ABSTRACT
In vision-based robotic systems, robot states and environment structure are measured by visual feedback. Compared with other sensing modalities, such as Global Positioning System, ultrasonic sensors, Light Detection and Ranging, Radars, vision sensors are more descriptive and can be widely used in various environments. For vision-based robotic systems, the image to be seen in the desired place is provided, and the captured image at the view of the robot is used to provide visual feedback to the visual controller. The vision-based robotic system coordinates vision system and control system to accomplish specific tasks for robots. The visual perception module includes the estimation of robot states and the perception of environment information. In the moving process of the robot, system states are estimated by comparing the images captured at current pose and the desired pose. The visual control module includes the trajectory planning and visual servo controller to regulate the robot for task accomplishment.
