Robust compressive sampling and connections with statistics

We show that compressive sampling is–perhaps surprisingly–robust vis a vis modeling and measurement errors.