DIPP - Distributed Parallel Pursuit

We consider a distributed compressed sensing scenario where multiple sensors observe correlated sparse signals and the sensors are connected through a network. The signal correlation is realized by a partial common support-set. For such a system, the main objective of this paper is to develop a greedy pursuit algorithm. To fulfill the objective, we develop distributed parallel pursuit algorithm and provide restricted isometry property based analysis on convergence and reconstruction performance. Through simulations we show that the distributed parallel pursuit provides a good performance for a reasonably connected network.