Title: Compressive Sensing: Theory, Applications and Implementation of Secure Nodes for Internet of Things


Riccardo Rovatti, University of Bologna, Italy

Gianluca Setti, University of Ferrara, Italy 


Compressed Sensing (CS) is a technique for the reconstruction of a waveform using a number of measurements that is potentially much smaller than the number of samples at the Nyquist rate. CS hinges on a simple early processing of signal samples and is the basis for the so called Analog-to-Information Converters (AIC) where one tries to match resources needed for acquisition with the actual amount of information which every sample is able to capture. Consequently, the most significant feature of these sampling procedures is that they allow a sensing node to capture the information content of a signal without going through the acquisition of its entire profile, thus performing acquisition and compression at the same time.

In summary, CS is a very simple and efficient procedure to sample signals at a reduced frequency, using (much) less resources with respect to standard sampling required for A/D conversion. These results have already caused a small revolution in the scientific community, since they have offered a completely new perspective in fundamental assumptions such as the sampling bound and the choice of the basis for the signal representation, and they are paving the way for the possibility of implementing ultra-low-power sensing nodes for applications in the framework of body-area networks or IoT.

This tutorial starts from CS basics and develops recent techniques for the joint design of hardware and algorithms for AICs based on lightweight signal adaptation. Design flows will be exemplified for analog and digital implementations with considerations on power consumption and effect of non-idealities validated by simulation and measurements.

Configurations and conditions in which AIC yields significant improvements over classical acquisition mechanisms will be identified. The discussion will also cover other advantages of the early-processing entailed by CS. For example, we will show how it is possible to embed partial but zero-cost security into the resulting acquisition processing, so that the use of dedicated cryptographic stages can be avoided in non-critical applications; finally we will also explore the possibility to embed CS in the realization of body-area-network/IoT nodes for biomedical application using commercial microcontrollers.