-time state-space model with bounded noise. This model is constructed using kinematics laws and it can be used for arbitrary type of ground vehicle. Bayesian approach is applied to obtain position estimates. The maximum a posteriori (MAP) estimation converts to the nonlinear programming. The paper also discusses a setting of initial conditions for successful running of estimation process.

Important Links:

Go Back