Objectives The unknown inertial measurement unit (IMU) mounting angles, i.e., the misalignment angles between the smartphone's IMU and the vehicle body frame, can seriously affect the smartphone-based global navigation satellite system (GNSS)/inertial navigation system integrated system for vehicular navigation.

Methods A real-time estimation of the IMU mounting angles is addressed through a Kalman filter whose measurement updates are realized by position variation between epochs derived by time-differenced carrier phase besides the smartphone GNSS/ micro-electro-mechanical system inertial measurement unit (MEMS IMU) tightly-coupled integration algorithm.

Results Two vehicular navigation experiments were carried out in urban challenging environments to evaluate the performance of the proposed algorithm. The results show that a large mounting angle can be initialized within 80 s under moderate occlusion conditions, then convergence of the estimated mounting angle is shown to occur within minutes, and the accuracy is in the order of 2°. When the mounting angle is estimated successfully, the positioning performance of smartphone vehicular positioning will be improved significantly, both horizontal and vertical accuracy can be maintained at about 5 m in terms of root mean square by applying non-holonomic constraint under severe occlusion conditions.

Conclusions The proposed algorithm can estimate large mounting angles accurately and rapidly without any auxiliary equipment such as an odometer.