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...