Indoor Integrated Navigation for Differential Wheeled Robot Using Ceiling Artificial Landmark

The article proposes a vision/dead reckoning integrated indoor navigation system with low cost and high precision for differential wheeled robot. A monocular infrared sensitive camera is utilized to identify the ceiling artificial landmark, and to measure distances between blind node and anchor nodes, thus achieving vision localization. In addition, the kinematics model for differential wheeled robot is established to reckon the robot position with wheel speeds input from optical encoder sensors. Every moment, the vision localization information is combined with the estimated position from kinematics model by an extended Kalman Filter (EKF). Simulation results show that the positioning accuracy is improved from decimeter(vision localization) to centimeter level.