This paper presents our work in developing an autonomous micromanipulation system. The originality of our system is that it takes advantage of adhesion forces to grip micro-objects using an AFM (Atomic Force Microscopy) probe. A theoretical analysis of rolling conditions is carried out in order to achieve precise release of an object picked-up by adhesion. Vision control, based on the specificities of optical microscopy, and force control, based on the analysis of the AFM probe, are established. Experiments validate the employed techniques and the proposed manipulation mode.

No abstract

Abstract-This paper proposes a real-time, robust and efficient 3D model-based tracking algorithm for visual servoing. A virtual visual servoing approach is used for 3D tracking. This method is similar to more classical non-linear pose computation techniques. Robustness is obtained by integrating an Mestimator into the virtual visual control law via an iteratively reweighted least squares implementation. The presented approach is also extended to the use of multiple cameras. Results show the method to be robust to occlusion, changes in illumination and miss-tracking.

Universie Pierre et Marie Curie (paris 6) -CNRS FRE 2507 Using by Rolling 18 route du Panorama, BP61, Abstract-This paper presents our work in developing an autonomous micromanipulation system. The originality of our system is that it takes advantage of adhesion to grip microobjects by using a single fingered gripper. This is in fact a tipless cantilever previously designed for Atomic Force Microscopy applications. We describe vision techniques employed to process images provided by an optical microscope, allowing to position accurately the end-effector for a gripping task. A theoretical study of the direct force measurement device and an experimental validation show how we can improve the measurement of impact and contact forces. Then we explain the strategy used to bring the gripper into contact with the object, based on force control and kinematic redundancy. Finally, a simplified model of the release task is proposed in order to determine conditions that aliow to roll the object, and then to place it with precision.