Autonomous mobile robot navigation using stereovision

An autonomous mobile robot navigation system based on stereovision perception is developed. The surroundings are recognized by constructing local 3D maps from binocular images. Minimizing the artificial potential energy of the mobile robot on a local 3D map plans a relay position and an approaching path. A self-learning controller using adaptive fuzzy systems is designed to manipulate the dynamic behavior of the mobile robot in tracking a planned path. Using Lagrange formalism, a mathematical model describing the autonomous mobile robot is derived for simulation study. Simulation and experimental results are presented.