A practical approach for EKF-SLAM in an indoor environment: fusing ultrasonic sensors and stereo camera

摘要: Improving the practical capability of SLAM requires effective sensor fusion to cope with large uncertainties from sensors and environment. Fusing ultrasonic vision possesses advantages both economical efficiency complementary cooperation. In particular, it can resolve false data association divergence problem an sensor-only algorithm overcome low frequency update caused by computational burden weakness illumination changes a algorithm. this paper, we propose VR-SLAM (Vision Range sensor-SLAM) combine stereo camera very effectively. It consists two schemes: (1) extracting robust point line features sonar (2) recognizing planar visual objects using multi-scale Harris corner detector its SIFT descriptor pre-constructed object database. We show that fusing these schemes through EKF-SLAM frameworks achieve correct via recognition high features. The performance proposed was verified experiments in various real indoor environments.