Emmanuel Piat

An electromagnetic micromanipulation system for single-cell manipulation

Biological objects were micromanipulated with a magnetic microactuator. These objects are pushed with a small ferromagnetic particle whose size can be as small as 10 × 10 × 5 μm3. This particle is called the manipulator and is moved thanks to a permanent magnet. This magnetic device allows the manipulation of objects in an extremely confined space. As biological objects are fragile, the force applied on them must be controlled during the manipulation. The model we present allows to determine the force applied by the device on the manipulated object. Several experimental measurements are presented in order to validate the model.

Control of a particular micro-macro positioning system applied to cell micromanipulation

Biological research requires new tools for cell micromanipulations. Currently, biological cell sizes range from a few to hundreds of micrometers, their manipulation therefore belonging to the field of microrobotics. This paper presents a new wireless micromanipulation system which allows cells placed in a droplet of liquid to be pushed on a glass slide. The cell micropusher is a ferromagnetic object which follows the movement of a permanent magnet located under the glass slide. It has been proved in previous works that two kinds of micropusher movements can induce a movement of the pushed object: turning the micropusher around the contact point (rotation), or moving the micropusher in translation. Rotation allows an object to be placed with a precision below 1 /spl mu/m, but acts within a narrow range. Translation allows placement of an object with lower accuracy, but within a wide range. We propose a specific coarse-fine control strategy to push an object, with good precision, within a wide range. Furthermore, experimentation on polystyrene balls of 50 /spl mu/m in diameter, and immature human oocytes of 150 /spl mu/m in diameter are presented. Note to Practitioners-Biological research requires new tools for cell micromanipulations. Currently, biological cell sizes range from a few hundred micrometers; their manipulation, therefore, belongs to the field of microrobotics. This paper presents a new wireless micromanipulation system which allows cells placed in a droplet of liquid to be pushed on a glass slide. The cell micropusher is a ferromagnetic object which follows the movement of a permanent magnet located under the glass slide. It has been proven in previous works that two kinds of micropusher movements can induce a movement of the pushed object: turning the micropusher around the contact point (rotation), or moving the micropusher in translation. Rotation allows an object to be placed with a precision below 1 /spl mu/m, but acts within a narrow range. Translation allows placement of an object with lower accuracy, but within a wide range. We propose a specific coarse-fine control strategy to push an object, with good precision, within a wide range. Furthermore, experimentation on polystyrene balls of 50 /spl mu/m in diameter, and immature human oocytes of 150 /spl mu/m in diameter are presented.

Smart Microrobots for Mechanical Cell Characterization and Cell Convoying

This paper deals with the effective design of smart microrobots for both mechanical cell characterization and cell convoying for in vitro fertilization. The first microrobotic device was developed to evaluate oocyte mechanical behavior in order to sort oocytes. A multi-axial micro-force sensor based on a frictionless magnetic bearing was developed. The second microrobotic device presented is a cell convoying device consisting of a wireless micropusher based on magnetic actuation. As wireless capabilities are supported by this microrobotic system, no power supply connections to the micropusher are needed. Preliminary experiments have been performed regarding both cell transporting and biomechanical characterization capabilities under in vitro conditions on human oocytes so as to demonstrate the viability and effectiveness of the proposed setups.

Passive diamagnetic levitation: theoretical foundations and application to the design of a micro-nano force sensor

Mechanical friction and more generally adhesion forces are some problems, which can severely limit the performances of micromechanical devices. One way to avoid friction problem, is to use levitation methods. Levitation in static magnetic field is very easy to achieve by the use of diamagnetic materials. Thus, it is possible to freely suspend a light magnet and let it in a stable equilibrium state. We have developed a prototype of a micro-nano force sensor using a passive levitation approach. This paper explains diamagnetic levitation in the simple case of a small cylindrical magnet with a mass and volume of 11 mg and 1.65 mm/sup 3/ respectively. The forces applied to the suspended magnet are presented and the natural stability of the diamagnetic levitation is explained. Finally, we present the design of our micro-nano force sensor using diamagnetic levitation.

