RTDGPS Accuracy Improvement Using PSO-LSWSVM and Low-Cost GPS Receivers

The position of mobile devices is determined by Real Time Differential Global Positioning System (RTDGPS). This system is composed of fixed and mobile station. The accuracy of this system depends on the speed of updating the correction data. Yet, this system often faces the problem of the reference GPS receiver's signal cut off and internal or non- internal delays which increase the latency time and decrease the updating speed. The prediction models are used to compensate these delays. In this paper, the Particle Swarm Optimization Least Square Wavelet Support Vector Machine (PSO-LSWSVM) is used for predicting the DGPS correction. An experimental setup is designed in order to carry out the operations of reference and user stations. Low-cost receivers were used in both stations. The accuracy of PSO-LSWSVM was evaluated through various simulations and experimental. The practical tests showed that the accuracy of the designed RTDGPS is 0.42 m with PSO-LSWSVM and is equal to 0.73 m without it. In comparing with other method that recently is introduced; it has a higher positioning accuracy.