The Research Of Binocular Measurement And Its Application In Self-adaptive Robotic Arm

Intelligence is the future direction of industrial robotics. Vision and mechanical arm are indispensable parts to ma ke it possible for industrial robotics to engage in all kinds of tasks, like human-being. This paper based on the research status of current robot technology all over the world, do research on how to make it better for vision systems and robotics systems w orking together when they are doing tasks. The target is to design a set of feasible solutions, realizing the use of robot vision complete system to identify and locate the target object, planning a reasonable manipulator trajectory to allow the robot autonomously complete tasks, such as target capture and placement through space pose transform relations.Focused on robots’ two-eyes-visual system, this topic mainly studied how to provide visual guidance for robot operation. It is follow these steps: to reco gnize and extract a target from two-dimension image through the image segmentation in human-computer interaction way, do feature detection on the divided foreground, get another view of the corresponding pixel coordinates through stereo matching algorithm, and lastly calculate the 3-D coordinates of these points using binocular measuring principle, and estimate the position and pose of target object in the visual system.Aimed at the problems that mechanical arm need to keep fast, steady, and avoid the singular configuration, this topic uses the d-h method to establish the geometric connecting rod model for all mechanical arms, finds out the kinematics equation, and then does job trajectory planning. Besides, mechanical arm has large amount of calculation an d singular configurations in Cartesian space trajectory planning. In order to avoid that problem, this method combined the Cartesian space with joint space, does the trajectory planning by introducing the NURBS algorithm, and effectively avoids the impact from the mutation of mechanical arm’s position, velocity and acceleration.To pose the next target of the visual system to the base coordinate system conversion robot requires a pre-calibrated binocular camera transformation matrix in the robot base coordinate system. In order to reduce the influence of machinery manufacturing error, installation error, etc of Vision system, this paper proposes a robot hand-eye calibration method based on mechanical arm end position, and verifies the feasibility and accurac y of this method though experiment.After accomplishing all the research on two ccd camera measure system, manipulator trajectory planning and robot hand-eye calibration, the validity and reliability of the research content are proved according to simulati on and experimental validation.