A line extraction algorithm for mobile robot using sonar sensor

This paper focuses on line features extraction and fusion problems in sensing the environment for indoor mobile robot with sonar sensor. A sonar model is firstly built using a large amount of sonar experimental data that is obtained through testing different ranges for the sonar sensor. Then the sonar data consisted of points detected from the sonar is classified into different line segments using improved IEPF (Iterative End-Point Fit) and variance of line slopes. An algorithm for line segment parameters estimation is performed combining Kalman filter and fusion of line segments is made use of to establish the indoor environment around the mobile. At last simulation experiment is presented for illustrating the efficiency of the proposed method including line extraction and fusion. Works in the paper would contribute to build environment in the simultaneous localization and mapping of the mobile robot.