HighlightsAn integrated GPS/INS/VNS navigation system was developed to improve navigation accuracy.An adaptive federal Kalman filter with information distribution factors was used to fuse navigation information.Detection of seedling row lines was achieved based on subregional feature points clustering.A modified rice transplanter was developed as an experimental platform for automatic navigation.Abstract. In this article, an integrated global positioning system (GPS), inertial navigation system (INS), and visual navigation system (VNS) navigation method based on an adaptive federal Kalman filter (KF) is presented to improve positioning accuracy for a rice transplanter operating in a paddy field. The proposed method used GPS/VNS to aid the INS and reduce the influence of the accumulated error of the INS on navigation accuracy. An adaptive federal KF algorithm was designed to fuse navigation information from different sensors. The information distribution factor of each local filter was obtained adaptively on the basis of its own error covariance matrix. Computer simulation and transplanter tests were conducted to verify the proposed method. Results showed that the proposed method provided accurate and reliable navigation information outputs and achieved better navigation performance compared with single GPS navigation and an integrated method based on a conventional federal KF. Keywords: Federal Kalman filter, GPS/INS/VNS, Information distribution factor, Information fusion, Integrated navigation.
Visual navigation method for indoor mobile robot based on extended BoW model
