Vehicle Trajectory Estimation in GPS-blocked Environments Using Inertial Measurement Units, Estimated Drive Speed, and bidirectional Kalman Filters

Abstract

Road roughness is a widely used index for evaluating pavement serviceability and determining maintenance needs. However, road roughness estimation methods utilizing GPS information encounter difficulties in GPS-blocked environments; existing non-GPS vehicle mapping techniques utilizing only inertial measurement unit (IMU), have a limited accuracy when IMU is not of high precision. To address this problem, this study introduces a vehicle mapping approach that relies solely on an IMU at GPS-blocked road sections. The proposed method consists of five steps utilizing GPS data at GPS-accessible sections and IMU data for all sections. The data are transformed into Universal Transverse Mercator (UTM) and body coordinate systems. The initial positions, velocities and accelerations at both ends of the GPS-blocked section are estimated by optimization based on kinematic information outside the GPS-blocked section. Then the drive speed is estimated by the inverse analysis of vehicle body dynamic responses. Finally, the vehicle states within the GPS-blocked section are estimated using the IMU data as well as the estimated drive speed through a bidirectional Kalman filter. The effectiveness of the method is evaluated at linear and curved road sections. The results show the high accuracy of the method.