Team

OMNI   (OMNI)

Description

Microelectromechanical systems (MEMS) have matured enough to now be mass produced and to be integrated into consumer products. Gyroscopes and accelerometers are now included in the ABS car, like the latest mobile phones and micro-mirror arrays (DMD) are now found in most video projectors. These MEMS can be used alone as is the case of accelerometers or they can be grouped together and do work together to achieve an overall objective, as is the case of DMD. This is called {\ bf} distributed MEMS.

It is also possible to add processing capacity of information directly related to a MEMS. This can be centralized on a PC or an FPGA, but then arises the problem of scalability. Indeed, the central processing capacity becomes a bottleneck, not only in purely physical connectivity but also in software in the management of communications and processing capability. A distributed architecture solves these problems of scalability, it is called {\ bf} MEMS distributed intelligent.

MEMS are intelligent distributed objects belonging to the macroscopic world but composed of cells belonging to the microscopic world. There is a difference in scale between the object "MEMS distributed intelligent" and the component cells. Similarly, a monolithic object, in the sense that it is not composed of cells MEMS, can interact with an object "MEMS distributed intelligent" which itself communicates with its cells, it is called {\ bf networks multi- scales}.

An intelligent distributed MEMS consists of cells that comprise a sensor and / or an actuator, a calculation unit and means for wired communication or hertziens.Les intelligent distributed MEMS are classified according to their characteristics: type of MEMS (sensor, actuator, sensor / actuator), type of network (wired or wireless) and network topology (static or dynamic). The combination of these characteristics define various research issues around the MEMS distributed intelligent. However, all types of distributed intelligent MEMS have common challenges.

The first is the management of extensibility. Indeed, as MEMS are manufactured in series, it is possible for a small fee to use a large number. All algorithms must therefore take into account and anticipate the scaling which is counted in millions of nodes.

The second challenge is to manage the density of communication. Imagine that in a few liters volume it is possible to have as many links and entities on the internet. We must therefore manage the density of communication especially in the case of wireless communications.

The third challenge lies in managing communications that are subject to many errors that can bring them closer to ad hoc networks.

Finally, it is not possible to have a computing capacity equivalent between the macro world and the micro world. A calculation unit of a distributed intelligent MEMS is at most one millimeter square. The calculations must therefore be adapted accordingly.

We propose to solve these challenges through two axes:
    Modeling and information management in intelligent distributed MEMS
    Dynamic optimization, positioning and timing for the multi-scale mobility