Map Building Based on Line Feature Matching

The autonomous navigation of mobile robot is the focus on robot research. In the autonomous navigation, the robot senses the environment and their own state by sensors and realizes autonomous movement in obstacle environment. Map building and localization are two key techniques of mobile robot navigation. The map building with line feature is an effective way. There are many researches with respect to state estimation and the environment described for robot map building. Most researchers like to use the extended Kalman filter. Recently, many researchers combine particle filter and Kalman filter to improve performance. These methods primarily realize robot state estimation and do not deal with environmental characteristics. In this paper, a kind of map building method is developed based on line matching of environment features. Firstly, the motion controlling model of mobile robot is presented. The mathematics model of odometer sensor and ultrasonic sensor are established. The method of sensor filtering is proposed with Wavelet Transform including constructions of sensor Wavelet Transform model and a kind of threshold function improving de-noising effect. Map building includes line feature extraction, environmental feature matching and association and the environment map update. Experiments show the effectiveness of the method proposed in this paper.