In this domain, we focus our efforts on the definition of model-based composition operators dedicated to data collection policies and large-scale sensor networks. These operators allow one to reuse an existing network by deploying unforeseen applications on top of it, instead of building a new one for a single purpose. This is very important in the context of Smart Cities for example.
For example, during Cyril Cecchinel*’s PhD, we supported the automatic deployment of multiple data collection policies on the same network, sharing resources to support constraints such as minimizing the energy consumption or reducing the consumed bandwidth. We also used machine learning algorithms to predict the required collection frequencies for the involved sensors, considering this data to extend applications’ lifetime: a policy with a weekly scale lifetime was extended to a yearly scale one, thanks to an adaptive model trained on top of the composed policies.
The group also designed of two experimental platforms related to CPSs and sensor networks: (i) the SENSAPP platform, extended into the (ii) SmartCampus. These two platforms were reused by industrial partners in the context of proof of concepts or more ambitious projects (e.g., SENSAPP was used in two European projects related to cloud computing: PaaSage and MODAClouds)
Two PhD students related to CPS are working in the gorup: Sebastien Bonnieux and Sami Lazreg.