Study on INS/GPS Sensor Fusion for Agricultural Vehicle Navigation System

This study was performed to investigate the effects of inertial navigation system (INS) / global positioning system (GPS) sensor fusion for agricultural vehicle navigation. An extended Kalman filter algorithm was adopted for INS/GPS sensor fusion in an integrated mode, and the vehicle dynamic model was used instead of the navigation state error model. The INS/GPS system was consisted of a low-cost gyroscope, an odometer and a GPS receiver, and its performance was tested through computer simulations. When measurement noises of GPS receiver were 10, 1.0, 0.5, and 0.2 m ( $1{\sigma}$ ), RMS position and heading errors of INS/GPS system at 5 m/s straight path were remarkably reduced with 10%, 35%, 40%, and 60% of those obtained from the GPS receiver, respectively. The decrease of position and heading errors by using INS/GPS rather than stand-alone GPS can provide more stable steering of agricultural equipments. Therefore, the low-cost INS/GPS system using the extended Kalman filter algorithm may enable the self-autonomous navigation to meet required performance like stable steering or more less position errors even in slow-speed operation.