| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 108456 | Journal of Transportation Systems Engineering and Information Technology | 2013 | 6 Pages |
It is difficult to establish a precise mathematical model of dynamic filtering for vehicle GPS navigation with Kalman filtering. And the traditional wavelet transformation has some shortness in real-time processing. This paper proposes a GPS dynamic filter algorithm based on the lifting wavelet transformation (LWT). It first decomposes the vehicle GPS navigation signal with the LWT. Then, it detects, eliminates and corrects the signal's gross error at different LWT resolution levels with statistical 3srule and polynomial interpolation method. The signal's noise is de-noised at different LWT resolution levels with modulus square soft-threshold de-noising method. Finally, the algorithm realizes real-time dynamic filtering for the vehicle GPS navigation system through reconstructing the de-noised LWT coefficients. The simulation results show that the algorithm is more effective than the Kalman filter in positioning precision. Compared with the traditional wavelet transformation, the speed is doubled for dynamic filtering even though the proposed algorithm only has little advantages on the positioning accuracy. Therefore, the algorithm is available for vehicle GPS navigation system.
