Abstract:

Precise vehicle heading information is of great importance for many Intelligent Transportation Systems (ITS) applications. GPS-based localization and heading estimation is widely used in almost every transportation systems. However, dense urban environment causes inconsistency in the reception of the GPS signals. Given the diverse sensors within mobile devices, i.e., Microelectromechanical System (MEMS) sensors such as gyroscope, accelerometer, magnetometer etc., they have a strong potential for sensing vehicle dynamics and can promote a broad range of applications associated with heading estimation. A magnetometer sensor of a smart mobile device can be utilized to obtain accurate vehicle heading estimation. However, ferromagnetic components of a vehicle significantly deforms the magnetic field measured by magnetometer sensor. In this study, it is demonstrated that an accurate vehicle heading estimation can numerically be achieved through identifying error parameters. These parameters were then transformed into a mathematical model and contributing errors were eliminated from raw sensor output. Simulation results show that the model produces a maximum error of 3.4%.

Keywords: