Multi-channel simultaneous data acquisition through a compressive sampling-based approach

Nowadays, a variety of measurement applications require the acquisition and processing of signals coming from many disseminated sensors. To reduce the cost of the overall measurement process, the above-mentioned tasks are typically performed by devices like microcontrollers, low cost data acquisition systems (DASs), field gate programmable arrays (FPGAs), and so on. This is particularly true when smart-sensors or wireless sensor networks are considered. These devices are characterized by some limitations mainly concerning with the reduced internal memory and the use of multiplexing circuits to share the same analog to digital converter (ADC) over different physical channels. The former limitation affects the maximum observation time the devices are able to cover with a single shot acquisition, the latter imposes unwanted phase shifting on the acquired signals. The use of modern acquisition and processing techniques could be of some help. One of the most promising is the compressive sampling (CS), which can assure good signal reconstruction starting from very few acquired samples. To this aim, the paper deals with the definition, implementation and experimental characterization of a CS-based acquisition approach specifically addressed to cost-effective multi-channel DASs, like those characterizing typical 8 or 32-bit microcontrollers. In particular, the paper aims at showing the reliable phase reconstruction capability also in the presence of multiplexed multichannel architectures. A number of tests conducted both on numerical and actual experiments are carried out to assess the performance of the proposed approach. In particular, the results obtained on real data acquired by means of the considered low-cost platforms shows the efficacy of the adopted approach; phase shift as low as few milliradians are, in fact, experienced