Distributed Control Architecture for smart surfaces

A
distributed control architecture is needed to perform part recognition
and close-loop control of the Smart Surface. This architecture is based
on decentralized cells able to communicate with their four neighbors
thanks to peer-to-peer links. Various original algorithms have been
proposed to reconstruct, recognize and convey the object levitating on
the Smart Surface.
Experimental results show that each algorithm does a good job for
itself and that all the algorithms together succeed in sorting
and conveying the objects to their
final destination.
Read more...
Video:
Experimental
validation (8 MB mp4 file)
Distributed Discrete State Acquisition and Concurrent Pattern Recognition

In
order to differentiate the parts put on top of the Smart Surface,
algorithms of distributed state acquisition and concurrent pattern
recognition have been developed. These algorithms have been tested in a
multithreaded Java Smart Surface Simulator, SSS, which runs on
multicore machines.
Read more...
Video:
Smart surface simulator (9 MB wmv file)
Decentralized Reinforcement Learning

Distributed-air-jet
MEMS-based systems have been proposed to manipulate small parts with
high velocities and without any friction problems. The control of
distributed-air-jet systems is very challenging and usual
approaches for contact arrayed system don't produce satisfactory
results. We investigate reinforcement learning control approaches in
order to position and convey an object. Reinforcement learning is a
popular approach to find controllers that are tailored exactly to the
system without any prior model. We show how to apply reinforcement
learning in a decentralized perspective and in order to address the
global-local trade-off.
Read
more...
Calibration of the Smart Surface

The
number of sensors that have to be embedded in the Smart Surface is a
parameter that has to be taken into account when designing the hardware
part of the Smart Surface. The Sensor
Network Calibrator (SNC) is a simulation framework which allows to
parameterize the Smart Surface and to determine the number of sensors
required on top of the Smart Surface.
Read more...
Distributed
Shape Differentiation

One
of the aims of the processing unit embedded in each cell of the Smart
Surface is to recognize the shape of the part that is put on top of the
smart surface. This recognition or more precisely this differentiation
is done through a distributed algorithm that is called a criterion. In
order to test exhaustively the efficiency of different differentiation
criteria, in terms of differentiation efficiency, memory and processing
power needed, a software framework called ECO (Exhaustive Comparison
Framework) has been developed.
Read more...
Multi-domain Simulation using VHDL-AMS

We
propose advance methods of behavioral modeling, allowing both easy
development and faster simulation for better integration of arrayed
MEMS into systems. Design and simulation are produced by solverbased
cost-effective solution using VHDL-AMS. A hierarchical circuit-level
design methodology has been followed to model and simulate a MEMS
array-based smart surface applied in air-fluid environment for 2-D
contactless micromanipulation. Using a V-shaped-based design approach,
a top-down VHDL-AMS-based modeling has been first achieved with
behavioral, structural behavioral, and component models, which include
a MEMS-based pneumatic microactuator. Then, a modeling bottom-up
approach has been developed to validate our design by comparing
simulations and experiments of the distributed surface.
Read more...