In Motion Low-Cost IMU-to-Vehicle Alignment for Intelligent Vehicle Applications Using Kalman Filter

The expansion of mobile devices and their sensors, like IMU and GNSS, increases the number of applications concerning vehicle dynamics and driver behavior monitoring. However, to make this monitoring possible, it is necessary to know the IMU's attitude to the vehicle's frame. To solve this problem, we proposed a novel technique for in motion IMU-to-Vehicle alignment that relies only on data collected from IMU and GNSS. The proposed approach is divided into two phases, leveling and heading. The leveling aligns the IMU to the horizontal plane of the vehicle's frame. It is updated iteratively using the measured gravity counter force when the vehicle is on steady-state. The heading, otherwise, corrects the yaw angle of the IMU relative to the vehicle's chassis. It employs a Kalman Filter to adjust, iteratively, the estimated heading based on vehicle accelerations observed in straight trajectories. The results show a reorientation capability, for any unknown pose, with errors below 7 deg for roll and pitch and below 15 deg for yaw. Although the observed results are similar to those from the literature, the proposed technique has the advantages of not requiring any initial conditions or constraints, being simpler and computationally more efficient.