Research of tracking robot based on vision and IMU

Considering the vulnerability of wheeled robot for the control and movement,a skid-steered mobile robot is designed and implemented in this paper,which has the characteristics of structural simplification,movement agility,high reliability and good controllability.Data fusion of monocular vision and IMU is implemented by EKF and feature detection and match of tracking object are fulfilled by SURF.The Range/Depth information as well as the information of object turning is determined combining with vertical line feature of circumstance.Experiments show that the system and the algorithm are real-time and robust.