An intelligent Differential GPS using Π-Σ Neural Network

The main component of Global Positioning System (GPS) positioning error results from time and space varying conditions of radio wave propagation, which depend on atmospherics, disturbances in the satellite constellation, orbit stability, and also due to the U. S. military intentionally such as Selective Availability (SA), among other things. The error caused by the transmitter and receiver operation precision or time-measurement accuracy is negligible. A method for GPS precision enhancement commonly used by civil users is the Differential GPS (DGPS). If DGPS service is interrupted, it will lead to the degraded navigation performance. This paper focuses on applying a Π-Σ Neural Network (PSNN) model to predict Pseudo-Range Corrections (PRC) for DGPS. A low cost commercial module (Rockwell single-frequency GPS receiver) is employed to represent the improvement in DGPS. Experimental results show the proposed NN can online predict the PRC precisely when the PRC signal is lost for a short period of time.