An Intelligent GPS/Ins Integrated System Using Co-Active Neuro-Fuzzy Inference System (CANFIS)

Abstract

The Global Positioning System (GPS) and Inertial Navigation System (INS) have complementary operational characteristics. To have the benefits of these systems and to overcome their disadvantages, GPS and INS are integrated together to provide accurate position, velocity, and attitude. In this paper, a novel technique based on Co-active Neuro-Fuzzy Inference System (CANFIS) is proposed to implement GPS/MEMS-INS integration. The proposed CANFIS is constructed of the neural network adaptive capabilities and the fuzzy logic qualitative approach. The performance of the CANFIS is examined. In fact, this is extremely difficult to achieve if the Micro-Electro-Mechanical-Systems (MEMS) based inertial sensors are involved in the integrated INS/GPS system due to their relatively high measurement noise. CANFIS is used to model the operation of the MEMS sensors so as to reduce the integrated system’s errors. Experimental results were reached and are outlined in this paper. The results of the study are highly promising and suggest that CANFIS modeling is a more flexible and enhanced alternative to the corresponding conventional models approach in combating MEMS errors.