Projected-gradient-descent in rakeness-based compressed sensing with disturbance rejection

Compressed Sensing (CS) has recently emerged as an effective tool to simultaneously acquire and compress analog waveforms in low-resource sensing devices. Its mechanisms have been also extended by both adapting the sensing stage to the actual class of input signals, and granting it the ability to reject disturbances. Regrettably, the resulting design flow entails the solution of two optimization problems with a potentially huge number of variables. This work overcomes this impasse by proposing a Project-Gradient-Descend method algorithm that drastically reduces the required CPU time to obtain a solution.