Message Passing (MP), also called Belief Propagation (BP), is a well known and widely applied approach to solve problems of high complexity like MAP detection / decoding in a an approximate way, e.g., in LDCP decoding. Similarly, BP can be applied in the context of compressed sensing to achieve approximate solutions of the resulting MAP detection problems. A recent development in this field are the approximate message passing algorithms (AMP), which further reduce the complexity of BP algorithms still keeping the asymptotic ideality and the predictability of BP.

Here, the basic task is to perform a literature recap of AMP, program AMP algorithms in MATLAB and extend these algorithms to provide the necessary input and output for soft-information processing. Additionally, a diploma or master thesis would require research of iterative soft-information processing with AMP.