A Class Of Mobile Robot Visual Perception And Control Technology Research

With the rapid development of robotics, mobile robots have been widely used in many areas in recent years. As aging society is coming, the intelligent mobile robots begin to service for families or others. The robots are designed to be able to do heavy work for human such as lifting a heavy object. They also can communicate with human. This paper does some research on robotics such as binocular calibration, distance measurement, path planning, object recognition and motion planning based on a mobile robot platform.This paper introduces the overall structure of the mobile robot, including control system, drive control modules, MFC programming platform and OpenCV vision library. MATLAB calibration toolbox is used to calibrate the binocular vision. With the calibration results, the robot can correct the vision, meaure the distance of obstacles or target by using OpenCV library functions, and also can build grid map depending on disparity map. After introducing global path planning methods, simulation of path planning based on the MRPT library is carried in this paper. The virtual robot can reach the destination from the original location with no collision, and the path selected is shortest.When reaching the designated area, the robot identifies the target object by using color-based image segmentation and SURF feature matching. Finally, the robot picks the object up by motion planning.