Levitated micro-nano force sensor using diamagnetic materials

Under suitable conditions, diamagnetic materials allow to achieve stable levitation of permanent magnets in entirely passive configuration. Using NdFeB magnets and diamagnetic materials such as graphite in a particular configuration, we build a passive levitated force sensor with a variable stiffness and linear output. The suspended part is used as the sensing device and two directions of force measurement are possible. The absence of friction makes the sensor highly sensitive and forces around nN can be measured. The established model of both magnetic and diamagnetic forces allows to calculate the applied force on the end point of the levitating device after measuring the position of the levitating part. This paper presents the description of the levitated sensor, force calculation and experimental results.

Learning mixed behaviours with parallel Q-learning

This paper presents a reinforcement learning algorithm based on a parallel approach of the Watkinss Q-learning. This algorithm is used to control a two axis micro-manipulator system. The aim is to learn complex behaviour such as reaching target positions and avoiding obstacles at the same time. The simulations and the tests with the real manipulator show that this algorithm is able to learn simultaneously opposite behaviours and that it generates interesting action policies with regard to global path optimization.

Microfabrication and scale effect studies for a magnetic micromanipulation system

The development of the biological research requires the implementation of new micromanipulation devices of single cell. The device presented here allows the manipulation of micro-objects where diameter varies from 20 /spl mu/m to 1 mm, in a drop of water between two glass slides. Micro-objects are moved thanks to a magnetic micromanipulator. Although magnetic force is a volumic force, we prove that it is still important in the microworld compared to adhesion forces, or weight. Magnetic micromanipulators are small nickel pieces whose sizes varies from 10 /spl times/ 10 /spl times/ 5 /spl mu/m/sup 3/ to 400 /spl times/ 400 /spl times/ 20 /spl mu/m/sup 3/. We present an original microfabrication process to fabricate and to isolate the microscopic nickel pieces.

Parallel Q-learning for a block-pushing problem

This paper presents an application of reinforcement learning to a block-pushing problem. The manipulator system we used is able to push millimeter size objects on a glass slide under a CCD camera. The objective is to automate high level tasks of pushing. Our approach is based on reinforcement learning algorithm (Q-learning) because the models of the manipulator and of the dynamics of objects are unknown. The system is too complex for a classic algorithm, so we propose an original architecture which realizes several learning processes at the same time. This method produces an almost optimal policy whatever the number of manipulated objects may be. Some simulations allowed us to optimize every parameter of the learning process. In particular, they show that the more objects there are, the faster the controller learns. The experimental tests show that, after the learning process, the controller fills his part perfectly.

Micromanipulation tasks using passive levitated force sensing manipulator

We present in this paper the development of a new kind of force sensing manipulator and use it to design a teleoperated micromanipulation station. The micromanipulation tasks are the two-dimensional positioning and forces interaction measurement of micrometer-size particles (50 /spl mu/m radius) in ambient condition. The manipulator used consists of a levitated glass probe with a tip of 20 /spl mu/m size. Since the levitation is achieved with totally passive approach, no control loop is then needed. Passive levitation is possible, under suitable conditions, using permanent magnets and diamagnetic materials. The established model of both magnetic and diamagnetic forces allows to calculate the applied force on the end-effector of the levitating device after measuring the position of the levitating part. The absence of friction forces and design configuration make the manipulator highly sensitive and forces around nN can be measured. This paper reports the description of the levitated sensing manipulator, model force calculation and experimental results of micromanipulation tasks.

Nanoforce estimation based on Kalman filtering and applied to a force sensor using diamagnetic levitation

Nanoforce sensors based on passive diamagnetic levitation with a macroscopic seismic mass are a possible alternative to classical Atomic Force Microscopes when the force bandwidth to be measured is limited to a few Hertz. When an external unknown force is applied to the levitating seismic mass, this one acts as a transducer that converts this unknown input into a displacement that is the measured output signal. Because the under-damped and long transient response of this kind of macroscopic transducer cannot be neglected for time-varying force measurement, it is then necessary to deconvolve the output to correctly estimate the unknown input force. The deconvolution approach proposed in this paper is based on a Kalman filter that use an uncertain a priori model to represent the unknown nanoforce to be estimated. The main advantage of this approach is that the end-user can directly control the unavoidable trade-off that exists between the wished resolution on the estimated force and the response time of the estimation.