Fast and Precise Micropipette Positioning System based on Continuous Camera-Robot Recalibration and Visual Servoing
Micro-biomanipulation typically deals with the positioning, injection and delivery of foreign material into diminutive and very delicate biological structures, such as cells and early embryos. The standard tools used for these operations are glass pipettes with fine tips, which are controlled under a high magnification microscope with the help of micromanipulators. The operations require very precise pipette positioning and high motion resolution. As a result, manually performing biomanipulations is difficult, requiring extensive operator training and meticulous work under fatiguing conditions. Automation is an alternative to create fast biomanipulation systems that offer high-throughput, high consistency and high efficiency. However, challenges have to be overcome for the successful automation of such systems, including interface design for integration and control of the biomanipulation equipment; the acquisition of feedback information for automatic control; and the problem of actually controlling the system based on the obtained feedback. Here we address new strategies to deal with the later problem, which has been traditional solved through visual servoing or through the definition of mapping functions from the image space to the task space. Visual servoing enables very precise tool positioning; however, it is totally dependent on image processing algorithms, which usually limit its performance and robustness. On the other hand, the use of mapping functions between coordinate frames does not suffer from processing limitations, but suffers from drift in tool position since it is based on “blind” motion control. The proposed strategy for creating a fast and precise pipette positioning system consisted of combining these two techniques. In our system, fast motions were achieved through camera-robot mapping, and precise positioning was realized by visual servoing. This effectively eliminated the robot speed limitations imposed by the image processing system while maintaining the precision and robustness of the visual servoing system. In addition, tool localization information gathered from the vision system was used for updating the camera-robot mapping parameters continuously, which effectively eliminated the impact of tool drift on the mapping precision.
Webmaster: Micky Rakotondrabe.