Regional feedback control of robot with application to optical manipulation of biological cells
Date of Issue2013
School of Electrical and Electronic Engineering
Centre for Intelligent Machines
Task-space sensory feedback information is used in many modern robot control systems as it improves robustness to model uncertainty. However, existing task-space sensory feedback control methods of robot are only valid locally in a finite task space within a limited sensing zone where singularity of the Jacobian matrix is avoided. The global stability problem of task-space control system has not been systemically solved. It is interesting to observe from human vision guided tasks that sensor information is not used for the entire movement, but only at end phases when our hand is near the target. We are able to move our hand from an initial position that is not within our field of view and transit smoothly and easily into visual feedback when the target is near. Moreover, we can reach and manipulate an object even when the vision is occluded. The exploration of a control method that mimics such human behaviour is an important step toward understanding dexterous movement of robot. In this research, a novel regional feedback control methodology is proposed for robots. Each feedback information is employed in a local region, and the combination of regional information ensures the convergence of robot motion. Instead of designing multiple controllers in different regions and switching between them, the regional feedback method integrates the use of dual task-space information in a single controller. The transition from one feedback information to another is embedded in the controllers without using any hard or discontinuous switching. It will be shown that the proposed control method is a unified formulation to address several open issues in task-space robot control systems, such as the singularity of the Jacobian matrix, the limited field of view of cameras, and the vision occlusion. Using the proposed regional feedback, a new task-space robot controller is proposed, which consists of a reaching task variable that drives the end effector of the manipulator from one task space to another and a desired task variable to move the end effector to the desired position at the ending stage. It enables the end effector to start from any initial position outside sensing zone and in the vicinity of singular configurations, and reach for a desired trajectory in the end. The dynamic stability of the closed-loop systems is analysed by using Lyapunov-like method. Numerous experimental results are presented to illustrate the performance of the proposed control methods. The concept of regional feedback is also extended to the optical manipulation of biological cells with robot-assisted tweezers. For the optical tweezers system, the optical trapping works only when the cell is located in a small region around the centre of the focused laser beam. To solve the problem, a unified robotic manipulation technique for optical tweezers is proposed to integrate trapping and manipulation of biological cells into a single control method. It allows the laser beam to start from an initial position that is far away from the cell and automatically trap then manipulate the cell, and it also works when the cell escapes from the optical trap during the course of manipulation. The dynamics of robotic manipulator is introduced into optical tweezers system so that a closed-loop manipulator control problem can be formulated and solved. The proposed formulation provides a theoretical framework that bridges the gap between traditional robotic manipulation techniques and optical manipulation techniques of cells.
DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics