An Improved In-Motion Alignment Method for SINS/GPS With the Sliding Windows Integration

In this paper, an improved in-motion alignment method for SINS/GPS integration is proposed. The alignment errors, which are caused by the maneuvering of the vehicle and the accumulated errors of inertial sensor errors, are suppressed. Firstly, the traditional in-motion alignment method for SINS/GPS integration is reviewed, and the existing problems, which are accumulated errors in the observed vectors, are investigated. To address these problems, a sliding windows operation for vector observation optimization is proposed. The derivation for the calculation of the new vector observations is given in detail. Using the new vector observations, the accumulated errors, which are existing in the observed vectors, are weakened. Thus, the alignment results are more stable whenever any maneuvering of the vehicle. Lastly, the simulation and field tests are designed for verification of the proposed method. Compared with the currently popular methods, the alignment errors are converged to a small value quickly.