| Summary: | 碩士 === 國立臺灣海洋大學 === 導航與通訊系 === 92 === The objective of this thesis is to utilize the neural networks aided Kalman filter on the design of GPS carrier tracking loop. The Neural Network is employed for compensating the dynamic modeling errors. In addition to the assumption that both the process and measurement are corrupted by zero-mean Gaussian white noises, the designers are required to have good knowledge on both dynamic process and measurement models to achieve good Kalman filtering solutions. Therefore, for obtaining good estimation on the state parameters the state space model needs to be clearly understood. In fact, most of the systems in the real world are non-linear and difficult to describe its state space model. In the meantime, the mathematical model of the carrier tracking loops in the GPS receiver should be non-linear no matter how small the sampling interval is selected. To resolve the problem mentioned above the adaptable learning ability based on the neural networks will be used. The well trained neural networks will be able to compensate the uncertainty of Kalman filter, and to prevent GPS signal being unlocked due to the dynamic modeling error caused by Doppler effect.
|
|---|
