부산대학교 도서관

A single microelectromechanical system (MEMS) inertial measurement unit (IMU) that is mounted on a wheel can achieve a higher vehicle dead reckoning accuracy than a conventional odometer-aided inertial navigation system (INS) because a wheel-mounted IMU can be equivalent to a combination of a virtual odometer and a rotary IMU that can utilize rotation modulation technology. However, the kinematic model used in previous works was simplified without considering the effects of wheel rotation and vehicle body motion on the accelerometers, resulting in very large localization errors, especially when the IMU is not strictly mounted in the center of the wheel and the vehicle speed is higher. Moreover, the mounting parameters, i.e., the misalignment angles and eccentric distance, markedly reduce the localization accuracy. Thus, the parameters should be well estimated and compensated. A kinematic model that considers all major influences, e.g., the wheel rotation, vehicle body motion, and mounting parameters, on the IMU outputs is revealed, and an IMU-wheel alignment and an extended Kalman filter (EKF) are applied to simultaneously estimate the pose of the vehicle and mounting parameters of the IMU with no additional sensors. The results of dozens of experiments show that the proposed method achieves higher accuracy with regards to the position, heading, and velocity than the state-of-the-art method and that the localization accuracy is increased by approximately 1/3.