3 Environment Modeling for Mobile Service Robot

An algorithm for environment modeling of mobile service robots based on the binocular stereo vision is presented. Three classical algorithms for environment modeling are emulated and compared. The structure of mobile service robots based on the stereo vision is divided into four sections: perception, modeling, planning, and action. The binocular camera is used to gather information of environment in the section of visual perception. Scale-invariant feature transform is used to collect pairs of stable feature points which are matched in one pair of pictures, and then three dimensional information of feature point will be obtained by using disparity to measure distance. The algorithms of environment modeling describe circumstance with the three dimensional information of environment feature points and match the current pictures being gained to the last pictures. The algorithms of environment modeling can guide the mobile service robots to avoid obstacles and get to the destination safely and reliably through path planning and vision navigation. This paper achieves different mapping methods through simulation experiment and gains the effect of map's use in robot vision navigation and obstacle avoidance experiment.