Application of ANFIS for Frequency Syntonization Using GPS Carrier-Phase Measurements

A new scheme of frequency syntonization based on the GPS carrier-phase measurements using an adaptive neural-fuzzy inference system (ANFIS) is proposed in this paper. A GPS carrier-phase disciplined oscillator (GPSCDO), which uses pure carrier-phase measurements to steer an oven-controlled crystal oscillator (OCXO) has been designed. High short-term frequency stability of the GPSCDO is obtained. However, like traditional GPSDO, the long-term performance of the GPSCDO is limited due to the atmospheric effect. Therefore, an ANFIS is employed in an attempt to improve the long-term frequency stability of the GPSCDO. By applying the off-line data, which are predicted by the ANFIS, the remote GPSCDO can correct its estimations of the frequency offsets, which is used as input of a fuzzy controller. Experimental results show that for averaging times of one day, the frequency stability of the GPSCDO can be improved from a few parts in 10¹² to 10¹⁴.