Benoît Larras, ISEN, Lille
LEOPAR: Low-Energy On-chip Pre-processing for Activity Recognition
The emerging “Internet of Things” and ambient intelligence growth is faced with entire data sets to be transferred from connected portable devices to central computation units that process data from multiple sensors. Decreasing the energy consumption of the devices must be a priority to increase the battery lifetime and to comply with a sustainable development line. To that end, the input data must be pre-processed on-chip in order to limit the amount of data to transmit. The LEOPAR project aims at designing the so-called pre-processing unit in order to determine the signal relevance for the targeted application, using classification functions. These functions are destined to be integrated on-chip using processing techniques in break with the state-of-the-art like simplified clique-based neural networks, leveraging the benefits of both low-connectivity analog and digital signal processing.